2 Turbo PMAC
^1 USER MANUAL
^2 Turbo PMAC
^3 Programmable Multi Axis Controller
^4 3Ax-602264-TUxx
^5 October 5, 2004
Single Source Machine Control
Power // Flexibility // Ease of Use
21314 Lassen Street Chatsworth, CA 91311 // Tel. (818) 998-2095 Fax. (818) 998-7807 // www.deltatau.com
Copyright Information
© 2004 Delta Tau Data Systems, Inc. All rights reserved.
This document is furnished for the customers of Delta Tau Data Systems, Inc. Other uses are
unauthorized without written permission of Delta Tau Data Systems, Inc. Information contained
in this manual may be updated from time-to-time due to product improvements, etc., and may not
conform in every respect to former issues.
To report errors or inconsistencies, call or email:
Delta Tau Data Systems, Inc. Technical Support
Phone: (818) 717-5656
Fax: (818) 998-7807
Email: [email protected]
Website: http://www.deltatau.com
Operating Conditions
All Delta Tau Data Systems, Inc. motion controller products, accessories, and amplifiers contain
static sensitive components that can be damaged by incorrect handling. When installing or
handling Delta Tau Data Systems, Inc. products, avoid contact with highly insulated materials.
Only qualified personnel should be allowed to handle this equipment.
In the case of industrial applications, we expect our products to be protected from hazardous or
conductive materials and/or environments that could cause harm to the controller by damaging
components or causing electrical shorts. When our products are used in an industrial
environment, install them into an industrial electrical cabinet or industrial PC to protect them
from excessive or corrosive moisture, abnormal ambient temperatures, and conductive materials.
If Delta Tau Data Systems, Inc. products are exposed to hazardous or conductive materials and/or
environments, we cannot guarantee their operation
Turbo PMAC User Manual
Table of Contents
USING THIS MANUAL .............................................................................................................................................1
What is in this Manual ..............................................................................................................................................1
Other Manuals to Use................................................................................................................................................1
Turbo PMAC Software Reference Manual ...........................................................................................................1
Hardware Reference Manuals ..............................................................................................................................1
UMAC Quick Reference Guide .............................................................................................................................1
PMAC Executive Program Manual ......................................................................................................................1
Software Development Package Manuals.............................................................................................................2
TURBO PMAC FAMILY OVERVIEW ...................................................................................................................3
Turbo PMAC vs. Non-Turbo PMAC ........................................................................................................................3
Turbo PMAC vs. Turbo PMAC2 ..............................................................................................................................3
Turbo PMAC is a Computer .....................................................................................................................................3
What Turbo PMAC Does..........................................................................................................................................4
Executing Motion Programs .................................................................................................................................4
Executing PLC Programs .....................................................................................................................................4
Servo Loop Update ...............................................................................................................................................4
Commutation Update ............................................................................................................................................4
Housekeeping........................................................................................................................................................4
Communicating with the Host...............................................................................................................................4
Task Priorities.......................................................................................................................................................5
Key Hardware Components ......................................................................................................................................5
CPU Section..........................................................................................................................................................5
Machine Interface ICs...........................................................................................................................................8
Communications Ports........................................................................................................................................12
Key Software Components......................................................................................................................................13
Operational Firmware........................................................................................................................................13
Bootstrap Firmware............................................................................................................................................13
Motors, Axes, and Coordinate Systems...............................................................................................................13
User Programs....................................................................................................................................................14
Turbo PMAC Configurations..................................................................................................................................15
Board-Level Designs...........................................................................................................................................15
UMAC Rack-Mounted Designs...........................................................................................................................15
QMAC Boxed Design..........................................................................................................................................16
TURBO PMAC/PMAC2 SYSTEM CONFIGURATION AND AUTO-CONFIGURATION ............................17
CPU Clock Frequency.............................................................................................................................................17
Turbo PMAC2 System Clock Source .....................................................................................................................18
Default Clock Source ..........................................................................................................................................18
IC Clock Generation Facilities ...........................................................................................................................18
Typical Clock-Source ICs ...................................................................................................................................19
External Clock Sources.......................................................................................................................................19
Distribution of Clock Signals..............................................................................................................................19
Missing Clock Signals.........................................................................................................................................20
Re-Initialization Actions .....................................................................................................................................20
User-Customized Clock-Source Specification ....................................................................................................21
Normal Reset Actions..........................................................................................................................................22
MACRO IC Selection .............................................................................................................................................22
Dual-Ported RAM IC Selection ..............................................................................................................................23
System Configuration Status Reporting ..................................................................................................................23
Servo IC Configuration.......................................................................................................................................23
MACRO IC Configuration ..................................................................................................................................23
DPRAM IC Configuration ..................................................................................................................................24
CPU Section Configuration ................................................................................................................................24
UBUS Accessory Board Identification................................................................................................................24
Table of Contents
i
Turbo PMAC User Manual
Setting System Clock Frequencies ..........................................................................................................................24
Setting Turbo PMAC Phase and Servo Clock Frequencies ................................................................................25
Setting Turbo PMAC2 Phase and Servo Clock Frequencies ..............................................................................26
Setting I10 Servo Update Time Parameter .........................................................................................................28
Setting up a Turbo PMAC2 on the MACRO Ring..................................................................................................28
MACRO Ring Frequency Control Variables ......................................................................................................29
I7: Phase Cycle Extension ..................................................................................................................................29
I6840: MACRO IC 0 Master Configuration .......................................................................................................29
I6890/I6940/I6990: MACRO IC 1/2/3 Master Configuration ............................................................................30
I6841/I6891/I6941/I6991: MACRO IC 0/1/2/3 Node Activation Control...........................................................30
I70/I72/I74/I76: MACRO IC 0/1/2/3 Node Auxiliary Function Enable..............................................................31
I71/I73/I75/I77: MACRO IC 0/1/2/3 Node Protocol Type Control ....................................................................31
I78: MACRO Master/Slave Auxiliary Communications Timeout .......................................................................31
I79: MACRO Master/Master Auxiliary Communications Timeout.....................................................................32
I80, I81, I82: MACRO Ring Check Period and Limits .......................................................................................32
MACRO Node Addresses ....................................................................................................................................32
Resetting and Re-Initializing Turbo PMAC............................................................................................................35
Methods of Resetting...........................................................................................................................................35
Actions on a Normal Reset..................................................................................................................................35
Actions on Reset with Re-Initialization...............................................................................................................36
Actions on Reset for Firmware Reload ...............................................................................................................36
Re-Initialization and Clear Command................................................................................................................37
TALKING TO TURBO PMAC ...............................................................................................................................39
Communications Ports ............................................................................................................................................39
Serial Communications Ports .............................................................................................................................39
Bus Communications Port ..................................................................................................................................42
Dual-Ported RAM ...............................................................................................................................................44
Giving Commands to Turbo PMAC........................................................................................................................45
Turbo PMAC Processing of Commands .............................................................................................................45
Command Acknowledgement ..............................................................................................................................45
Data Response ....................................................................................................................................................45
Data Integrity......................................................................................................................................................46
Data Response Format .......................................................................................................................................46
On-Line (Immediate) Commands ...........................................................................................................................46
Types of On-Line Commands..............................................................................................................................46
Buffered (Program) Commands ..............................................................................................................................48
SETTING UP FEEDBACK AND MASTER POSITION SENSORS ...................................................................49
Setting Up Quadrature Encoders.............................................................................................................................49
Signal Format .....................................................................................................................................................49
Hardware Setup ..................................................................................................................................................49
Turbo PMAC Hardware-Control Parameter Setup ............................................................................................51
Conversion Table Processing Setup – Turbo PMAC Interface...........................................................................53
Conversion Table Processing Setup – MACRO Station Interface ......................................................................54
Scaling the Feedback Units ................................................................................................................................54
Setting Up Digital Hall Sensors ..............................................................................................................................54
Signal Format .....................................................................................................................................................54
Hardware Setup ..................................................................................................................................................55
Turbo PMAC Hardware-Control Parameter Setup ............................................................................................55
Power-Up Phasing Usage...................................................................................................................................56
Conversion Table Processing Setup – Turbo PMAC Interface...........................................................................56
Conversion Table Processing Setup – MACRO Station Interface ......................................................................56
Scaling the Feedback Units ................................................................................................................................56
Setting up Sinusoidal Encoders...............................................................................................................................57
Hardware Setup ..................................................................................................................................................57
Turbo PMAC Hardware-Control Parameter Setup ............................................................................................57
ii
Table of Contents
Turbo PMAC User Manual
Conversion Table Processing Setup – Turbo PMAC Interface...........................................................................58
Conversion Table Processing Setup – MACRO Station Interface ......................................................................59
Scaling the Feedback Units ................................................................................................................................59
Setting up Resolvers................................................................................................................................................60
Hardware Setup ..................................................................................................................................................60
Turbo PMAC Hardware-Control Parameter Setup ............................................................................................60
Conversion Table Processing Setup – Turbo PMAC Interface...........................................................................61
Conversion Table Processing Setup – MACRO Station Interface ......................................................................61
Setting Up for Power-On Absolute Position .......................................................................................................61
Scaling the Feedback Units ................................................................................................................................62
Setting Up MLDTs..................................................................................................................................................62
MLDT Interface Type..........................................................................................................................................62
Signal Formats....................................................................................................................................................62
Hardware Setup ..................................................................................................................................................63
Turbo PMAC Hardware-Control Parameter Setup ............................................................................................63
Conversion Table Processing Setup – Turbo PMAC Interface...........................................................................66
Conversion Table Processing Setup – MACRO Station Interface ......................................................................66
Setting Up for Power-On Absolute Position .......................................................................................................66
Scaling the Feedback Units ................................................................................................................................66
Setting Up LVDTs, RVDTs, & Other Analog ........................................................................................................67
Hardware Setup ..................................................................................................................................................67
Turbo PMAC Hardware-Control Parameter Setup ............................................................................................67
Conversion Table Processing Setup – Turbo PMAC Interface...........................................................................68
Conversion Table Processing Setup – MACRO Station Interface ......................................................................68
Setting Up for Power-On Absolute Position .......................................................................................................69
Scaling the Feedback Units ................................................................................................................................69
Setting Up Absolute Encoders ................................................................................................................................69
BASIC MOTOR SETUP...........................................................................................................................................71
Hardware Setup.......................................................................................................................................................71
Parameters to Set up Basic Motor Operation ..........................................................................................................71
Initial Setup Parameters ..........................................................................................................................................72
Activating the Motor: Ixx00 ................................................................................................................................72
Activating PMAC Motor Commutation: Ixx01 ...................................................................................................72
Motor Address Setup Parameters ............................................................................................................................73
Command Output Address: Ixx02 .......................................................................................................................74
Position-Loop Feedback Address: Ixx03 ............................................................................................................74
Velocity-Loop Feedback Address: Ixx04 ............................................................................................................74
Flag Addresses: Ixx25, Ixx42, Ixx43...................................................................................................................74
Flag Modes: Ixx24 ..............................................................................................................................................75
Absolute Power-Up Position Address and Format: Ixx10 and Ixx95 .................................................................76
Power-Up Mode: Ixx80.......................................................................................................................................76
Is Turbo PMAC Commutating or Closing the Current Loop for This Motor?........................................................77
Setting up Turbo PMAC for Velocity or Torque Control .......................................................................................77
Hardware Setup ..................................................................................................................................................77
Turbo PMAC Parameter Setup ...........................................................................................................................77
Setting Up Turbo PMAC2 for Pulse-and-Direction Control...................................................................................79
Hardware Setup ..................................................................................................................................................80
Signal Timing......................................................................................................................................................80
Turbo PMAC Parameter Setup ...........................................................................................................................81
Testing the Setup.................................................................................................................................................85
SETTING UP TURBO PMAC-BASED COMMUTATION AND/OR CURRENT LOOP ................................89
Beginning Setup of Commutation...........................................................................................................................89
Commutation Enable: Ixx01 ...............................................................................................................................89
Commutation Cycle Size: Ixx70 & Ixx71 ............................................................................................................89
Commutation Feedback Address: Ixx83 .............................................................................................................90
Table of Contents
iii
Turbo PMAC User Manual
Current Loop in Turbo PMAC or Not.................................................................................................................91
Setting Up for Direct PWM Control .......................................................................................................................91
Introduction ........................................................................................................................................................91
Digital Current Loop Principle of Operation .....................................................................................................91
Turbo PMAC Parameter Setup ...........................................................................................................................94
Special Instructions for Direct-PWM Control of Brush Motors .......................................................................100
Testing PWM and Current Feedback Operation ..............................................................................................101
Position Feedback and Polarity Test ................................................................................................................102
Establishing Basic Current Loop Operation.....................................................................................................107
Setting Up Turbo PMAC for Sine-Wave Output Control .....................................................................................110
Hardware Setup ................................................................................................................................................110
Turbo PMAC Parameter Setup .........................................................................................................................111
Establishing Basic Output Operation ...............................................................................................................114
Finishing Setting up Turbo PMAC Commutation (Direct PWM or Sine Wave), Synchronous Motors...............115
Confirming Commutation Polarity Match ........................................................................................................115
Establishing a Phase Reference........................................................................................................................116
Finishing Setting up Turbo PMAC Commutation (Direct PWM or Sine Wave), Asynchronous (Induction) Motors
...............................................................................................................................................................................127
Calculating Ixx78 Slip Constant .......................................................................................................................127
Setting Ixx77 Magnetization Current................................................................................................................128
Direct Microstepping of Motors (Open-Loop Commutation) ...............................................................................129
Setting the I-Variables ......................................................................................................................................130
What to Do Next ...................................................................................................................................................132
User-Written Phase Algorithms ............................................................................................................................133
Writing the Algorithm .......................................................................................................................................133
SETTING UP THE ENCODER CONVERSION TABLE ..................................................................................135
Conversion Table Execution .................................................................................................................................136
Conversion Table Structure...................................................................................................................................136
Setup Structure..................................................................................................................................................136
Result Structure.................................................................................................................................................138
Using the Results ..............................................................................................................................................139
Conversion Methods .............................................................................................................................................140
Digital Quadrature Encoders ...........................................................................................................................140
Analog Sine/Cosine Encoders...........................................................................................................................142
Acc-28 Analog-to-Digital Converters...............................................................................................................143
Parallel Feedback.............................................................................................................................................143
Exponential Filter .............................................................................................................................................145
Sum and Difference...........................................................................................................................................145
Time Base (Differentiation) ..............................................................................................................................146
SETTING UP THE SERVO LOOP.......................................................................................................................149
Servo Update Rate.................................................................................................................................................149
Reasons to Change Servo Update Rate.............................................................................................................149
Ramifications of Changing the Servo Update Rate...........................................................................................149
Types of Amplifiers ..............................................................................................................................................150
Amplifiers for Which Servo Produces Velocity Command................................................................................150
Amplifiers for Which Servo Produces Torque/Force Command.......................................................................151
PID/Feedforward/Notch Servo Filter ....................................................................................................................152
PID Feedback Filter .........................................................................................................................................152
Feedforward Filter ...........................................................................................................................................153
Actual PID/Feedforward Algorithm .................................................................................................................154
Notch Filter.......................................................................................................................................................154
Servo-Loop Modifiers .......................................................................................................................................159
Extended Servo Algorithm....................................................................................................................................161
Cascading Servo Loops.........................................................................................................................................163
Selecting Turbo PMAC Motors to Use .............................................................................................................164
iv
Table of Contents
Turbo PMAC User Manual
Inner Loop General Setup.................................................................................................................................164
Outer Loop General Setup ................................................................................................................................164
Joining the Loops..............................................................................................................................................165
Tuning the Outer Loop......................................................................................................................................166
Programming the Outer Loop Motor................................................................................................................166
Setup Example...................................................................................................................................................167
Changing the Mode of Control .........................................................................................................................167
Special Instructions for Extended Single-Loop Setup .......................................................................................168
User-Written Servo Algorithms ............................................................................................................................169
Open Servo Compiled Algorithms.....................................................................................................................169
Assembled User-Written Servo Algorithms.......................................................................................................181
MOTOR COMPENSATION TABLES AND CONSTANTS ..............................................................................183
Position Compensation Tables ..............................................................................................................................183
Source, Target, and Assigned Motors...............................................................................................................183
Dimension of the Table .....................................................................................................................................184
Using Desired vs. Actual Position ....................................................................................................................185
Multiple Tables per Motor ................................................................................................................................185
Table Range ......................................................................................................................................................186
Determining Compensation Values ..................................................................................................................186
Entering Tables.................................................................................................................................................186
Enabling and Disabling Tables ........................................................................................................................188
Active Calculation of Corrections.....................................................................................................................189
Reporting Table Information ............................................................................................................................189
Deleting Tables .................................................................................................................................................189
Backlash Compensation ........................................................................................................................................190
Constant Backlash Parameter ..........................................................................................................................190
Backlash Take-Up Rate.....................................................................................................................................190
Backlash Hysteresis ..........................................................................................................................................190
Backlash Compensation Tables ........................................................................................................................190
Enabling and Disabling Backlash.....................................................................................................................191
Backlash Table Example...................................................................................................................................192
Torque Compensation Tables................................................................................................................................193
Entering Tables.................................................................................................................................................193
Reporting Table Information ............................................................................................................................194
Enabling and Disabling Tables ........................................................................................................................194
Active Calculation of Corrections.....................................................................................................................194
Deleting Tables .................................................................................................................................................194
TURBO PMAC GENERAL PURPOSE I/O USE ................................................................................................195
Turbo PMAC General-Purpose I/O (JOPTO) Port................................................................................................195
Hardware Characteristics ................................................................................................................................195
Software Access ................................................................................................................................................195
Turbo PMAC Multiplexed I/O (JTHW) Port ........................................................................................................196
Turbo PMAC Control Panel Port ..........................................................................................................................196
Control-Panel Inputs ........................................................................................................................................196
Control-Panel Outputs......................................................................................................................................196
Turbo PMAC2 General-Purpose I/O (JIO) Port....................................................................................................197
Hardware Characteristics ................................................................................................................................197
Suggested M-Variables .....................................................................................................................................197
Direction Control..............................................................................................................................................197
Inversion Control..............................................................................................................................................198
Alternate Uses...................................................................................................................................................198
Turbo PMAC2 Multiplexed I/O Port (JTHW) ......................................................................................................198
Multiplexer Port Accessories ............................................................................................................................199
Hardware Characteristics ................................................................................................................................199
Suggested M-Variables .....................................................................................................................................199
Table of Contents
v
Turbo PMAC User Manual
Direction Control..............................................................................................................................................199
Inversion Control..............................................................................................................................................200
Alternate Uses...................................................................................................................................................200
Turbo PMAC Analog Input (JANA) Port .............................................................................................................200
Hardware Characteristics ................................................................................................................................200
Multiplexing Principle ......................................................................................................................................201
De-multiplexing I-Variables .............................................................................................................................201
UMAC Digital I/O Boards ....................................................................................................................................202
Addressing UMAC I/O Boards .........................................................................................................................202
Setting up UMAC I/O Boards ...........................................................................................................................203
MAKING AN APPLICATION SAFE ...................................................................................................................207
Following Error Limits..........................................................................................................................................207
Fatal Following Error Limit.............................................................................................................................207
Warning Following Error Limit........................................................................................................................207
Integrated Following Error Protection.............................................................................................................208
Position (Overtravel) Limits..................................................................................................................................208
Hardware Overtravel Limit Switches ...............................................................................................................208
Software Overtravel Limit Variables ................................................................................................................209
Velocity Limits......................................................................................................................................................210
Vector Velocity Limit ........................................................................................................................................210
Motor Velocity Limit.........................................................................................................................................210
Acceleration Limits ...............................................................................................................................................211
Command Output Limits.......................................................................................................................................211
2
Integrated Current (I T) Protection.......................................................................................................................211
Amplifier Enable and Fault Lines .........................................................................................................................213
Encoder-Loss Detection ........................................................................................................................................213
User-Written Safety Algorithms ...........................................................................................................................215
Watchdog Timer....................................................................................................................................................215
Actions on Watchdog Timer Trip ......................................................................................................................215
Diagnosing Cause of Watchdog Timer Trip .....................................................................................................215
Hardware Stop Command Inputs ..........................................................................................................................216
Host-Generated Stop Commands ..........................................................................................................................216
Program Checksums..............................................................................................................................................217
Firmware Checksum .........................................................................................................................................217
User-Program Checksum..................................................................................................................................217
Communications Data Integrity ............................................................................................................................218
EXECUTING INDIVIDUAL MOTOR MOVES .................................................................................................219
Jogging Move Control...........................................................................................................................................219
Jog Acceleration ...............................................................................................................................................219
Jog Speed ..........................................................................................................................................................219
Jog Commands..................................................................................................................................................221
Triggered Motor Moves ........................................................................................................................................222
Types of Triggered Moves.................................................................................................................................222
Types of Trigger Conditions .............................................................................................................................222
Capturing the Trigger Position.........................................................................................................................223
Post-Trigger Move............................................................................................................................................224
Homing Search Moves ......................................................................................................................................224
Jog-Until-Trigger Moves ..................................................................................................................................231
Motion Program Move-Until-Trigger...............................................................................................................232
Open-Loop Moves.................................................................................................................................................233
TURBO PMAC COMPUTATIONAL FEATURES.............................................................................................235
Computational Priorities .......................................................................................................................................235
Single Character I/O.........................................................................................................................................235
Commutation Update ........................................................................................................................................235
vi
Table of Contents
Turbo PMAC User Manual
Servo Update ....................................................................................................................................................235
Real-Time Interrupt Tasks ................................................................................................................................236
VME Mailbox Processing .................................................................................................................................236
Background Tasks.............................................................................................................................................236
Priority Level Optimization ..............................................................................................................................238
Evaluating the Turbo PMAC’s Computational Load ............................................................................................238
Phase Interrupt Tasks .......................................................................................................................................238
Servo Interrupt Tasks........................................................................................................................................239
Real-Time Interrupt Tasks ................................................................................................................................239
Total Interrupt Tasks ........................................................................................................................................239
Sample Monitoring Program ............................................................................................................................239
Background Cycle Time....................................................................................................................................240
Numerical Values..................................................................................................................................................240
Internal Formats ...............................................................................................................................................240
Receiving Values...............................................................................................................................................241
Reporting Values...............................................................................................................................................241
Addresses ..............................................................................................................................................................242
Variables ...............................................................................................................................................................242
I-Variables ........................................................................................................................................................242
P-Variables .......................................................................................................................................................243
Q-Variables.......................................................................................................................................................244
M-Variables ......................................................................................................................................................245
Operators ...............................................................................................................................................................247
Arithmetic Operators ........................................................................................................................................247
Logical Operators.............................................................................................................................................248
Functions ...............................................................................................................................................................248
Expressions ...........................................................................................................................................................248
The {DATA} Syntax .............................................................................................................................................248
Variable Value Assignment Statement..................................................................................................................249
I-Variable Default Value Assignment ...............................................................................................................249
Synchronous M-Variable Value Assignment.....................................................................................................249
Comparators ..........................................................................................................................................................251
Conditions .............................................................................................................................................................252
Simple Conditions .............................................................................................................................................252
Compound Conditions ......................................................................................................................................252
Single-Line Condition Actions ..........................................................................................................................252
Multiple-Line Conditions ..................................................................................................................................253
Timers ...................................................................................................................................................................253
Computational Considerations ..............................................................................................................................253
SETTING UP A COORDINATE SYSTEM..........................................................................................................255
What is a Coordinate System?...............................................................................................................................255
What is an Axis? ...................................................................................................................................................255
Single-Motor Axes.............................................................................................................................................255
Multiple-Motor Axes .........................................................................................................................................256
Phantom Axes ...................................................................................................................................................256
Axis Definition......................................................................................................................................................256
Matching Motor to Axis ....................................................................................................................................256
Scaling and Offset.............................................................................................................................................256
Axis Types .........................................................................................................................................................257
Conversion from Axis to Motor Position ..........................................................................................................258
Conversion from Motor to Axis Positions.........................................................................................................258
Coordinate-System Kinematic Calculations..........................................................................................................259
Creating the Kinematic Program Buffers .........................................................................................................260
Coordinate System Transformations with Kinematics ......................................................................................267
Executing the Kinematic Programs ..................................................................................................................267
Table of Contents
vii
Turbo PMAC User Manual
Coordinate System Time-Base..............................................................................................................................268
WRITING AND EXECUTING MOTION PROGRAMS....................................................................................271
Sequenced Motion Program Execution .................................................................................................................271
Flow Control .........................................................................................................................................................271
G-Codes ................................................................................................................................................................271
Modal Commands .................................................................................................................................................272
Move Commands ..................................................................................................................................................272
Motion Program Trajectories ................................................................................................................................272
Linear Blended Moves ..........................................................................................................................................272
Position or Distance Specification....................................................................................................................272
Feedrate or Move-Time Specification...............................................................................................................272
Acceleration Parameters ..................................................................................................................................274
Acceleration Limits ...........................................................................................................................................275
Minimum Move Time ........................................................................................................................................276
Maximum Move Time........................................................................................................................................277
The Blending Function......................................................................................................................................277
Circular Blended Moves........................................................................................................................................282
Specifying the Interpolation Plane....................................................................................................................282
Circle Modes.....................................................................................................................................................282
Center Vector....................................................................................................................................................283
Radius Size Specification ..................................................................................................................................284
No Center Specification ....................................................................................................................................285
Vector Feedrate Axes........................................................................................................................................285
Circle-Radius Errors ........................................................................................................................................285
Move Segmentation Mode.................................................................................................................................285
Rapid-Mode Moves...............................................................................................................................................286
Move Time ........................................................................................................................................................286
Move Path.........................................................................................................................................................286
No Blending ......................................................................................................................................................287
Motion Program Move-Until-Trigger...............................................................................................................287
Altered-Destination Moves ...............................................................................................................................288
Spline-Mode Moves ..............................................................................................................................................289
How They Work ................................................................................................................................................289
Added Pieces.....................................................................................................................................................290
Quantifying the Position Adjustment ................................................................................................................290
5-Point Spline Correction .................................................................................................................................290
Non-Uniform Spline..........................................................................................................................................290
PVT-Mode Moves.................................................................................................................................................292
Mode Statement.................................................................................................................................................292
Move Statements ...............................................................................................................................................292
Turbo PMAC Calculations ...............................................................................................................................292
Problems in Stepping ........................................................................................................................................292
Use of PVT to Create Arbitrary Profiles...........................................................................................................292
Use of PVT in Contouring.................................................................................................................................293
Cutter Radius Compensation.................................................................................................................................294
Defining the Plane of Compensation ................................................................................................................294
Defining the Magnitude of Compensation ........................................................................................................295
Turning On Compensation................................................................................................................................295
Turning Off Compensation ...............................................................................................................................295
How Turbo PMAC Introduces Compensation ..................................................................................................295
Treatment of Compensated Inside Corners.......................................................................................................297
Treatment of Outside Corners ..........................................................................................................................297
Treatment of Full Reversal ...............................................................................................................................299
Note on Full Circles..........................................................................................................................................299
Speed of Compensated Moves...........................................................................................................................300
viii
Table of Contents
Turbo PMAC User Manual
Changes in Compensation ................................................................................................................................300
Failures in Cutter Compensation......................................................................................................................302
Block Buffering for Cutter Compensation.........................................................................................................303
Single-Stepping While in Compensation...........................................................................................................304
Three-Dimensional Cutter Radius Compensation.................................................................................................305
Defining the Magnitude of 3D Compensation ..................................................................................................305
Turning on 3D Compensation...........................................................................................................................305
Turning Off 3D Compensation..........................................................................................................................306
Declaring the Surface-Normal Vector ..............................................................................................................306
Declaring the Tool-Orientation Vector.............................................................................................................306
How 3D Compensation is Performed ...............................................................................................................307
Turbo PMAC Lookahead Function.......................................................................................................................307
Quick Instructions: Setting Up Lookahead .......................................................................................................309
Detailed Instructions: Setting up to use Lookahead .........................................................................................310
Running a Program with Lookahead ................................................................................................................315
Stopping While In Lookahead...........................................................................................................................317
Reversal While in Lookahead ...........................................................................................................................319
Axis Transformation Matrices...............................................................................................................................320
Setting up the Matrices .....................................................................................................................................321
Using the Matrices............................................................................................................................................321
Calculation Implications...................................................................................................................................322
Examples...........................................................................................................................................................322
Entering a Motion Program...................................................................................................................................323
Learning a Motion Program.............................................................................................................................324
Motion Program Structure................................................................................................................................324
Running a Motion Program ..............................................................................................................................328
Implementing a Machine-Tool Style Program ......................................................................................................329
G, M, T, and D-Codes.......................................................................................................................................329
Standard G-Codes.............................................................................................................................................329
Spindle Programs .............................................................................................................................................333
Standard M-Codes ............................................................................................................................................334
Default Conditions............................................................................................................................................336
Rotary Motion Program Buffers............................................................................................................................337
Defining a Rotary Buffer...................................................................................................................................337
Preparing to Run ..............................................................................................................................................337
Opening for Entry .............................................................................................................................................337
Staying Ahead of Executing Line ......................................................................................................................338
Closing and Deleting Buffers............................................................................................................................338
How PMAC Executes a Motion Program .............................................................................................................339
Calculation of Subsequent Moves .....................................................................................................................340
When No Calculation Ahead.............................................................................................................................341
SYNCHRONIZING TURBO PMAC TO EXTERNAL EVENTS......................................................................345
Position Following (Electronic Gearing)...............................................................................................................345
Position Following I-Variables.........................................................................................................................345
Changing Ratios on the Fly ..............................................................................................................................346
Superimposing Following on Programmed Moves...........................................................................................346
External Time-Base Control (Electronic Cams)....................................................................................................347
What Is Time-Base Control?.............................................................................................................................347
Instructions for Using an External Time-Base Signal ......................................................................................348
Time-Base Example ..........................................................................................................................................350
Triggered Time Base.........................................................................................................................................352
Synchronizing Turbo PMAC to other Turbo PMACs...........................................................................................354
Clock Timing.....................................................................................................................................................354
Motion Program Timing ...................................................................................................................................356
Hardware Position-Capture Functions ..................................................................................................................357
Table of Contents
ix
Turbo PMAC User Manual
Requirements for Hardware Capture ...............................................................................................................357
Setting the Trigger Condition ...........................................................................................................................357
Automatic Move-Until-Trigger Functions ........................................................................................................358
Manual Use of the Capture Feature .................................................................................................................358
Converting to Motor and Axis Coordinates ......................................................................................................361
Using the Position-Compare Feature on Turbo PMAC.........................................................................................362
Scaling and Offset of Position-Compare Registers...........................................................................................362
Setup on a PMAC-Style Servo IC......................................................................................................................362
Setup on a PMAC2-Style Servo IC....................................................................................................................363
Converting from Motor and Axis Coordinates..................................................................................................368
Synchronous M-Variable Assignment Outputs.....................................................................................................369
WRITING AND EXECUTING PLC PROGRAMS.............................................................................................371
What are PLC Programs? ......................................................................................................................................371
When To Use.....................................................................................................................................................371
Common Uses ...................................................................................................................................................371
64 PLC Programs .............................................................................................................................................371
Entering a PLC Program .......................................................................................................................................371
Opening the Buffer............................................................................................................................................372
Downloading the Program ...............................................................................................................................372
Closing the Buffer .............................................................................................................................................372
Erasing the Program ........................................................................................................................................372
Example ............................................................................................................................................................372
PLC Program Structure .........................................................................................................................................373
Calculation Statements .....................................................................................................................................373
Conditional Statements .....................................................................................................................................373
Compiled PLC Programs ......................................................................................................................................375
Execution of Compiled PLCs ............................................................................................................................376
Writing Compiled PLC Programs.....................................................................................................................376
Optimization for Speed .....................................................................................................................................380
Memory Utilization ...........................................................................................................................................380
Compiling the PLCs..........................................................................................................................................381
Running Compiled PLCs...................................................................................................................................382
WRITING A HOST COMMUNICATIONS PROGRAM...................................................................................383
Turbo PMAC Command/Response Format ..........................................................................................................383
Response Types .................................................................................................................................................383
Variations .........................................................................................................................................................384
Clearing the Port ..............................................................................................................................................384
Serial Port Communications..................................................................................................................................385
Setting Up the Interface ....................................................................................................................................385
Sending a Character .........................................................................................................................................385
Reading a Character.........................................................................................................................................385
ISA/PCI Host Port Communications.....................................................................................................................386
Host Port Structure...........................................................................................................................................386
Register Functions ............................................................................................................................................386
Registers for Simple Polled Communications...................................................................................................386
Setting up the Port ............................................................................................................................................387
Sending a Character .........................................................................................................................................387
Reading a Character.........................................................................................................................................387
ISA/PCI Interrupts ............................................................................................................................................387
Initializing the Interrupt Controller..................................................................................................................391
VME Bus Communications ..................................................................................................................................391
Setting up VME Communications .....................................................................................................................392
VME Mailbox Register Communications..........................................................................................................392
Dual-Ported RAM Communications .....................................................................................................................397
Physical Configuration and Connection...........................................................................................................398
x
Table of Contents
Turbo PMAC User Manual
Host Address Setup ...............................................................................................................................................398
ISA Bus Setup....................................................................................................................................................398
VME Bus Setup .................................................................................................................................................399
Mapping of Memory Addresses ........................................................................................................................401
DPRAM Automatic Functions...........................................................................................................................402
DPRAM Data Format.......................................................................................................................................402
DPRAM Motor Data Reporting Buffer .............................................................................................................403
DPRAM Background Data Reporting Buffer....................................................................................................404
DPRAM ASCII Communications ......................................................................................................................405
DPRAM Communications Interrupts ................................................................................................................406
DPRAM Background Variable Read Buffer .....................................................................................................407
DPRAM Background Variable Data Write Buffer............................................................................................409
DPRAM Binary Rotary Program Transfer Buffers...........................................................................................410
DPRAM Data Gathering Buffer........................................................................................................................412
Turbo PMAC Ethernet Protocol............................................................................................................................412
PMAC Ethernet Protocol Command Packet Description .................................................................................413
Turbo PMAC Ethernet Protocol Command Set ................................................................................................414
Data Gathering ......................................................................................................................................................420
Table of Contents
xi
Turbo PMAC User Manual
xii
Table of Contents
Turbo PMAC User Manual
USING THIS MANUAL
This is the User Manual for the Turbo PMAC family of motion and machine controllers from Delta Tau
Data Systems, Inc. The Turbo PMAC family combines power, flexibility, and ease of use in a wide
variety of configurations to provide optimal solutions for machine builders.
What is in this Manual
This manual contains a step-by-step guide to setting up the Turbo PMAC application, starting with the
overall system configuration, moving on to the setup of the individual motors and the grouping of the
motors into coordinate systems, followed by the writing of motion programs, PLC programs, and host
communications programs.
The User Manual is divided into sections that are put in the order that most new users will need during the
basic setup of their application. In the later stages of development, the typical user will be going back and
forth in this manual for more detailed information about particular subjects of interest.
Other Manuals to Use
Other manuals should be consulted during the development of the Turbo PMAC project. The most
important of these manuals are listed below:
• Turbo PMAC Software Reference Manual
• Hardware Reference Manuals
• UMAC Quick Reference Guide
• PMAC Executive Program Manual
• Software Development Package Manuals
Turbo PMAC Software Reference Manual
The Software Reference Manual for the Turbo PMAC family of controllers contains detailed descriptions
of all commands and setup variables, listed in alphabetical or numerical order for easy reference. It also
contains a detailed memory and I/O map for the Turbo PMAC family, plus suggested M-variable pointer
definitions and a list of updates to the Turbo PMAC embedded firmware.
If there is a specific question about the implementation of a particular Turbo PMAC setup variable,
command, or register, consult the appropriate section of the Software Reference Manual for the quickest
answer.
Hardware Reference Manuals
A particular Turbo PMAC hardware configuration will have its own Hardware Reference Manual for the
controller, and any accessories will each have its own Hardware Reference (User) Manual. These will be
among the first manuals to consult, because they explain how to configure and install the Turbo PMAC.
UMAC Quick Reference Guide
New users of the UMAC rack-mounted configuration of the Turbo PMAC family will want to consult the
UMAC Quick Reference Guide. This guide provides a good overview of the hardware and software
setup for the UMAC and its accessories, so less use is required of the more detailed manuals.
PMAC Executive Program Manual
Virtually all users will utilize a PMAC Executive program on a PC to start the development, even if there
will not be a PC in the final application. The most commonly used Executive program as of this writing
is Pewin32 Pro (PMAC Executive program for 32-bit Windows, Pro Suite). While the installation and
use of this program suite is very intuitive and in accordance with industry standards, consult the manual
for this package.
Using this Manual
1
Turbo PMAC User Manual
Actually, the Pewin32 Pro package is a suite of software programs, including step-by-step tutorial setup
programs, tuning programs (interactive and auto-tuning), and plotting programs.
The Pewin32 Pro suite automates or simplifies many of the setup steps that are explained in detail for
low-level setup in this manual. Utilize the Pro Suite tools for automated setup wherever possible.
Software Development Package Manuals
Delta Tau offers several software development packages to enable the quick and easy development of
front-end software for communications with the Turbo PMAC and interface with the machine operator.
Each of these packages has its own manual.
The two most commonly used packages are the PComm32 Pro (PMAC Communications Library for 32bit Windows, Pro suite) library of communications routines, and the PHMI (PMAC Human-Machine
Interface) GUI development package.
2
Using this Manual
Turbo PMAC User Manual
TURBO PMAC FAMILY
The Turbo PMAC family of controllers is the newest generation of motion and machine controllers from
Delta Tau Data Systems. It is available in a wide variety of configurations, permitting the user to
optimize the controller hardware and software to particular application needs. This section provides a
brief overview of the Turbo PMAC structure; all items mentioned here are covered in more detail
elsewhere in the User Manual or in related reference manuals.
Turbo PMAC vs. Non-Turbo PMAC
The Turbo PMAC family of controllers is the next generation after the original PMAC family of
controllers, differing only in having a faster and more powerful CPU section, but able to maintain the
same interface circuitry. Indeed, in some configurations it is possible to remove the old non-Turbo
PMAC CPU board from a controller and install the Turbo PMAC CPU as a field upgrade.
The Turbo PMAC family retains the same capabilities and programming styles as the original PMAC
family, but adds important new capabilities as well. These new capabilities are fully listed in a section of
the Software Reference manual, and detailed in the appropriate sections of the User Manual and Software
Reference; highlights are listed here:
• 32-motor capability vs. 8-motor
• 16-coordinate-system capability vs. 8-coordinate-system
• Forward and inverse kinematic algorithms
• Standard multi-move lookahead algorithms
• 32,768 user variables vs. 4096
• Simultaneous communication over multiple ports
Turbo PMAC vs. Turbo PMAC2
As with the original PMAC family, Turbo PMACs can come as Turbo PMAC or Turbo PMAC2
controllers. The difference is due to the nature of the “core” servo-interface ICs and in a few aspects of
the firmware used to set them up. Fundamentally, PMAC2-style Servo ICs permit programmable
configuration of the frequency of the key clock signals that drive the hardware and software processes on
the controllers; PMAC-style Servo ICs do not.
Unlike the older non-Turbo PMACs, it is possible to mix PMAC-style and PMAC2-style Servo ICs in a
single Turbo PMAC system. The Turbo PMAC CPU will automatically recognize which types of ICs are
present, and configure the setup I-variables appropriately.
In this documentation, Turbo PMAC refers generically to both Turbo PMAC and Turbo PMAC2
controllers. Turbo PMAC and Turbo PMAC2 refer specifically to one type of controller. If a specific
product name (e.g. Turbo PMAC-PC) does not specify 1 or 2, it is a PMAC controller.
Turbo PMAC is a Computer
It is important to realize that Turbo PMAC is a full computer in its own right, capable of standalone
operation with its own stored programs. Furthermore, it is a real-time, multitasking computer that can
prioritize tasks and have the higher priority tasks pre-empt those of lower priority (most personal
computers are not capable of this).
Even when used with a host computer, the communications should be thought of as those from one
computer to another, not as computer to peripheral. In these applications, Turbo PMAC’s ability to run
multiple tasks simultaneously, properly prioritized, can take a tremendous burden off the host computer
(and its programmer), both in terms of processor time, and of task-switching complexity.
Turbo PMAC Family Overview
3
Turbo PMAC User Manual
What Turbo PMAC Does
Turbo PMAC can handle all of the tasks required for machine control, constantly switching back and
forth between the different tasks thousands of times per second. The major tasks involved in machine
control are summarized here.
Executing Motion Programs
The most obvious task of Turbo PMAC is executing sequences of motions given to it in a motion
program. When told to execute a motion program, Turbo PMAC works through the program one move at
a time, performing all the calculations up to that move command (including non-motion tasks) to prepare
for actual execution of the move. Turbo PMAC is always working ahead of the actual move in progress,
so it can blend properly into the upcoming move, if required. See the Writing a Motion Program section
for more details.
Executing PLC Programs
The sequential nature of motion program suits it well for commanding a series of moves and other
coordinated actions; however these programs are not good at performing actions that are not directly
coordinated with the sequence of motions. For these types of tasks, Turbo PMAC provides the capability
for users to write “PLC programs.” These are named after Programmable Logic Controllers because they
operate in a similar manner, continually scanning through their operations as fast as processor time
allows. These programs are very useful for any task that is asynchronous to the motion sequences. See
the Writing a PLC Program section for more details.
Servo Loop Update
In an automatic task that is essentially invisible to the Turbo PMAC user, Turbo PMAC performs a servo
update for each motor at a fixed frequency (usually around 2 kHz). The servo update for a motor consists
of incrementing the commanded position (if necessary) according the equations generated by the motion
program or other motion command, comparing this to the actual position as read from the feedback
sensor, and computing a command output based on the difference. This task occurs automatically without
the need for any explicit commands. See the Setting up the Servo Loop section for more details.
Commutation Update
If Turbo PMAC is requested to perform the commutation for a multiphase motor, it will perform
commutation updates automatically at a fixed frequency (usually around 9 kHz). The commutation, or
phasing, update for a motor consists of measuring and/or estimating the rotor magnetic field orientation,
then apportioning the command that was calculated by the servo update among the different phases of the
motor. This task occurs automatically without the need for any explicit commands. See the Setting up
Commutation section for more details.
Housekeeping
Turbo PMAC regularly and automatically performs housekeeping tasks that make sure the system is in
good working order. These tasks include the safety checks, such as following error limits, hardware
overtravel limits, software overtravel limits, and amplifier faults, plus general status updates. They also
include the update of the watchdog timer. If any problem in hardware or software keeps these tasks from
executing, the watchdog timer will trip, and the card will shut down. See the Making an Application Safe
section for more details.
Communicating with the Host
Turbo PMAC can communicate with one or more host computers at any time, even in the middle of a
sequence of motions. Turbo PMAC will accept a command, and take the appropriate action – putting the
command in a program buffer for later execution, providing a data response to the host, starting a motor
move, etc. If the command is illegal, it will report an error to the host.
4
Turbo PMAC Family Overview
Turbo PMAC User Manual
Task Priorities
These tasks are ordered in a priority scheme that has been optimized to keep applications running
efficiently and safely. While the priority levels are fixed, the frequency at which various tasks are
performed is under user control. See Computational Features for more details.
Key Hardware Components
A Turbo PMAC controller is a combination of a computer processor section and specialized interface
circuitry for motion, I/O, and communications. Different configurations of Turbo PMAC controllers
provide different combinations of these components. The following block diagram shows the basic
possibilities for combinations of the concepts in PMAC and Turbo PMAC control systems.
PMAC Family Main Components
Serial Communications
Interface
y
y
I/O Interface
Circuitry
CPU
y
y
y
PMAC(1)-style
PMAC2-style
RS-232
and / or
RS-422
y
y
y
y
y
y
PMAC(1)–style ICs
PMAC2–style ICs
PMAC(1)–style connectors
PMAC2–style connectors
Mini / Lite / (full)
Non-Turbo: DSP5600x
(16-bit addressing)
Turbo: DSP5630x
(24-bit addressing)
Bus Communications
Interface
y
y
y
y
Servo Interface
Circuitry
MACRO Ring
Interface
y
y
y
PMAC2 Ultralite
ACC-24P
ACC-5E (UMAC)
ISA / VME / PCI
PC104 / STD
USB
ETHERNET
(optional DPRAM)
CPU Section
The computer portion of a Turbo PMAC is often called the CPU (central processing unit). It consists of a
microprocessor, memory, and associated support circuitry. In most Turbo PMAC configurations, the
CPU is on a separate circuit board from the interface circuitry – a piggyback board on top of expansionslot controllers, or a dedicated board in the case of the rack-mounted UMAC controllers. However, in the
Lite and Ultralite controllers, this circuitry is integrated into the main circuit board.
Processor
Turbo PMAC controllers use a processor from Motorola’s DSP56300 family of digital signal processors
(DSPs). A DSP is a special type of microprocessor optimized for fast and repeated mathematical
operations of the type found commonly in audio, video, and motion control. They provide a more costeffective solution to these specific types of computations than do general-purpose microprocessors.
Different members of the DSP56300 family are used depending on the CPU option ordered for the Turbo
PMAC.
Turbo PMAC Family Overview
5
Turbo PMAC User Manual
•
Option 5Cx (the x specifies the external memory size – see below) provides an 80 MHz DSP56303
with 8k 24-bit words of internal memory. This is the default processor.
• Option 5Dx provides a 100 MHz DSP56309 with 34k 24-bit words of internal memory.
• Option 5Ex provides a 160 MHz DSP56311 with 128k 24-bit words of internal memory. These
options require firmware revision V1.939 or newer.
• Option 5Fx (expected release in 3rd quarter 2003) provides a 240 MHz DSP56321 with 192k 24-bit
words of internal memory. These options require firmware revision V1.940 or newer.
Not all of these options are available in all Turbo PMAC configurations.
Note:
Just because a processor is capable of operating at a particular frequency does not
mean that it will actually be operating at that frequency. The value of Turbo
PMAC variable I52 at power-up/reset controls what the operating frequency of the
processor will be: 10 MHz * (I52+1). The default value of I52 is 7, for 80 MHz
operation, regardless of the CPU speed option ordered. The TYPE command will
show the frequency at which the CPU is actually operating.
To locate the CPU part number, issue the CPU on-line command. Turbo PMAC will respond with the
part number (e.g. 56311). Internally, global status bits X:$000006 bit 21 and Y:$000006 bit 21 (part of
the ??? global status query word) indicate which type of CPU is present.
Active Memory
Turbo PMAC uses static RAM (SRAM) ICs for its active memory. This type of RAM is faster and more
robust than the dynamic RAM (DRAM) that forms the bulk of the PC’s memory. (SRAM ICs are used
for the PC’s fast cache memory bank.)
As with any RAM ICs, the contents of these SRAM ICs (as well as memory registers internal to the DSP,
and memory-mapped registers in the ASICs) are lost when power is removed. Settings that should be
retained through a power-down must first be copied to non-volatile flash memory with the SAVE
command.
The Motorola DSPs employ a Harvard architecture, which uses separate memory banks for program
(compiled or assembled machine code instructions) storage and data (everything else, including
interpreted program commands) storage. This is in contrast to the more common von Neumann
architecture that the PC uses, in which any memory can be used for program or data storage.
The standard memory configuration for a Turbo PMAC (Option 5x0, where “x” specifies the CPU speed,
as explained above) provides 128k 24-bit words of program memory, and 128k 24-bit words (organized
as 64k 48-bit words) of data memory, in addition to what is internal to the DSP. This is the default
configuration. With the standard memory configuration, the total addressable memory is the sum of
memory internal to the DSP and the external memory in the SRAM ICs; getting a DSP with more internal
memory adds to your total memory capacity.
If the extended memory configuration is ordered (Option 5x3), the Turbo PMAC comes with a total of
512k 24-bit words of program memory, and 512k 24-bit words (organized as 256k 48-bit words) of data
memory. With the extended memory configuration, the total addressable memory is limited by the
addressing space of the DSP; getting a DSP with more internal memory does not add to the total memory
capacity (although it does substitute faster internal memory for slower external memory).
6
Turbo PMAC Family Overview
Turbo PMAC User Manual
Status variable I4908 contains the address of the next register past last word of unreserved data memory.
If no UBUFFER has been defined, this is the address one greater than the last word of data memory.
With the standard memory configuration, no UBUFFER is defined on re-initialization. With the extended
memory configuration, a UBUFFER of 65,536 words – from X/Y:$030000 - $03FFFF – is defined
automatically on re-initialization.
Flash Memory
Turbo PMAC’s non-volatile memory storage is contained in a flash-memory IC. Flash-memory ICs
retain their contents without applied power. However, they are relatively slow in access times, so Turbo
PMAC does not use the flash IC in ongoing operation. Instead, during power-up/reset, it copies the
contents of the flash memory into fast SRAM so the processor can access it quickly. The flash memory is
only written to upon a SAVE command, or when new firmware is downloaded.
The flash memory provides non-volatile storage for both the firmware created by Delta Tau, and the
programs, variable values, and other settings created by the user. The flash memory IC is sized
depending on the active memory capacity so as to be able to store the entire contents of the active
memory. Status variables I4904 and I4909 both contain information as to what size of flash memory is
present.
Copying user settings to flash memory with the SAVE command takes several seconds during which some
other tasks, including several safety checks, are not performed; this should not be done while the machine
is in actual operation. As Turbo PMAC is copying saved data from flash memory to active SRAM during
a power-up/reset, it is evaluating the checksum of this data and comparing it to the checksum calculated
during the last SAVE operation. If there is a discrepancy, it will revert the settings in active memory to
factory default values, as if the card had been re-initialized, and set the Flash Read Error (a.k.a. “EAROM
error) status bit at X:$000006 bit 21, part of the ??? global status query word.
Optional Battery-Backed Parameter Memory
If Option 16A is ordered, the Turbo CPU section is provided with an additional bank of 32k 24-bit words
(mapped as 16k 48-bit words) of data-memory SRAM that is battery-backed (BBRAM) and can be used
for parameter storage. Typically, this memory bank is used to store machine-state information without
the need for a lengthy SAVE command.
The low-power SRAM that can be powered from batteries for lengthy periods is slower than the SRAM
ICs used for the main memory, so the user should be aware that significant use of this memory bank
might incur a noticeable computational speed penalty. P-variable and/or Q-variable storage can be moved
from fast main memory to the battery-backed memory by changing I46 from its default value of 0. This
provides automatic retention of the variable values at the cost of slower access. More commonly,
registers in the BBRAM are just accessed with M-variable as pointers. BBRAM registers start at
X/Y:$050000.
Although the expected average battery life for the lithium battery that retains this memory is five years or
more, a yearly replacement schedule is recommended. There is a super-capacitor capable of retaining the
contents of the memory for several minutes without the battery, so a battery change can be done without
power applied to the controller.
Turbo PMAC register X:$00003F contains the address of the end register of BBRAM. It will report a
value of $050000 if there is no BBRAM present. It will report a value of $054000 if the Option 16A
BBRAM is present. It will report a value of $060000 if extended BBRAM is present (Turbo PMAC
designs support this but it is not being sold at this time).
Turbo PMAC Family Overview
7
Turbo PMAC User Manual
Machine Interface ICs
The Turbo PMAC CPU communicates to the physical machine through several types of special ICs that
have memory-mapped registers for easy processor access, and application-specific circuitry for the
machine interface. The most common of these are the Servo ICs, the MACRO ICs, and the I/O ICs,
application-specific ICs (ASICs) designed by Delta Tau and manufactured in gate array technology to
create a full feature set in a cost-effective manner.
Servo ICs
The Servo ICs contain all of the digital logic to provide the interface between the CPU and the motion
(servo or stepper) channels. Each Servo IC provides the interface to four motion channels. There are
presently two types of Servo ICs, one supporting the older PMAC-style interface (analog amplifier
interface only), and the other supporting the newer PMAC2-style interface (analog or digital amplifier
interfaces).
Servo Channels: Servo channels are a hardware structure in Turbo PMAC systems, a set of interfaces
and registers in the Servo ICs and surrounding systems. A channel consists of the interface and registers
for a single amplifier, encoder, and set of flags. While these channels are usable by Turbo PMAC motors,
axes, and coordinate systems (which are software structures – see Key Software Components below) for
various purposes, they can exist independently of any of those structures.
Reporting Servo ICs Present: Turbo PMAC variable I4900 reports how many Servo ICs are present,
and at which addresses. I4901 reports which type each Servo IC is. The following table shows which
Servo ICs can be present, their addresses, and the I-variables used to configure them:
Servo IC #
Base Address
I Variables
Servo IC #
Base Address
I Variables
0
1
2
3
4
$078000
$078100
$078200
$078300
$079200
I7000 – I7049
I7100 – I7149
I7200 – I7249
I7300 – I7349
I7400 – I7449
5
6
7
8
9
$079300
$07A200
$07A300
$07B200
$07B300
I7500 – I7549
I7600 – I7649
I7700 – I7749
I7800 – I7849
I7900 – I7949
Servo ICs 0 and 1 are on board a board-level Turbo PMAC or in a 3U-format stack. Servo ICs 2 through
9 are accessed through the expansion port, either the UMAC’s UBUS backplane expansion port, or a
board-level Turbo PMAC’s JEXP cable expansion port, on an Acc-24x axis board, or an Acc-51x
interpolator board.
PMAC-Style DSPGATE ASIC: The PMAC-style Servo IC is labeled the DSPGATE. It is a 4-channel
part with 32 memory-mapped registers. Each channel supports the following features:
• Serial output for 16-bit digital-to-analog converter
• Input for digital quadrature or pulse-and-direction feedback with index
• Four input flags (home, +/-limit, amp-fault) that can trigger hardware encoder capture
• Amplifier-enable output
• Hardware position-compare output
• Input from 16-bit analog-to-digital converter (from accessory board)
8
Turbo PMAC Family Overview
Turbo PMAC User Manual
HOME 1+LIMIT 14-LIMIT 1FAULT
14
4
PMAC
CUSTOM
AENA 1-4
EQU 1-4
SELECTABLE-FREQUENCY CLOCK INPUTS
ENCODER
SAMPLE
LD
DAC 1
SERVO
PHASE
LD
DAC 2
INPUT
FLAGS
4
ENCODER
INPUTS
ENCODER 1
A
B
C
ENCODER 2
A
B
C
DAC 3
DAC SHIFT REGISTERS (4)
ENCODER 3
LD
SERIAL
DATA OUT
DAC 4
OPTO
ISOLATION
DSP-GATE
SERIAL
DATA IN
1
ADC
16 BIT
MUX
2
3
4
ANALOG
INPUTS
4
ENCODER 4
ANALOG
CONTROL
CLOCK
MUX
CONTROL
ENCODER
CONTROL
24-BIT
DATA BUS
16 BIT
RESOLUTION
OUTPUT
FLAGS
ADC SHIFT REGISTERS (4)
A
B
C
LD
FLAG
CONTROL
A
B
C
4
ANALOG
OUTPUTS.
ACCESSORY BOARD
16-BIT
ADDRESS BUS
The PMAC-style DSPGATE IC is provided on the following Turbo PMAC products:
• Turbo PMAC PC
• Turbo PMAC VME
• Turbo PMAC PCI
• Turbo PMAC PCI Lite
• Acc-24P ISA-Format Axis-Expansion Board
• Acc-51P ISA-Format Interpolator Board
PMAC2-Style DSPGATE1 ASIC: The PMAC2-style Servo IC is labeled the DSPGATE1. It is a 4channel part with 64 memory-mapped registers. Each channel supports the following features:
• Three output command signal sets, configurable as either:
•
Two serial data streams to digital-to-analog converters of up to 18 bits, and one pulse-anddirection pair, or
•
Three pulse-width-modulated (PWM) top and bottom pairs
• Input for digital quadrature with index, pulse-and-direction, or MLDT feedback
• Four input flags (home, +/-limit, user) that can trigger hardware encoder capture
• Amplifier-fault input
• Four supplemental input flags (T, U, V, W) for hall commutation sensors, sub-count data, fault codes,
or general use
• Amplifier-enable output
•
Hardware position-compare output
•
Input from two analog-to-digital converters of up to 18 bits (from amplifier or accessory board)
Turbo PMAC Family Overview
9
Turbo PMAC User Manual
The DSPGATE1 IC also has on-board software-configurable clock generation circuitry. It can generate
the servo and phase clocks for the entire Turbo PMAC system (only one IC will do this; the others will
accept these as inputs). It also generates the clock signals that drive its own circuitry: DAC, ADC, PWM
and PFM (pulse-frequency-modulation).
PMAC2 Gate Array IC
“DSPGATE1”
3
Encoder 1
Output1 A,B,C
6
9
2
Flag 1
ADCIN1 A,B
2
3
Encoder 2
Output2 A,B,C
6
9
2
Flag 2
ADCIN2 A,B
2
3
Encoder 3
Output3 A,B,C
6
9
2
Flag 3
ADCIN3 A,B
2
3
Encoder 4
Output4 A,B,C
6
9
2
Flag 4
ADCIN4 A,B
2
Encoder/Flag
Status & Control
Output/ADCIN
Control
2
2
Clock, PLL
24-bit
DATA BUS
6-bit
ADDRESS BUS
The PMAC2-style DSPGATE1 IC is provided on the following Turbo PMAC products:
• Turbo PMAC2 PC
• Turbo PMAC2 VME
• Turbo PMAC2 PCI
• Turbo PMAC2 PCI Lite
• Acc-24P2 ISA-format Axis Expansion Board
• Acc-24E2 UMAC PWM Axis Board
• Acc-24E2A UMAC Analog Axis Board
• Acc-24E2S UMAC Encoder/Stepper Board
• Acc-51E UMAC Interpolator Board
• Acc-24C2 UMAC-CPCI PWM Axis Board
• Acc-24C2A UMAC CPCI Analog Axis Board
• Acc-51C UMAC CPCI Interpolator Board
• Acc-2E 3U-Format Stack Axis Board
MACRO ICs
The Turbo PMAC has two types of ASICs that support communication over the MACRO ring, the openarchitecture high-speed real-time communications ring that Delta Tau developed to simplify connection
of distributed control hardware.
10
Turbo PMAC Family Overview
Turbo PMAC User Manual
Turbo PMAC variable I4902 reports how many MACRO ICs are present, and at which addresses. I4903
reports which type each MACRO IC is, a DSPGATE2 or a MACROGATE (see below). Variables I20 –
I23 specify the addresses of the four MACRO ICs that are automatically configured with I-variables. The
following table summarizes this configuration:
MACRO IC #
I Variables
Usual Base Address
Usual Type
0
1
2
3
I6800 – I6849
I6850 – I6899
I6900 – I6949
I6950 – I6999
$078400
$079400
$07A400
$07B400
DSPGATE2
MACROGATE
MACROGATE
MACROGATE
DSPGATE2 MACRO IC
The DSPGATE2 IC provides both MACRO-ring functions and general-purpose I/O functions. The I/O
functions are described in the next section. For the MACRO ring, the DSPGATE2 provides a 16-node bidirectional interface. Of these nodes, eight can be used as servo nodes, each of which can transfer all of
the command and feedback data required for the servo and commutation of a motor. Six of the nodes can
be used for general-purpose I/O, each node supporting 72 bits of hard real-time I/O in each direction.
Two of the nodes are for non-real-time communications, including broadcast mode in which a master
controller can talk to all of its slave devices simultaneously.
Turbo PMAC2 Ultralite controllers come standard with a single DSPGATE2 IC, for a 16-node MACRO
interface. The UMAC Acc-5E also comes standard with a single DSPGATE2 IC.
MACROGATE MACRO IC
The MACROGATE IC provides the same 16-node MACRO-ring interface as does the DSPGATE2 IC,
but it does not have the I/O capability of a DSPGATE2. Subsequent optional MACRO ICs (Options 1U1,
1U2, and 1U3) on a Turbo PMAC2 Ultralite are MACROGATE ICs.
I/O ICs
The Turbo PMAC CPU accesses general-purpose digital I/O through a variety of I/O interface ICs. On
Turbo PMAC boards, these are modern equivalents of the classic 8255 IC design originally made by
Intel. These ICs are addressed by a Turbo PMAC CPU at Y:$078800 – Y:$078803.
On Turbo PMAC2 boards, and UMAC I/O boards, these are Delta Tau-designed I/O ASICs, as described
below.
DSPGATE2 I/O IC
The DSPGATE2 IC provides both MACRO-ring functions and general-purpose I/O functions. The
MACRO functions are described in the previous section. Each board-level Turbo PMAC2 board and the
UMAC’s Acc-5E board use the DSPGATE2 to support the non-servo I/O ports. These ports are:
• The JTHW multiplexer port
• The JIO general-purpose I/O port
• The JDISP display port
• The JHW handwheel port
The DSPGATE2 IC on a board-level Turbo PMAC2 is addressed at Y:$078400 – Y:$07843F. On a
UMAC Acc-5E board with the address DIP-switches in the default configuration, the DSPGATE2 IC is
addressed at this same location.
Turbo PMAC Family Overview
11
Turbo PMAC User Manual
IOGATE I/O IC
The IOGATE IC is used to access general-purpose digital I/O on most of the UMAC I/O boards. It
provides 48 I/O points, addressed as 6 bytes in consecutive registers. Different boards use different
buffers and drivers around the IOGATE to provide the specific I/O features desired. While on the
IOGATE itself, each I/O point is individually selectable as to direction, on most of the I/O boards, each
point’s direction is fixed by the external circuitry for that point. The IOGATE must be set up at poweron/reset to support the particular direction configuration of the board it is used on.
Potentially, up to 16 boards with the IOGATE or similar ICs can be installed in a UMAC system. The
following table shows the possible addresses of these boards:
DIP
Switch
Base
Address
DIP
Switch
Base
Address
DIP
Switch
Base
Address
DIP
Switch
Base
Address
0
1
2
3
$078C00
$078D00
$078E00
$078F00
4
5
6
7
$079C00
$079D00
$079E00
$079F00
8
9
10
11
$07AC00
$07AD00
$07AE00
$07AF00
12
13
14
15
$07BC00
$07BD00
$07BE00
$07BF00
Communications Ports
Turbo PMAC controllers provide multiple communications ports. These ports can be used
simultaneously, permitting the use of a second port for debugging while the primary port is used in the
actual application, or of a second port for an operator pendant.
Bus Ports
Each configuration of the Turbo PMAC supports a bus port for fast parallel communications. These
utilize the host port on the DSP, and optionally dual-ported RAM for communications through shared
memory. The buses can be either the traditional backplane buses, or the newer wire buses.
The backplane buses currently supported are:
• ISA (or the PC/104 stack version)
• PCI
• VME
The wire buses currently supported are:
• USB (1.1 or 2.0)
• Ethernet (TCP/IP or UDP/IP protocols) at 10 Mbps
Serial Ports
Turbo PMACs also support serial-port communications, always with a single serial port, optionally with a
second serial port as well.
Main Serial Port
Turbo PMAC controllers always come with a standard serial communications port. In most
configurations, this port can be used with either RS-232 signal format, or RS-422 signal format, jumper
selectable. Even if there are separate connectors for the two signal formats, this is a single port; only one
of the connectors can be used at a time.
Optional Auxiliary Serial Port
If the Option 9T is ordered, a second serial port, always at RS-232 levels, is provided. This port can be
used for general communications. It is possible to turn off the command parser for this port by setting
variable I43 to 1, so Turbo PMAC does not try to interpret incoming characters as commands, making it
possible for the user to parse incoming data as he pleases. This port is required for setup and diagnostics
if the IEC-1131 PMAC Ladder programs are used.
12
Turbo PMAC Family Overview
Turbo PMAC User Manual
Key Software Components
In any application, the Turbo PMAC will have software components provided both by Delta Tau and by
the user. The Delta Tau software components are known as the firmware (software embedded in the
hardware). These are not application-specific. The user software components – primarily motion
programs and PLC programs – are application-specific.
Operational Firmware
The operational firmware of the Turbo PMAC serves as both the operating system for the controller and
as the PMAC-language interpreter. As an operating system, it provides a hard-real-time multi-tasking
environment with extremely fast task switching, providing a very high level of determinism for the highpriority tasks such as servo-loop closure. It also handles the sequencing of operations within motion
programs automatically, so this sequencing of move-by-move operations does not have to be handled by
the user.
Typically, the operational firmware is updated once or twice a year to add new features and to fix
problems. Field upgrades of the firmware are easy to perform. Simply power up the Turbo PMAC with
the firmware reload jumper installed, and then launch the PMAC Executive program on the PC. The
Executive program will notice automatically that the Turbo PMAC is ready to install new firmware and
query for the file to be downloaded. It will download the new firmware automatically from the specified
file into the Turbo PMAC.
Bootstrap Firmware
The operation of the bootstrap firmware is invisible to virtually all Turbo PMAC users. Its function is
very similar to that of a PC’s BIOS, providing a link between the hardware and the operating system, and
permitting the operating system to load. While Delta Tau may upgrade the bootstrap firmware from time
to time (and each new hardware configuration of the CPU may require a different variation of the
bootstrap firmware), Delta Tau does not support the field upgrade of the bootstrap firmware.
In normal operation, the bootstrap firmware operates invisibly underneath the operational firmware,
automatically loading the operational firmware on top of it at each power-up/reset. However, if the Turbo
PMAC is powered up or reset with the firmware-reload jumper on, the operational firmware will not be
loaded from flash memory, and the host computer can communicate directly with the bootstrap firmware
(through the main serial port or a bus that uses the CPU’s host port only). If the operational firmware has
not been loaded, Turbo PMAC’s response to the ?, ??, or ??? status query command is BOOTSTRAP
PROM.
Motors, Axes, and Coordinate Systems
Motors, axes, and coordinate systems are software structures in a Turbo PMAC system. While they
usually utilize hardware channels to interface to the outside world, it is possible to use them as virtual
structures, just for their computational features.
Motors
A motor in Turbo PMAC consists of the software structures necessary for basic moves, commandedposition interpolation, servo-loop closure, commutation, and current-loop closure. Every Turbo PMAC
has the capability for executing the algorithms for 32 motors, even if only a few servo channels are
provided. These motors are specified by number (1 through 32), and their attributes are specified in terms
of raw “counts”, milliseconds, and related units.
A Turbo PMAC motor accesses hardware-channel functions for its various input and output needs
through setup I-variables that contain the address of the register to be used. This mapping functionality
provides great power and flexibility in a simple manner.
Turbo PMAC Family Overview
13
Turbo PMAC User Manual
Axes
Generally, motions in a Turbo PMAC system are commanded through the use of axes. An axis in Turbo
PMAC consists of the software structures for programmed moves. Axes are specified by letter (A, B, C,
U, V, W, X, Y, and Z), and their attributes are specified in terms of user-specified units (e.g. millimeters,
inches, degrees, seconds, minutes).
Axes are assigned to motors through axis definition statements or kinematic subroutines. While there is
usually a one-to-one mapping between axes and motors (e.g. Motor 1 assigned to the X-axis and Motor 2
assigned to the Y-axis), this does not have to be the case. Multiple motors can be assigned to a single
axis, as in a gantry configuration; there can be no motors assigned to an axis, creating a virtual axis; or
there can be a complex relationship of multiple motors to multiple axes in a kinematic algorithm.
An axis belongs to a coordinate system (see below). Up to 9 axes may be used in a single coordinate
system, one of each letter name.
Coordinate Systems
The coordinate system is Turbo PMAC’s structure for achieving tightly coordinated motion of multiple
motors. Turbo PMAC supports up to 16 separate coordinate systems. A motor is assigned to an axis in a
particular coordinate system. Multiple axes in the same coordinate system that are commanded on the
same line of a motion program are automatically coordinated.
To make the motion of more than one motor coordinated, assign them to axes in the same coordinate
system. To have independent motion of motors (i.e. starting stopping, and changing speeds at arbitrary
times with respect to each other, assign the motors to axes in separate coordinate systems.
User Programs
Turbo PMAC users can install several types of programs into the controller, each type with a specific
purpose.
Motion Programs
Turbo PMAC motion programs provide an easy way to specify sequences of coordinated motion and the
execution of any calculations that are synchronous with the programmed motions. The motion program
language combines features of the RS-274 standard G-code machine-tool programming language, which
is good for specifying sequences of coordinated motion, and BASIC, which is good for the associated
math and logic.
PLC Programs
PLC programs in Turbo PMAC are intended for actions and calculations that are asynchronous to the
programmed motion. PLC programs repeatedly scan in the fashion of regular programmable logic
controllers. They can be used for I/O control as a dedicated PLC would be, but because they have direct
access to all registers in Turbo PMAC, they have many other uses as well.
Turbo PMAC programs can either be written in a BASIC-like text language, or in IEC-1131 ladder logic
or sequential function charts (extended memory option and special PC software required). Text PLC
programs can be run either as interpreted code, or as compiled code for greater efficiency. See the User
Manual section on PLC programs for more details.
User-Written Servo and Phase Algorithms
Virtually all users will be able to utilize Turbo PMAC’s built-in servo-loop-closure and phasecommutation algorithms. However, it is possible for the user to install custom algorithms for either or
both of these to accomplish tasks not possible with the standard algorithms. Some people will use these
algorithms simply for high-speed, high-priority I/O or calculations by activating them on an otherwise
unused motor.
User-written servo algorithms can be written either in the high-level PMAC language (for ease of use) or
in DSP56300 assembly language (for maximum efficiency). User-written phase algorithms must be
written in assembly language. See the User Manual sections on commutation and servo loops for details.
14
Turbo PMAC Family Overview
Turbo PMAC User Manual
Turbo PMAC Configurations
Turbo PMAC controllers can come in a variety of different physical configurations. Fundamentally, these
break into three different types: board-level, rack-mounted, and boxed. Each of these types is described
below.
Board-Level Designs
There are multiple board-level implementations of the Turbo PMAC. These designs can be used as
expansion cards in a backplane bus, but they do not have to be installed in the bus, either for initial setup,
or in the actual application. All have serial communications links as well.
Presently, there are Turbo PMAC designs for the ISA, VME and PCI buses. These come with PMACstyle or PMAC2-style interfaces. Lite versions of the Turbo PMAC have only one on-board Servo IC,
providing four channels of servo interface circuitry. Ultralite versions of the Turbo PMAC2 replace onboard servo interface circuitry with MACRO-ring interfaces; the actual servo interface circuitry is located
in a remote MACRO Station or embedded in a MACRO drive.
The following table summarizes the board-level Turbo PMAC implementations available as of this
writing (mid-2003):
Product
Bus Type
Max # On-Board
Servo ICs (Channels)
On-Board
MACRO?
Turbo PMAC PC
Turbo PMAC VME
Turbo PMAC PCI
Turbo PMAC PCI Lite
Turbo PMAC2 PC
Turbo PMAC2 PC Ultralite
Turbo PMAC2 VME
Turbo PMAC2 VME Ultralite
Turbo PMAC2 PCI
Turbo PMAC2 PCI Lite
Turbo PMAC2 PCI Ultralite
ISA
VME
PCI
PCI
ISA
ISA
VME
VME
PCI
PCI
PCI
2 (8)
2 (8)
2 (8)
1 (4)
2 (8)
0 (0)
2 (8)
0 (0)
2 (8)
1 (4)
0 (0)
No
No
No
No
No
Yes
No
Yes
No
No
Yes
UMAC Rack-Mounted Designs
Rack-mounted implementations of the Turbo PMAC are called UMAC (Universal Motion and
Automation Control). These implementations consist of a modular set of 3U-format (100mm x 160mm)
Euro-cards installed in a common backplane and mounted in a Euro-rack. The modular style permits the
easy customization of the controller to the User precise needs. The rack mounting provides easy
installation and connection in industrial machines, without the difficulties of wiring into computermounted controllers. High-speed wire communications links such as USB and Ethernet provide
communications speeds close to those of backplane buses.
UMAC Turbo
The UMAC Turbo is a rack-mounted Turbo PMAC system in which the 3U-format boards are connected
across a backplane board that has the physical format of a VME bus (96-pin DIN connectors), but is not
electrically or software compatible with the VME bus. The field wiring comes out of the top, bottom, and
(sometimes) front sides of the cards.
There are a wide variety of UMAC boards. This includes:
• Turbo CPU board with PC/104 interface
• Axis boards: PWM, analog, stepper/encoder
• Sensor boards: Sinusoidal Encoder, SSI, Absolute Encoder, resolver
• Digital I/O boards: TTL and Isolated 24V, sinking and sourcing
• A/D converter boards: 12-bit and 16-bit
• High-speed communications board: USB or Ethernet
Turbo PMAC Family Overview
15
Turbo PMAC User Manual
•
•
•
Fieldbus interface: DeviceNet and Profibus, master or slave
Backplanes for 4 to 18 boards plus power supply
AC-input (85 – 240VAC) and DC-input (24VDC) power supplies
Compact UMAC Turbo
The Compact UMAC Turbo (formerly called UMAC-CPCI Turbo) is a rack-mounted Turbo PMAC
system in which the 3U-format boards are connected across a backplane board that has the physical
format of a Compact PCI (CPCI) bus (110-pin high-density connectors), but is not electrically or software
compatible with the CPCI bus. The field wring comes out of pass-through connectors behind the
backplane. It is the User responsibility to devise a distribution scheme for the field wiring. For this
reason, the Compact UMAC is intended for high-volume users who can afford the investment in an
application-specific distribution scheme.
There is a small family of Compact UMAC boards. This family presently includes:
• Turbo CPU Board with optional USB or Ethernet interface
• Acc-24C2 Digital (PWM/Stepper) 4-Axis Interface Board
• Acc-24C2A Analog 4-Axis Interface Board
• Acc-11C 48/96 Isolated Digital I/O Board
• Acc-51C Sinusoidal Encoder Interpolator Board
• Acc-C8 8-slot backplane board (plus power-supply slot)
QMAC Boxed Design
The QMAC is a boxed 4-channel Turbo PMAC2 with built-in power supply and front-panel connector
board. It is available with pulse-and-direction, analog, or direct-PWM outputs to amplifiers.
QMAC has the following standard features:
• 4 channels basic axis-interface circuitry, each including:
•
Quadrature encoder input with index and hall commutation sensor
•
Differential pulse-and-direction stepper outputs
•
Four isolated input flags (home, /-limit, user)
•
Isolated position-compare output flag
•
Encoder/stepper interface on DB-15 connector
•
Flag interface on removable terminal block
• Two supplemental channels on DB-25 connector, each including:
•
Quadrature encoder input
•
Pulse output, configurable as PWM or PFM
• General-purpose isolated digital I/O: 8 out, 16 in, on removable terminal blocks
• RS-232 communications port
• Multiplexer-port interface for I/O accessories
• Display port interface
QMAC has the following optional features:
• Four channels single or dual analog amplifier interface
• Four channels direct-PWM amplifier interface
• USB communications interface
• Ethernet communications interface
16
Turbo PMAC Family Overview
Turbo PMAC User Manual
TURBO PMAC/PMAC2 SYSTEM CONFIGURATION AND
AUTO-CONFIGURATION
Turbo PMAC, and especially Turbo PMAC2, boards have extensive capabilities for automatically
identifying and self-configuring their systems. This is particularly important for UMAC Turbo systems,
with their wide variety of configurations. These capabilities provide ease of use and flexibility in getting
started with a particular configuration.
CPU Clock Frequency
In any Turbo PMAC system, the clock frequency at which the CPU actually operates is set by software
variable I52, expressed as a multiplication factor from the fixed clock-crystal frequency for the system.
The clock-crystal frequency on a Turbo PMAC is 19.66 MHz (sometimes called 20 MHz as a close
approximation). The CPU divides this in half, to 9.83 MHz, and then multiplies the resulting frequency
by an integer factor in a circuit called a phase-locked loop (PLL) to obtain its operating frequency. The
precise equation for the CPU operating frequency is:
CPUfreq( MHz ) = 9.83 * ( I 52 + 1 )
In approximate terms, the equation is:
CPUfreq( MHz ) = 10 * ( I 52 + 1 )
Note that the frequency at which the CPU is operating is not necessarily the same as the maximum rated
frequency for the CPU. Depending on the CPU option ordered, the maximum rated frequency will be
different. The following table shows the maximum rated frequencies for each CPU option and the
maximum I52 values that can be used without exceeding these frequencies.
CPU Option
CPU Max. Rated Freq.
Max. Rated I52 Value
Operating Freq. at
Max. Rated I52
5Cx
5Dx
5Ex
5Fx
80 MHz
100 MHz
160 MHz
240 MHz
7
9
15
23
78.64 MHz
98.30 MHz
157.28 MHz
235.92 MHz
The factory default value for I52 is 7, producing an 80 MHz CPU frequency, regardless of the CPU
option present. I52 is used only at power-up/reset time, so to change the CPU frequency, change the
active value of I52, store this new value to non-volatile flash memory with a SAVE command, and reset
the card (usually with a $$$ command). During the power-up/reset cycle, Turbo PMAC will read the
saved value of I52, set the PLL circuitry to generate the proper operating frequency, and make other
settings (e.g. baud rate dividers, memory and I/O wait states) appropriate for that frequency.
Usually, I52 is set for the CPU to run at the maximum rated frequency. However, there are some cases in
which a lower frequency may be desired. If serial communications using either the main or auxiliary
serial port at 115,200 baud is desired, the CPU frequency must be an exact multiple of 30 MHz (29.49
MHz) to generate the baud rate accurately enough. So the Option 5Dx 100 MHz CPU might be operated
at 90 MHz to do this.
Note:
It may be possible to operate a CPU at a frequency higher than its rated frequency,
particularly at moderate ambient temperatures. However, safe operation cannot be
guaranteed under these conditions, and any such operation is done entirely at the
user’s own risk.
Turbo PMAC System Configuration and Auto Configuration
17
Turbo PMAC User Manual
System clock frequencies such as the phase and servo clocks, plus the clock frequencies driving hardware
circuits, are generated directly from the clock-crystal frequency through a Servo IC or a MACRO IC and
are not dependent on the CPU operating frequency. These clock frequencies are covered in the following
sections.
Turbo PMAC2 System Clock Source
In a Turbo PMAC2 system, the system phase and servo clocks, which interrupt the processor and latch
key input and output data for the servos, come from one (and only one) of the Servo ICs or MACRO ICs
in the system, or possibly from an external source. There must be a unique source of the phase and servo
clocks for an entire Turbo PMAC2 system. This section explains how to specify that clock source. A
later portion of this section, Setting System Clock Frequencies, explains how to set the frequencies once
the source has been determined.
Default Clock Source
The factory-default source for these clock signals is appropriate in almost all applications. Only in
specialized cases will another source be used. The Turbo Setup program will walk you through the setup
of the frequencies for the source selected. Setting the frequencies is discussed in the next section.
Note:
A Turbo PMAC board uses fixed, discrete logic to generate its phase and servo
clocks. If accessory boards with Servo ICs or MACRO ICs that can generate their
own clock signals are added to a Turbo PMAC, they must be set up to use the
Turbo PMAC’s phase and servo clock signals. Turbo PMAC systems also do
support external clock sources in the same way that Turbo PMAC2 systems do.
IC Clock Generation Facilities
Each PMAC2-style Servo IC (DSPGATE1 IC) and DSPGATE2-type MACRO IC has the capability for
generating its own phase and servo clock signals, or for accepting external phase and servo clock signals.
(Exception: MACROGATE-type MACRO ICs can generate their own phase clock, but not servo clock.
Therefore, they cannot be used to generate clocks for the entire system.) At most one of these ICs in a
system may generate its own clock signals – none if the signals come from an external source.
Variables I7m07 and I7m57 control the clock direction for Servo ICs ‘m’ and ‘m*’, respectively.
Variables I6807, I6857, I6907, and I6957 control the clock direction for MACRO ICs 0, 1, 2, and 3,
respectively. If the variable value is 0, the IC generates its own clock signals and outputs them. If the
variable value is 3, the IC accepts the clock signals from a source external to it. At most one of these ICs
can have this variable at a value of 0; the rest must be set to 3.
Note:
If more than one of these ICs is set up to use its own clock signals and to output
them, the processor will be interrupted by multiple sources and will not operate
normally – it is possible that the watchdog timer will trip. (Because the outputs are
open-collector types, there will be no hardware damage from signal contention, but
system software operation will be compromised.)
18
Turbo PMAC System Configuration and Auto Configuration
Turbo PMAC User Manual
Typical Clock-Source ICs
On a board-level Turbo PMAC2 controller that is not an Ultralite MACRO-only controller, usually the
system clock source is Servo IC 0, the first on-board Servo IC. This means that I7007 is set to 0, so that
I7000, I7001, and I7002 control the system clock frequencies. Other clock-direction I-variables should be
set to 3. The same is true for 3U Turbo Stack controllers.
On a Turbo PMAC2 Ultralite controller, or a UMAC Turbo system utilizing the MACRO ring, usually
the system clock source is MACRO IC 0, the first MACRO IC. This means that I6807 is set to 0, so that
I6800, I6801, and I6802 control the system clock frequencies. Other clock-direction I-variables should be
set to 3.
On a UMAC Turbo system that does not utilize the MACRO ring (which is most UMAC Turbo systems),
usually the system clock source is Servo IC 2, the first Servo IC on the UBUS backplane. This means
that I7207 is set to 0, so that I7200, I7201, and I7202 control the system clock frequencies. Other clockdirection I-variables should be set to 3.
External Clock Sources
It is possible to set up a Turbo PMAC or Turbo PMAC2 system to use externally generated phase and
servo clock signals brought in through extra pins on the main serial port. Sometimes these clock signals
will come from another Turbo PMAC that is part of the same system in order to keep all of the controllers
fully synchronized. Other times they may come from a video system in order to fully synchronize the
motion to the video frames. Still other times they may come from a reference clock much more precise
than Turbo PMAC’s own clock crystal; this is common in the tracking control of telescopic systems.
On a Turbo PMAC system, if any of the jumpers E40 – E43 are removed, it will expect externally
generated phase and servo clocks. On a Turbo PMAC2 system, this function is determined by jumper E1
on board-level Turbo PMACs and jumpers E1A and E1B on the CPU board of a UMAC system.
It is recommended that the servo and phase clock signals provided to a Turbo PMAC each are differential
5V pairs (RS-422 levels). However, single-ended 5V signals can be accepted into the “+” inputs; the “-”
inputs should be left unconnected and will be held at 2.5V internally. The phase clock signal should have
a 50% duty cycle. The servo clock must be synchronous with the phase clock, with a period of n phase
clock periods, where n is a positive integer. The falling edge of the SERVO+ signal must be coincident
with the falling edge of the PHASE+ signal, and the SERVO+ signal should be low for at most 1 phaseclock cycle.
For more details on the use of external clock signals, refer to the Synchronizing Turbo PMAC to External
Events section of this manual.
Distribution of Clock Signals
Whatever the source of the phase and servo clock signals, these signals must be available to the processor
and all Servo ICs and MACRO ICs, plus any other circuits that use these signals in their functioning
(such as I/O cards that are used for parallel or serial feedback). Note that the hardware clock signals – the
DAC clock, ADC clock, encoder sample clock, and PFM clock – are generated locally inside each Servo
IC and MACRO IC, and are not shared between ICs.
Board-Level Turbo PMACs
On a board-level Turbo PMAC, the clock source must be one of the on-board Servo ICs or MACRO ICs,
and the clock signals are just transmitted over circuit traces on the board, and if there is a piggyback CPU
board, up to the CPU board. The servo and phase clock signals are always present as outputs (never
inputs) on the JEXP expansion port for accessory boards connected to this port. These signals are also
available as buffered differential outputs on extra pins on the RS-422 serial port connector for other
devices.
Turbo PMAC System Configuration and Auto Configuration
19
Turbo PMAC User Manual
UMAC Systems
In UMAC systems (including Compact UMAC), the phase and servo clocks are shared across the UBUS
backplane board among the different 3U-format cards inserted into that backplane. Each card has buffer
ICs for these signals as they interface to the backplane. On cards that are potential sources of the phase
and servo clock signals, such as the Acc-24E/C axis boards or the Acc-5E MACRO board, these buffers
can be configured as either inputs or outputs.
On each UMAC board that could be a clock source, there is a jumper that controls the configuration of the
clock-direction buffers. In one setting, the board can only input the clock signals. This setting is required
for the UMAC MACRO, in which the clock signals always come from the MACRO interface board. It is
permissible, but not recommended, for boards in UMAC Turbo systems that will be not be generating
their own phase and servo clock signals. This setting is not permissible for the UMAC board that is
generating the system phase and servo clocks.
In the other setting, the direction of the clock-signal buffers can be reversed by the Turbo CPU. This
setting is required for the board that is generating the system clocks; it is recommended for the other
boards as well (so the source can be changed without moving any buffers). At power-up/reset, the CPU
will configure the buffers the board containing the Servo IC or MACRO IC that is specified by I19 to
generate the system clocks as outputs to the UBUS backplane; it will configure the buffers on all other
boards to be inputs from the UBUS backplane.
Missing Clock Signals
If the phase and/or servo clock signals are not present, this is a serious failure, and the Turbo PMAC
system will not be able to operate properly. This section explains how Turbo PMAC systems respond to
such a failure.
Board-Level Turbo PMACs
On board-level Turbo PMACs, if the CPU does not receive the phase and servo clock signals at any time,
the watchdog timer will immediately trip and shut down the system completely. If this happens
immediately at power-up/reset due to improper configuration of the clock-source setup, install the reinitialization jumper and power up the system again to restore a valid clock source from one of the onboard Servo ICs or MACRO ICs.
UMAC Systems
In a UMAC system (including Compact UMAC), if the CPU does not receive any phase or servo clock
signals over the UBUS backplane immediately after configuring the clock source, it will go into a special
mode in which it generates its own phase and servo clocks at the factory default frequencies of 9.03 kHz
and 2.26 kHz, respectively. In this case, it sets the global status bit No hardware clocks found at
X:$000006 bit 3 to 1.
This mode is intended to keep the system alive to permit the user to set up the clock configuration
information properly for their setup. It is not intended for the UMAC to be able to do actual control in
this mode. In firmware revisions V1.940 and newer, no motor can be enabled if this error bit is set, with
the Turbo PMAC reporting an ERR018 to an enabling command if I6 is set to 1 or 3.
If the UMAC CPU stops receiving the phase or servo clock signals after this time, the watchdog timer
will trip and immediately shut down the system completely.
Re-Initialization Actions
On re-initialization of a Turbo PMAC2 ($$$*** command, or power-on/reset with re-initialization
jumper E3 ON), the CPU searches all possible locations of Servo ICs and MACRO ICs to see which are
present. If the system is not set up for an external clock source, it then makes a decision as to which of
these ICs it will use to generate the system’s phase and servo clocks, using the first IC that it finds in the
following list:
20
Turbo PMAC System Configuration and Auto Configuration
Turbo PMAC User Manual
1.
2.
3.
4.
…
11.
12.
…
19.
20.
21.
22.
Servo IC 0
MACRO IC 0
Servo IC 1
Servo IC 2
(On-board or 3U Stack)
(On-board or Acc-5E)
(On-board or 3U Stack)
(Acc-24E2, 51E)
(I19=7007)
(I19=6807)
(I19=7107)
(I19=7207)
Servo IC 9
Servo IC 2*
(Acc-24E2, 51E)
(Acc-24E2, 51E)
(I19=7907)
(I19=7257)
Servo IC 9*
MACRO IC 1
MACRO IC 2
MACRO IC 3
(Acc-24E2, 51E)
(On-board or Acc-5E)
(On-board or Acc-5E)
(On-board or Acc-5E)
(I19=7957)
(I19=6857)
(I19=6907)
(I19=6957)
(MACRO ICs must be DSPGATE2 ICs to be used as a clock source.)
Next, it sets I19 to the number of the clock-direction I-variable for the first of these ICs it finds
automatically. For example, if it finds Servo IC 0, it will set I19 to 7007. Finally, it will set this clockdirection I-variable to 0, and the clock-direction I-variable for all of the other Servo and MACRO ICs that it
finds to 3.
However, if it finds on re-initialization that the E1 jumpers are configured to use external phase and servo
clocks, I19 is set to 0 and all of the clock-direction I-variables are set to 3.
Note:
Once the system has been set up to take external phase and servo clock signals,
these signals must always be present while the system is powered, or the watchdog
timer will trip immediately.
To change a system set up for external clocks back to internal clocks, it is necessary to power it up with
the E3 re-initialization jumper ON, the E1 external-clock jumper OFF, and with the external clock signals
present. Once the new configuration for internal-clock source is established, either by the automatic reinitialization actions, or user-set configuration (see below), these settings must be stored to flash memory
with the SAVE command. Then the system can be powered down, the E3 re-initialization jumper
removed, and the external clock signals removed. Finally, the system can be powered up again normally.
Note:
If a Turbo PMAC2 is on a MACRO ring, but it is not the ring controller
synchronizing master, it must be set up to have its phase clock adjusted by receipt
of the sync packet over the MACRO ring. This is done by setting bit 7 of I6840 to
1, and bits 16 – 19 of I6841 to the node number of the sync packet (usually $F
[=15]). In this case, the phase and servo clocks are still generated internally,
although they are locked to receipt of this sync packet. Systems of this type should
have I6807 set to 0 (I19 set to 6807) to use MACRO IC 0 as the source of the
phase and servo clocks. If re-initialization does not automatically cause this to
happen, it must be done manually (see below).
User-Customized Clock-Source Specification
The procedure to change the clock source is slightly different for 3U-format Turbo PMAC2 systems and
other Turbo PMAC2 systems.
In a 3U-format Turbo PMAC2 system (UMAC Turbo or 3U Turbo Stack), to change the clock source,
simply follow this 3-step procedure:
1. Set I19 to the number of the clock-source I-variable of the IC you want to be the source.
2. Store this value to non-volatile flash memory with the SAVE command.
3. Reset the system normally (not re-initialization).
Turbo PMAC System Configuration and Auto Configuration
21
Turbo PMAC User Manual
Do not try to set the clock-direction I-variables directly.
In other Turbo PMAC2 systems, change the clock-direction I-variables themselves in a single command
(e.g. I6807=0 I7007=3). It is best to change I19 to the number of the I-variable that just set to 0
(I19=6807 in this example), but this is not necessary. Store these new values to non-volatile flash
memory with the SAVE command. They will then be used automatically on every subsequent powerup/reset.
Servo and phase clock lines are bi-directional on the UBUS backplane expansion port in UMAC Turbo
systems, so these signals can go either from or to the CPU board. However, on the JEXP flat-cable
expansion port, these clock lines are uni-directional and can only be output from the main PMAC board
or CPU board.
If I19 is set to an improper value, the watchdog timer will trip immediately on reset. To recover, power
down, install the E3 re-initialization jumper, and power up again.
The most common reason to change from the default setting is tying a Turbo PMAC2 that has Servo IC 0
and/or 1 to a MACRO ring where it is not the ring controller. In this case, MACRO IC 0 should be the
clock source, but the re-initialization procedure will decide on Servo IC 0. In this case, change I19 from
7007 to 6807, SAVE, and reset.
Normal Reset Actions
On a normal power-up or reset sequence, the Turbo PMAC2 CPU reads the value of I19 that was
previously saved to flash memory and sets the I-variable whose number it finds there to 0, specifying that
this IC uses its own phase and servo clocks and outputs them to the system. For example, if I19 were
6807, I6807 would be set to 0. It would then set the clock-direction I-variables automatically for all of
the other Servo and MACRO ICs that it finds at power-up/reset to 3, so these ICs accept servo and phase
clock signals as inputs.
MACRO IC Selection
Starting in Turbo PMAC firmware version 1.936, I-variables I20 – I23 must be set to specify the
addresses of the MACRO IC(s) used for automatic firmware functions. This is not compatible with older
firmware versions. If updating an application from an older version, after loading the old I-variable file,
issue the command I20..23=*, followed by a SAVE command, followed by a $$$ reset command.
Some Turbo PMAC2 systems (presently UMAC Turbo) can address up to 16 MACRO ICs. However,
there is automatic firmware support for only four of these ICs at any given time. These ICs are referred to
as MACRO ICs 0, 1, 2, and 3. Variables I20 through I23 specify the base addresses of MACRO ICs 0
through 3, respectively. These variables must be set properly to use the desired ICs for any automatic
firmware functions.
MACRO IC 0, specified by I20, has several functions that require automatic firmware support:
• Display port functions (can be changed dynamically)
• Multiplexer port functions (can be changed dynamically)
• I-variables I6800 – I6849 (values automatically assigned only at power-up/reset)
• MACRO nodes 0 – 15 (can be changed dynamically)
• MACRO Type 1 auxiliary communications if I84=0
MACRO IC 1, specified by I21, has several functions that require automatic firmware support:
I-variables I6850 – I6899 (values automatically assigned only at power-up/reset)
MACRO nodes 16 – 31 (can be changed dynamically)
MACRO Type 1 auxiliary communications if I84=1
•
•
•
MACRO IC 2, specified by I22, has several functions that require automatic firmware support:
I-variables I6900 – I6949 (values automatically assigned only at power-up/reset)
•
22
Turbo PMAC System Configuration and Auto Configuration
Turbo PMAC User Manual
•
•
MACRO nodes 32 – 47 (can be changed dynamically)
MACRO Type 1 auxiliary communications if I84=2
MACRO IC 3, specified by I23, has several functions that require automatic firmware support:
I-variables I6950 – I6999 (values automatically assigned only at power-up/reset)
MACRO nodes 48 – 63 (can be changed dynamically)
MACRO Type 1 auxiliary communications if I84=3
•
•
•
On re-initialization, Turbo PMAC2 searches for the MACRO ICs with the lowest base addresses. I20 is
assigned the lowest base address (if one is found); I21 is assigned the next (if found), and so on. Also, the
same action is taken when assigning the default value to one of these variables (e.g. I20=*).
Dual-Ported RAM IC Selection
Starting in Turbo PMAC firmware version 1.936, it is possible to specify the base address of the dualported RAM IC used for automatic firmware communications functions. This permits support for new
accessories such as the Acc-54E USB/Ethernet communications board. In firmware versions V1.935 and
older, the base address was fixed at $060000 for the “on-board” DPRAM.
New variable I24 specifies the base address of the DPRAM IC used for the automatic firmware
communications functions. For backward compatibility, if I24 is set to 0, the DPRAM will be assumed to
have a base address of $060000.
I24 is used only at power-up/reset. To use other than the on-board DPRAM IC, follow the instructions of
the accessory such as the Acc-54E to set the value of I24 to match the hardware settings on the accessory.
Then issue the SAVE command and the $$$ command so the accessory can be used for communications.
On re-initialization, I24 is set to the lowest base address of any DPRAM IC found. Most commonly, I24
will be set to $060000 for use with backplane-bus communications, or to $06C000 for use with the
DPRAM on a UMAC Acc-54E USB/Ethernet Communications Board.
System Configuration Status Reporting
Turbo PMAC systems can detect and report significant information about their configuration
automatically. They do this by having the processor query possible address locations for interface ICs –
Servo ICs, MACRO ICs, DPRAM ICs, and I/O ICs. This information can be very useful in the initial
setup of a Turbo PMAC system, and subsequently to verify that the configuration has not changed.
Servo IC Configuration
On power-up/reset, the Turbo PMAC CPU automatically tests for the presence and type of all possible
Servo ICs and reports the results in I4900 and I4901. I4900 is a collection of 20 independent bits, in
which bits 0 – 9 report the presence of Servo ICs 0 – 9, respectively, and bits 10 – 19 report the presence
of Servo ICs 0* to 9*, respectively. A bit value of 1 indicates the IC is present; a bit value of 0 indicates
the IC is absent.
I4901 is also a collection of 20 independent bits, in which bits 0 – 9 report the type of Servo ICs 0 – 9,
respectively, and bits 10 – 19 report the type of Servo ICs 0* to 9*, respectively. A bit value of 1
indicates a PMAC2-style “DSPGATE1” IC; a bit value of 0 indicates a PMAC1-style “DSPGATE” IC
(or no IC present if the corresponding bit of I4900 is 0).
MACRO IC Configuration
On power-up/reset, the Turbo PMAC CPU tests for the presence and type of all possible MACRO ICs
automatically and reports the results in I4902 and I4903. I4902 is a collection of 16 independent bits,
each reporting the presence of a MACRO IC at one of the 16 possible locations. A bit value of 1
indicates the IC is present; a bit value of 0 indicates the IC is absent.
Turbo PMAC System Configuration and Auto Configuration
23
Turbo PMAC User Manual
I4903 is also a collection of 16 independent bits, each reporting the type of MACRO IC at one of the 16
possible locations. A bit value of 1 indicates a DSPGATE2 IC; a bit value of 0 indicates a
MACROGATE IC (or no IC present if the corresponding bit of I4900 is 0).
While it is possible for up to 16 MACRO ICs to be installed in a Turbo PMAC system, only four of these
can be supported at any time by automatic firmware functions. I20 – I23 contain the base addresses of
these four ICs. When the system is re-initialized, these variables are set to values for the four ICs found
with the lowest base addresses.
DPRAM IC Configuration
On power-up/reset, the Turbo PMAC CPU automatically tests for the presence of all possible dual-ported
RAM ICs and reports the results in I4904. I4904 is a collection of eight independent bits, each reporting
the presence of a DPRAM IC at one of the eight possible locations. Only one of these ICs can be
supported at any time by automatic firmware functions. I24 contains the base address of this IC.
CPU Section Configuration
On power-up/reset, the Turbo PMAC automatically tests for the configuration of its own CPU section and
reports the results in I4908. I4908 is a 36-bit value reporting the CPU type, active memory size, DPRAM
size, battery-backed RAM size, flash memory size, presence of auxiliary serial port, part number, and
vendor ID.
UBUS Accessory Board Identification
The Turbo PMAC can report detailed information about accessory boards installed on the UBUS
expansion port in UMAC Turbo systems. This information is reported in variable I4910 – I4965. Each is
a 36-bit variable with the following contents:
Vendor ID: 8 bits
Options present: 10 bits
Revision number: 4 bits
Card ID (part number): 14 bits
Each variable can report one part or all parts of this information, depending on the setting of I39. If I39 is
set to 5, the variable reports the base address of the accessory board instead.
I4910 – I4925 report this information for the 16 possible accessory boards with Servo ICs, such as the
Acc-24E2, 24E2A, 24E2S, and 51E.
I4926 – I4941 report this information for the 16 possible accessory boards with MACRO ICs, such as the
Acc-5E.
I4942 – I4949 report this information for the 8 possible accessory boards with DPRAM ICs, such as the
Acc-54E USB/Ethernet interface.
I4950 – I4965 report this information for the 16 possible accessory boards with I/O ICs, such as the Acc14E, 28E, 36E, 53E, and 59E. (The Acc-9E, 10E, 11E, and 12E I/O boards currently cannot provide this
information.)
Setting System Clock Frequencies
The phase clock and servo clock signals set the heartbeat for the entire Turbo PMAC system,
synchronizing both hardware and software operations. While the factory default frequencies – 9.04 kHz
for the phase clock and 2.26 kHz for the servo clock – are suitable for most applications, some
applications will either require changes, or could benefit from changes in one or both of these frequencies.
The hardware tasks that are driven by the phase and servo clock signals include:
• Latching of encoder counters
• Latching of parallel feedback registers
• Strobing of A/D converters and latching of resulting data
• Output to D/A converters
24
Turbo PMAC System Configuration and Auto Configuration
Turbo PMAC User Manual
•
•
Output to PWM circuits
Communication over MACRO ring
The software tasks that are driven by the phase and servo clock signals include:
• Digital current loop closure (phase clock)
• Motor phase commutation (phase clock)
• Demuxing of muxed A/D converters (phase clock)
• Encoder conversion table pre-processing (servo clock)
• Trajectory interpolation (servo clock)
• Position/velocity loop closure (servo clock and Ixx60)
• PLC 0 and PLCC 0 execution (servo clock and I8)
• Checking for motion program move planning (servo clock and I8)
The sections on commutation and servo-loop closure discuss the selection of optimal frequencies for these
signals. In general, higher frequencies can lead to higher performance (although there are points of
diminishing returns, and other system limitations can cap performance no matter how fast the Turbo
PMAC is running), at the cost of increase processor usage.
Setting Turbo PMAC Phase and Servo Clock Frequencies
On a Turbo PMAC controller, the system clock frequencies are determined by the setting of E-point
jumpers. While some E-point numbers are specific to a particular controller, the clock-frequency jumper
numbers are common to all Turbo PMAC controllers and so can be summarized here. Consult the
Hardware Reference Manual for the particular controller to get the location for these jumpers on the
board.
PMAC(1) Clock Dividing Circuitry
E34A
Crystal
Oscillator
1/2
19.6608 MHz
1/4
1/8
1/16
Encoder Sample Clock
(SCLK)
E34
Analog Channel Select
(ASEL)
E35
1/4
1/17
E36
E33
E37
E98
1
2
3
1/2
1/4
DAC/ADC Clock
(DCLK)
1/8
1/16
Phase Clock
(Phase)
E32
E31
E30
E29
Servo Clock
(Servo)
1/n
E3
E4
E5
E6
n = 1 to 16
E98 DCLK Frequency Control
E98 controls the frequency of the DCLK clock signal that controls the bit rate into all the D/A converters
on the board and out of all the Acc-28 A/D converters connected to the board. The factory default setting
connects pins 1 and 2 of E98, setting a DCLK frequency of 2.46 MHz. It should be left in this setting
unless any Acc-28 A/D converter is connected to the board. In this case, the jumper should be moved to
connect pins 2 and 3, reducing the frequency to 1.23 MHz.
Turbo PMAC System Configuration and Auto Configuration
25
Turbo PMAC User Manual
E29 – E33 Phase Clock Frequency Control
The jumper set E29 – E33 determines the phase clock frequency, controlling its division from the DCLK
signal set by E98. Only one jumper in this set may be installed at any time. The maximum possible
phase clock frequency is 1/68 of the DCLK frequency – 36.14 kHz for the 2.46 MHz DCLK, or 18.07
kHz for the 1.23 MHz DCLK. This is the frequency selected by E33. The following table shows the
possible phase-clock frequencies that can be selected:
Jumper Selected
2.46 MHz DCLK (E98 1-2)
1.23 MHz DCLK (E98 2-3)
E33
E32
E31 (default)
E30
E29
36.14 kHz
18.07 kHz
9.04 kHz
4.52 kHz
2.26 kHz
18.07 kHz
9.04 kHz
4.52 kHz
2.26 kHz
1.13 kHz
E3 – E6 Servo Clock Frequency Control
The jumper set E3 – E6 determines the servo clock frequency, controlling its division from the phase
clock signal set by E29 – E33 and E98. These four jumpers create a binary number N specifying that the
phase clock frequency be divided by N+1 to obtain the servo clock frequency. A jumper ON creates a 0bit; a jumper OFF creates a 1-bit, with E3 as the least significant bit. The following table shows the
possible settings:
E3
E4
E5
E6
Division
Factor N+1
E3
E4
E5
E6
Division
Factor N+1
ON
ON
OFF
ON
ON
OFF
OFF
OFF
ON
ON
OFF
ON
ON
OFF
OFF
OFF
* Default
ON
ON
ON
ON
OFF
OFF
OFF
OFF
ON
ON
ON
ON
ON
ON
ON
ON
1
2
3
4*
5
6
7
8
ON
OFF
ON
OFF
ON
OFF
ON
OFF
ON
ON
OFF
OFF
ON
ON
OFF
OFF
ON
ON
ON
ON
OFF
OFF
OFF
OFF
OFF
OFF
OFF
OFF
OFF
OFF
OFF
OFF
9
10
11
12
13
14
15
16
Setting Turbo PMAC2 Phase and Servo Clock Frequencies
On a Turbo PMAC2 controller (including UMAC and QMAC systems), the system clock frequencies are
determined by the setting of I-variables for the Servo IC or MACRO IC that has been selected as the
clock source for the system. Selection of this IC (usually the first Servo IC in a system without the
MACRO ring, or the first MACRO IC in a system with the MACRO ring) is covered in the preceding
section, and the factory default selection is usually acceptable.
When Servo IC m is selected as the clock source, I-variables I7m00, I7m01, and I7m02 control these
system clock frequencies. When MACRO IC 0 is selected as the clock source, as on Ultralite Turbo
PMAC2 boards, I-variables I6800, I6801, and I6802 control these system clock frequencies.
26
Turbo PMAC System Configuration and Auto Configuration
Turbo PMAC User Manual
PMAC2 Gate Array IC Clock Control
20 MHz
PWM COUNT
120 MHz
x6 Phase
Locked Loop
PWM
Up/Down
Counter
16
16
SIGN
DIR
Max Phase
PWM Max
Count
PWM Dead Time/
PFM Pulse Width
24
Data
8
6
Address
DT / PW
40 MHz
1
3
Encoder Sample n = 0 - 7
Clock Control
PFM
n=0-7
Clock Control
DAC
n=0-7
Clock Control
n=0-7
ADC
Clock Control
Int/Ext Phase
Int/Ext Servo
n = 0 - 15
Phase
Clock Control
n = 0 - 15
Servo
Clock Control
1
2n
SLCK
1
2n
PFMCLK
DACCLK
1
2n
1
2n
1
2n
ADCCLK
PHASE
1
2n
External
Phase
SERVO
External
Servo
I7m00/I6800 MaxPhase Frequency Control
I7m00 for Servo IC m (or I6800 for MACRO IC 0) for the IC selected as system clock source sets the
frequency of the internal MaxPhase clock signal, the maximum possible phase clock frequency. It does
so according to the formula:
MaxPhase( kHz ) =
117 ,964.8
2 * I 7 m00 + 3
At the default value of 6527, it sets a 9.04 kHz frequency for MaxPhase.
I7m00 also sets the PWM output frequency for all channels on Servo IC m; this frequency is equal to
exactly half of the MaxPhase frequency.
I7m01/I6801 Phase Clock Frequency Control
I7m01 for Servo IC m (or I6801 for MACRO IC 1) for the IC selected as system clock source sets the
frequency of the phase clock signal for the system, controlling its division from the MaxPhase clock
frequency. It does so according to the formula:
Phase( kHz ) =
MaxPhase( kHz )
I 7 m01 + 1
At the default value of 0 (divide by 1) and the default MaxPhase frequency of 9.04 kHz, this sets a phase
clock frequency of 9.04 kHz (110 µsec period).
I7m02/I6802 Servo Clock Frequency Control
I7m02 for Servo IC m (or I6802 for MACRO IC 1) for the IC selected as system clock source sets the
frequency of the servo clock signal for the system, controlling its division from the phase clock
frequency. It does so according to the formula:
Servo( kHz ) =
Phase( kHz )
I 7 m02 + 1
Turbo PMAC System Configuration and Auto Configuration
27
Turbo PMAC User Manual
At the default value of 3 (divide by 4) and the default phase clock frequency of 9.04 kHz, this sets a servo
clock frequency of 2.26 kHz (442 µsec period). The following diagram shows the relationship between
the PWM counter, whose frequency is set by the I7m00/I6800 Max Count parameter, the resulting
MaxPhase clock signal, and the Phase and Servo clock signals that are derived from MaxPhase.
PMAC2 Clock Signal Example
+ Max
Count
PWM
Counter
t
- Max
Count
Max
Phase
Phase
Servo
Setting I10 Servo Update Time Parameter
On any Turbo PMAC system, set the I10 servo update time parameter to match the servo frequency
selected. I10 is used by the interpolation algorithms to decide how far to increment position each servo
cycle and it must be set correctly to get trajectories at the proper speed. I10 is scaled such that a value of
223 (8,388,608) matches a servo update type of 1 millisecond (1 kHz servo clock frequency).
The equation for setting I10 as a function of the servo update period is:
I 10 = 8 ,388 ,608 * ServoUpdatePeriod ( m sec) =
8 ,388 ,608
ServoUpdateFreq( kHz )
The default value for I10 of 3,713,707 matches the default servo period of 442 µsec, so if you keep this
default period, you will not have to change I10. Refer to the description of I10 in the Software Reference
manual for detailed instructions on the setting of I10. To get a new value of I10 to take effect, issue a
%100 command for each coordinate system used or a SAVE command and reset the Turbo PMAC.
Setting up a Turbo PMAC2 on the MACRO Ring
Turbo PMAC2 controllers can command axes and I/O over the MACRO ring. Most commonly, this is
done with an Ultralite board-level Turbo PMAC2 controller that is installed as an expansion card in the
host computer, communicating to MACRO Stations, MACRO-based drives, and/or MACRO peripheral
devices over the MACRO ring. However, it is also possible for a UMAC Turbo with the Acc-5E
MACRO interface, or a QMAC with the MACRO-interface option installed, to command devices over
the MACRO ring.
Several variables must be set up properly for proper ring operation. Usually, this is done automatically
through use of the Turbo Setup program on a PC. The following instructions permit direct manual setting
of these variables.
28
Turbo PMAC System Configuration and Auto Configuration
Turbo PMAC User Manual
MACRO Ring Frequency Control Variables
The MACRO ring update frequency is the phase clock frequency of the ring master controller. If there is
more than one Turbo PMAC2 controller on the ring, only one of them can be the ring master controller
(others are masters, but not ring masters). Of course, if there is only one Turbo PMAC2 controller on the
ring, it will be the ring master controller. Determining which Turbo PMAC2 (technically, which
MACRO IC on a Turbo PMAC2) is ring master is explained below.
While the ring master has the capability to force the clock generation of other devices on the ring into
synchronization, it is strongly recommended that all devices on the ring, both other Turbo PMAC2
controllers, and any slave devices, be set up for the same phase clock frequency. Determining which IC
sets the phase clock frequency and the actual setting the phase clock frequency for a Turbo PMAC2
controller is explained above.
For a Turbo PMAC2 driving a MACRO ring, MACRO IC 0 should generate the phase clock signal. This
means that I19 should be set to 6807 (which it will be by default on virtually any Turbo PMAC2 capable
of driving a MACRO ring), and that I6800 and I6801 set the phase clock frequency.
I7: Phase Cycle Extension
On the Turbo PMAC2 board, it is possible to skip hardware phase clock cycles between executions of the
phase update software. A Turbo PMAC2 board will execute the phase update software – commutation
and/or current-loop closure – every (I7+1) hardware phase clock cycles. The default value for I7 is 0, so
normally Turbo PMAC2 executes the phase update software every hardware phase clock cycle.
If the Turbo PMAC2 board is closing the current loop for direct PWM control over the MACRO ring, it is
desirable to have two hardware ring update cycles (which occur at the hardware phase clock frequency)
per software phase update. This eliminates one ring cycle of delay in the current loop, which permits
slightly higher gains and performance. To do this, I7 would be set to 1, so the phase update software
would execute every second hardware phase clock cycle, and ring update cycle.
Normally it is desirable to close the current loop at an update rate of about 9 kHz (the default rate). If two
ring updates were desired per current loop update, the ring update frequency would need to be set to 18
kHz. This is possible if there are no more than 40 total active nodes on the ring. To implement this,
I6800 would be set to one-half of the default value, and I6801 to the default value of 0.
Note:
When making this change, change the Turbo PMAC2’s I6800 variable first, then
the MACRO Station’s MI992. Changing the MACRO Station’s MI992 alone,
followed by an MSSAVE command and an MS$$$, could cause the Station’s
watchdog timer to trip.
I6840: MACRO IC 0 Master Configuration
Any MACRO IC on a Turbo PMAC2 talking to a MACRO Station must be configured as a master on the
ring. For purposes of the MACRO protocol, each MACRO IC is a separate logical master with its own
master number, even though there may be multiple MACRO ICs on a single physical Turbo PMAC2.
Each ring must have one and only one ring controller (synchronizing master). This should be the
MACRO IC 0 one and only one of the Turbo PMAC2 boards on the ring.
On a Turbo PMAC2, set I6840 to $30 to make the card’s MACRO IC 0 the ring controller. This sets bits
4 and 5 of the variable to 1. Setting bit 4 to 1 makes the IC a master on the ring; setting bit 5 to 1 makes
the IC the ring controller” starting each ring cycle by itself.
On a Turbo PMAC2 whose MACRO IC 0 will be a master but not ring controller, I6840 should be set to
$90. This sets bits 4 and 7 of the variable to 1. Setting bit 4 to 1 makes the IC a master on the ring;
setting bit 7 to 1 will cause this IC to be synchronized to the ring controller IC every time it receives a
ring packet specified by I6841.
Turbo PMAC System Configuration and Auto Configuration
29
Turbo PMAC User Manual
I6890/I6940/I6990: MACRO IC 1/2/3 Master Configuration
A Turbo PMAC2 Ultralite may have additional MACRO ICs if Options 1U1, 1U2, and/or 1U3 are
ordered. A UMAC Turbo system may have additional MACRO ICs if Option 1 on an Acc-5E is ordered,
or if multiple Acc-5E boards are ordered. These additional ICs should be set to be masters but not ring
controllers by setting I6890, I6940, and I6990, respectively to $10. This sets bit 4 of the variable to 1,
making the IC a master on the ring. These ICs should never be synchronizing masters, and since they do
not control the clock signals on their own board, their internal clocks do not need to be synchronized to
the ring (only MACRO IC 0 needs to do this).
I6841/I6891/I6941/I6991: MACRO IC 0/1/2/3 Node Activation Control
I6841, I6891, I6941, and I6991 on Turbo PMAC2 control which of the 16 MACRO nodes for MACRO
ICs 0, 1, 2, and 3, respectively, on the card are activated. They also control the master station number for
their respective ICs, and the node number of the packet that creates a synchronization signal. The bits of
these I-variables are arranged as follows:
Bits 0-15: Activation of MACRO Nodes 0 to 15, respectively (1 = active, 0 = inactive). These 16 bits
(usually read as four hex digits) individually control the activation of the MACRO nodes in the MACRO
IC on a Turbo PMAC2. Each node that is active on the matching MACRO Station, whether for servo,
I/O, or auxiliary communications, should have its node activation bit set to 1.
When working with a Delta Tau MACRO Station, Node 15 of each MACRO IC on a Turbo PMAC2
must be activated to permit auxiliary communications, so bit 15 of this variable should always be set to 1
if the IC is used to communicate with a MACRO Station.
Bits 16-19: Packet Sync Node Slave Number. These four bits together (usually read as one hex digit)
form the slave number (0 to 15) of the packet whose receipt by the PMAC2 will set the Sync Packet
Received status bit in the MACRO IC. Usually, this digit is set to $F (15), because Node 15 is always
activated.
Turbo PMAC2 must see this bit set regularly; otherwise it will assume ring problems and shut down servo
and I/O outputs on the ring. Bit 7 of I6840 must be set to 1 on the MACRO IC 0 of all Turbo PMAC2s
that are not ring controllers to enable the synchronization of their phase clocks to that of the ring
controller based on receipt of the sync packet.
Bits 20-23: Master Number. These four bits together form the master number (0 to 15) of the MACRO
IC on the MACRO ring. Each MACRO IC acting as a master on the ring, whether on the same card or
different cards, must have its own master number, and acts as a separate master station for the purposes of
the ring protocol. This master number forms half of the address byte with each packet sent by the
PMAC2 over the MACRO ring.
The master number can be the same number as the MACRO IC number (e.g. MACRO IC 0 has master
number 0, MACRO IC 1 has master number 1, and so on), and if there is only one Turbo PMAC2 in the
ring, this probably will be the case. However, this is not required. The MACRO IC that is the ring
controller must have master number 0 if Type 1 master-to-master auxiliary communications are to be used.
Hex ($)
0
0
0
0
0
0
Bit
Slave node Enables
Sync node Address (0-15)
Master Address (0-15)
30
Turbo PMAC System Configuration and Auto Configuration
Turbo PMAC User Manual
The table shown in an above section and in the Hardware Reference Manual for the 3U MACRO
Station’s SW1 switch setting provides a starting point for the Turbo PMAC2’s I6841/I6891/I6941/I6991
value. Additional bits of these I-variables may be set to 1 if I/O nodes are enabled or if more than one 3U
MACRO station is commanded from a single MACRO IC.
I70/I72/I74/I76: MACRO IC 0/1/2/3 Node Auxiliary Function Enable
I70, I72, I74, and I76 are 16-bit I-variables (bits 0 - 15) in which each bit controls the enabling or
disabling of the auxiliary flag function for the MACRO node number matching the bit number for
MACRO ICs 0, 1, 2, and 3, respectively. A bit value of 1 enables the auxiliary flag function; a bit value
of 0 disables it. If the function is enabled, PMAC automatically copies information between the MACRO
interface flag register and RAM register $00344n, $00345n, $00346n, and $00347n (where n is the IC’s
node number 0 – 15) for MACRO ICs 0, 1, 2, and 3, respectively.
Note that Turbo PMAC MACRO node numbers (as opposed to individual MACRO IC node numbers) go
from 0 to 63, with board nodes 0 – 15 on MACRO IC 0, board nodes 16 – 31 on MACRO IC 1, board
nodes 32 – 47 on MACRO IC 2, and board nodes 48 – 63 on MACRO IC 3.
Each MACRO node n that is used for servo functions should have the corresponding bit n of I70, I72,
I74, or I76 set to 1. Ixx25 for the Motor x that uses Node n should then address $00344n, $00345n,
$00346n, or $00347n, not the address of the MACRO register itself (see below). If Register 3 of a
MACRO node n is used for other purposes, such as direct I/O, the corresponding bit n of I70, I72, I74, or
I76 should be set to 0, so this copying function does not overwrite these registers.
Typically, non-servo I/O functions with a MACRO Station do not involve auxiliary flag functions, so this
flag copy function should remain disabled for any node used to transmit I/O between the Turbo PMAC2
and the MACRO Station. If any auxiliary communications is done between the Turbo PMAC2 and the
MACRO Station on Nodes 14 and/or 15, bits 14 and 15 of these variables must be set to 0.
Examples:
I70=$3
I72=$30
I74=$3300
I76=$3333
;
;
;
;
Enabled
Enabled
Enabled
Enabled
for
for
for
for
MACRO
MACRO
MACRO
MACRO
IC
IC
IC
IC
0
1
2
3
Nodes
Nodes
Nodes
Nodes
0 and 1
4 and 5
8,9,12,13
0,1,4,5,8,9,12,13
I71/I73/I75/I77: MACRO IC 0/1/2/3 Node Protocol Type Control
I71, I73, I75, and I77 are 16-bit I-variables (bits 0 - 15) in which each bit controls whether PMAC uses
the uses MACRO Type 0 protocol or the MACRO Type 1 protocol for the node whose number matches
the bit number for the purposes of the auxiliary servo flag transfer for MACRO ICs 0, 1, 2, and 3,
respectively. A bit value of 0 sets a Type 0 protocol; a bit value of 1 sets a Type 1 protocol.
All 3U MACRO Station nodes use the Type 1 protocol, so each MACRO node n used for servo purposes
with a MACRO Station must have bit n of I1002 set to 1. Generally I71 = I70, I73 = I72, I75 = I74, and
I77 = I76 on a Turbo PMAC2 communicating with a MACRO Station.
Remember that if servo nodes for more than one MACRO Station are commanded from a single MACRO
IC, the protocol must be selected for all of the active servo nodes on each station.
I78: MACRO Master/Slave Auxiliary Communications Timeout
If I78 is set greater than 0, the MACRO Type 1 Master/Slave Auxiliary Communications protocol using
Node 15 is enabled. Turbo PMAC implements this communications protocol using the MACROSLAVE
(MS), MACROSLVREAD (MSR), and MACROSLVWRITE (MSW) commands.
If this function is enabled, I78 sets the timeout value in PMAC servo cycles. In this case, if PMAC does
not get a response to a Node 15 auxiliary communications command within I78 servo cycles, it will stop
waiting and register a MACRO auxiliary communications error, setting Bit 5 of global status register
X:$000006.
Turbo PMAC System Configuration and Auto Configuration
31
Turbo PMAC User Manual
I78 must be set greater than 0 if any auxiliary communications is desired with a MACRO Station. This
reserves Node 15 for the Type 1 Auxiliary Communications. A value of 32 is suggested. If I78 is set
greater than 0, bit 15 of I70, I72, I74, and I76 must be set to 0, so Node 15 is not used for flag transfers also.
I79: MACRO Master/Master Auxiliary Communications Timeout
If I79 is set greater than 0, the MACRO Type 1 Master/Master Auxiliary Communications protocol using
Node 14 is enabled. Turbo PMAC implements this communications protocol using the MACROMASTER
(MM), MACROMSTREAD (MMR), and MACROMSTWRITE (MMW) commands. Only the Turbo PMAC that is
the “ring controller” can execute these commands; other Turbo PMACs that are masters on the ring can
respond to these commands from the ring controller.
If this function is enabled, I79 sets the “timeout” value in PMAC servo cycles. In this case, if the Turbo
PMAC does not get a response to a Node 14 master/master auxiliary communications command within
I79 servo cycles, it will stop waiting and register a “MACRO auxiliary communications error,” setting Bit
5 of global status register X:$000006.
I79 must be set greater than 0 if any auxiliary communications is desired with a MACRO Station. A
value of 32 is suggested. If a value of I79 greater than 0 has been saved into PMAC’s non-volatile
memory, then at subsequent power-up/resets, bit 14 of I70 is set to 0, the node-14 broadcast bit (bit 14 of
I6840) is set to 1, and activation bit for node 14 (bit 14 of I6841) is set to 1, regardless of the value saved
for these variables. This reserves Node 14 of MACRO IC 0 for the Type 1 Master/Master Auxiliary
Communications.
I80, I81, I82: MACRO Ring Check Period and Limits
If I80 is set to a value greater than zero, Turbo PMAC will monitor for MACRO ring breaks or repeated
MACRO communications errors automatically. A non-zero value sets the error detection cycle time in
Turbo PMAC servo cycles. Turbo PMAC checks to see that “sync node” packets (see I6840 and I6841)
are received regularly, and that there have not been regular communications errors.
The limits for these checks can be set with variables I81 and I82. If less than I82 sync node packets have
been received and detected during this time interval, or if I81 or more ring communications errors have
been detected in this interval, Turbo PMAC will assume a major ring problem, and all motors will be shut
down. Turbo PMAC will set the global status bit “Ring Error” (bit 4 of X:$000006) as an indication of
this error.
Turbo PMAC looks for receipt of sync node packets and ring communications errors once per real-time
interrupt – every (I8 + 1) servo cycles). The time interval set by I80 must be large enough that I82 realtime interrupts in PMAC can execute within the time interval, or false ring errors will be detected.
Remember that long motion program calculations can cause skips in the real-time interrupt. Typically
values of I80 setting a time interval of about 20 milliseconds are used. I80 can be set according to the
formula:
I80 = Desired cycle time (msec) * Servo update frequency (kHz)
For example, with the default servo update frequency of 2.26 kHz, to get a ring check cycle interval of 20
msec, I80 would be set to 20 * 2.26 ≅ 45.
MACRO Node Addresses
The MACRO ring operates by copying registers at high speed across the ring. Therefore, each Turbo
PMAC2 master controller on the ring communicates with its slave stations by reading from and writing to
registers in its own address space. MACRO hardware handles the data transfers across the ring automatically.
32
Turbo PMAC System Configuration and Auto Configuration
Turbo PMAC User Manual
Starting in Turbo firmware version 1.936, the base addresses of the up to 4 MACRO ICs must be
specified in I20 – I23, for MACRO IC 0 – 3 respectively. Before this, the base addresses were fixed at
$078400, $079400, $07A400, and $07B400, respectively. Only UMAC Turbo systems can support any
other configuration, and only rarely will another configuration be used.
The following table gives the addresses of the MACRO ring registers for Turbo PMAC2 controllers.
Note:
It is possible, although unlikely, to have other addresses in a UMAC Turbo system.
In these systems, the fourth digit does not have to be 4; it can also take the values
5, 6, and 7.
Register Addresses for MACRO IC 0 with I20=$078400 (default)
Node #
Turbo PMAC2
Reg. 0
Addresses:
Reg. 1
MACRO IC 0
Reg. 2
Reg. 3
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Y:$078420
Y:$078424
X:$078420
X:$078424
Y:$078428
Y:$07842C
X:$078428
X:$07842C
Y:$078430
Y:$078434
X:$078430
X:$078434
Y:$078438
Y:$07843C
X:$078438
X:$07843C
Y:$078421
Y:$078425
X:$078421
X:$078425
Y:$078429
Y:$07842D
X:$078429
X:$07842D
Y:$078431
Y:$078435
X:$078431
X:$078435
Y:$078439
Y:$07843D
X:$078439
X:$07843D
Y:$078422
Y:$078426
X:$078422
X:$078426
Y:$07842A
Y:$07842E
X:$07842A
X:$07842E
Y:$078432
Y:$078436
X:$078432
X:$078436
Y:$07843A
Y:$07843E
X:$07843A
X:$07843E
Y:$078423
Y:$078427
X:$078423
X:$078427
Y:$07842B
Y:$07842F
X:$07842B
X:$07842F
Y:$078433
Y:$078437
X:$078433
X:$078437
Y:$07843B
Y:$07843F
X:$07843B
X:$07843F
Register Addresses for MACRO IC 1 with I21=$079400 (default)
Node #
Turbo PMAC2
Reg. 0
Addresses:
Reg. 1
MACRO IC 1
Reg. 2
Reg. 3
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Y:$079420
Y:$079424
X:$079420
X:$079424
Y:$079428
Y:$07942C
X:$079428
X:$07942C
Y:$079430
Y:$079434
X:$079430
X:$079434
Y:$079438
Y:$07943C
X:$079438
X:$07943C
Y:$079421
Y:$079425
X:$079421
X:$079425
Y:$079429
Y:$07942D
X:$079429
X:$07942D
Y:$079431
Y:$079435
X:$079431
X:$079435
Y:$079439
Y:$07943D
X:$079439
X:$07943D
Y:$079422
Y:$079426
X:$079422
X:$079426
Y:$07942A
Y:$07942E
X:$07942A
X:$07942E
Y:$079432
Y:$079436
X:$079432
X:$079436
Y:$07943A
Y:$07943E
X:$07943A
X:$07943E
Y:$079423
Y:$079427
X:$079423
X:$079427
Y:$07942B
Y:$07942F
X:$07942B
X:$07942F
Y:$079433
Y:$079437
X:$079433
X:$079437
Y:$07943B
Y:$07943F
X:$07943B
X:$07943F
Turbo PMAC System Configuration and Auto Configuration
33
Turbo PMAC User Manual
Register Addresses for MACRO IC 2 with I22=$07A400 (default)
Node #
Turbo PMAC2
Reg. 0
Addresses:
Reg. 1
MACRO IC 2
Reg. 2
Reg. 3
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Y:$07A420
Y:$07A424
X:$07A420
X:$07A424
Y:$07A428
Y:$07A42C
X:$07A428
X:$07A42C
Y:$07A430
Y:$07A434
X:$07A430
X:$07A434
Y:$07A438
Y:$07A43C
X:$07A438
X:$07A43C
Y:$07A421
Y:$07A425
X:$07A421
X:$07A425
Y:$07A429
Y:$07A42D
X:$07A429
X:$07A42D
Y:$07A431
Y:$07A435
X:$07A431
X:$07A435
Y:$07A439
Y:$07A43D
X:$07A439
X:$07A43D
Y:$07A422
Y:$07A426
X:$07A422
X:$07A426
Y:$07A42A
Y:$07A42E
X:$07A42A
X:$07A42E
Y:$07A432
Y:$07A436
X:$07A432
X:$07A436
Y:$07A43A
Y:$07A43E
X:$07A43A
X:$07A43E
Y:$07A423
Y:$07A427
X:$07A423
X:$07A427
Y:$07A42B
Y:$07A42F
X:$07A42B
X:$07A42F
Y:$07A433
Y:$07A437
X:$07A433
X:$07A437
Y:$07A43B
Y:$07A43F
X:$07A43B
X:$07A43F
Register Addresses for MACRO IC 3 with I23=$07B400 (default)
Node #
Turbo PMAC2
Reg. 0
Addresses:
Reg. 1
MACRO IC 3
Reg. 2
Reg. 3
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Y:$07B420
Y:$07B424
X:$07B420
X:$07B424
Y:$07B428
Y:$07B42C
X:$07B428
X:$07B42C
Y:$07B430
Y:$07B434
X:$07B430
X:$07B434
Y:$07B438
Y:$07B43C
X:$07B438
X:$07B43C
Y:$07B421
Y:$07B425
X:$07B421
X:$07B425
Y:$07B429
Y:$07B42D
X:$07B429
X:$07B42D
Y:$07B431
Y:$07B435
X:$07B431
X:$07B435
Y:$07B439
Y:$07B43D
X:$07B439
X:$07B43D
Y:$07B422
Y:$07B426
X:$07B422
X:$07B426
Y:$07B42A
Y:$07B42E
X:$07B42A
X:$07B42E
Y:$07B432
Y:$07B436
X:$07B432
X:$07B436
Y:$07B43A
Y:$07B43E
X:$07B43A
X:$07B43E
Y:$07B423
Y:$07B427
X:$07B423
X:$07B427
Y:$07B42B
Y:$07B42F
X:$07B42B
X:$07B42F
Y:$07B433
Y:$07B437
X:$07B433
X:$07B437
Y:$07B43B
Y:$07B43F
X:$07B43B
X:$07B43F
Note:
With the MACRO station, only nodes that map into Turbo PMAC2 Y registers (0,
1, 4, 5, 8, 9, 12, and 13) can be used for servo control. These nodes are unshaded
in the above table. The nodes that map into X registers (2, 3, 6, 7, 10, 11, and 14)
can be used for I/O control. Node 15 is reserved for Type 1 auxiliary
communications. Node 14 is often reserved for broadcast communications.
34
Turbo PMAC System Configuration and Auto Configuration
Turbo PMAC User Manual
Resetting and Re-Initializing Turbo PMAC
It is important to understand how a Turbo PMAC system can be reset or re-initialized and what actions
are performed in each case.
Methods of Resetting
There are fundamentally three ways a Turbo PMAC system can be reset:
1. Cycling power to the digital circuits
2. Hardware reset
3. Software reset
Cycling Power
Removing power to the digital circuits of a Turbo PMAC, then re-applying that power, forces the Turbo
PMAC to go through its reset cycle. On most Turbo PMACs, this supply is 5V power, regulated down as
needed for lower-voltage components such as the CPU and RAM. As of this writing, only the UMACCPCI takes a separate 3.3V supply for the CPU section.
Removal and re-application of the +/-12V to +/-15V supply for analog circuits, whether or not these
circuits are optically isolated from the digital circuitry, does not cause a Turbo PMAC system to reset
(although it may cause a fault condition).
Hardware Reset
A hardware reset of a Turbo PMAC system is accomplished by taking a special digital input line (INIT/)
low, then releasing it high again (it has a internal pull-up resistor). There are several possible sources for
this signal. On board-level Turbo PMAC controllers, this signal is available on the JPAN control-panel
port (Turbo PMAC only) and the JTHW multiplexer ports. On the modular UMAC and UMAC-CPCI
systems, it is an input on the CPU board.
For Turbo PMAC boards that install in an ISA or VME bus, the bus reset line can be tied to the Turbo
PMAC’s reset line through the installation of jumper E39 on the Turbo PMAC, so that a reset of the main
computer also forces a hardware reset of the Turbo PMAC.
While the hardware reset line is held low, the CPU ceases to operate, and the Servo, MACRO, and I/O
ASICs in the system are held in their reset state, which forces all discrete outputs to their OFF state, and
all continuously variable outputs to their zero state.
Software Reset
A software reset of a Turbo PMAC system is accomplished by issuing the $$$ on-line reset command.
Actions on a Normal Reset
If any of the above methods for resetting is used when the Turbo PMAC is configured for a normal reset,
the standard actions described in this section will occur. A Turbo PMAC is configured for a normal reset
when both of the following conditions are true:
1. The re-initialization jumper (E51 on a Turbo PMAC, E3 on a Turbo PMAC2) is not installed.
2. The four bootstrap mode jumpers for the CPU (E4 – E7 on the piggyback Turbo CPU board, E20 –
E23 on UMAC and UMAC-CPCI CPU boards) are in their standard configuration (outer two jumpers
OFF, inner two jumpers ON).
For a Turbo PMAC configured in this manner, when digital power is applied, or the hardware reset line is
released to go high, or the $$$ software reset command is given, the following actions occur:
1. The installed firmware is loaded from the flash memory into active memory.
Turbo PMAC System Configuration and Auto Configuration
35
Turbo PMAC User Manual
2. The last-saved user configuration – variable values and definitions, user programs, tables, and buffers
– are loaded from the flash memory into active memory and registers. During this loading, the
checksums of the saved data are evaluated. If the checksum for the saved I-variables does not match
the data, all I-variables in active memory are returned to their factory default values. If the checksum
for the programs and buffers does not match the data, all of these programs and buffers are
completely cleared from active memory.
3. The basic configuration of the system – memory capacity, ASIC presence, location, and type – is
checked and logged. Counters in all ASICs are cleared.
4. All motors with Ixx80 bit 0 set to 1 are enabled.
5. All existing PLC programs whose operation is permitted by the saved value of I5 are activated.
Actions on Reset with Re-Initialization
If any of the above methods for resetting is used when the Turbo PMAC is configured for re-initialization,
the actions described in this section will occur. A Turbo PMAC is configured for re-initialization when
both of the following conditions are true:
1. The re-initialization jumper (E51 on a Turbo PMAC, E3 on a Turbo PMAC2) IS installed.
2. The four bootstrap mode jumpers for the CPU (E4 – E7 on the piggyback Turbo CPU board, E20 –
E23 on UMAC and UMAC-CPCI CPU boards) are in their standard configuration (outer two jumpers
off, inner two jumpers on).
For a Turbo PMAC configured in this manner, when digital power is applied, or the hardware reset line is
released to go high, or the $$$ software reset command is given, the following actions occur:
1. The installed firmware is loaded from the flash memory into active memory.
2. The factory default I-variables are loaded from firmware into active memory and registers. (The last
saved values in flash are not lost; they are simply not used.) The last saved user programs, table and
buffers are loaded into active memory, but none will be active because of the default I-variable
settings. If the checksum for the programs and buffers does not match the data, all of these programs
and buffers are completely cleared from active memory.
3. The basic configuration of the system – memory capacity, ASIC presence, location, and type – is
checked and logged. The CPU will make some decisions about default I-variable values based on this
configuration information. Counters in all ASICs are cleared.
4. Because of the default I-variable configuration, no motors are enabled, and no programs are activated.
Actions on Reset for Firmware Reload
If any of the above methods for resetting is used when the Turbo PMAC is configured for firmware
reload, the actions described in this section will occur. A Turbo PMAC is configured for firmware reload
when the four bootstrap mode jumpers for the CPU (E4 – E7 on the piggyback Turbo CPU board, E20 –
E23 on UMAC and UMAC-CPCI CPU boards) are in their firmware-reload configuration (first jumper
OFF, last three jumpers ON).
For a Turbo PMAC configured in this manner, when digital power is applied, or the hardware reset line is
released to go high, or the $$$ software reset command is given, only the bootstrap firmware is loaded
into active memory. At this point, it is ready to accept the download of new operational firmware into its
flash memory through either the main serial port or a bus port that operates through the host port of the
DSP (not VME or DPRAM).
The PMAC Executive program, when it establishes communications with a Turbo PMAC reset into this
mode, will detect that the Turbo PMAC is in bootstrap mode and ready to accept new firmware
automatically. It will ask for the name of the file containing the firmware to download.
36
Turbo PMAC System Configuration and Auto Configuration
Turbo PMAC User Manual
Updating the firmware will cause Turbo PMAC to revert to default I-variables. Make sure the
configuration is backed up before the new firmware is downloaded.
If writing a custom application to perform this function, detect that the Turbo PMAC is in bootstrap mode
by sending the ? query command. In this mode, it will respond with the string BOOTSTRAP PROM
instead of a hexadecimal status word. Sending the <CTRL-O> character will prepare the Turbo PMAC
for the downloading of the binary firmware file. Wait at least five seconds after sending this character to
make sure that the flash IC is ready to accept data. Sending the <CTRL-R> command will take the Turbo
PMAC out of bootstrap mode and cause it to take the actions of a normal reset.
Re-Initialization and Clear Command
The $$$*** command causes a reset and full re-initialization of Turbo PMAC. In addition to loading
default I-variable values, it also clears out all of the buffers in active RAM: motion program, PLC
program, tables, etc.
Some users will always have the card set up to re-initialize and clear during the reset cycle; they then
download all parameter settings and programs immediately after each cycle. The logic behind this
strategy is that the same startup sequence of operations is used even if a new replacement board has just
been put in. It is also useful for those applications that do not wish to rely in any way on Turbo PMAC’s
own non-volatile flash storage.
For a complete re-initialization of Turbo PMAC to known state, the following commands can be added:
P0..8191=0
Q0..8191=0
M0..8191->*
UNDEFINE ALL
Remember that these commands directly affect only active memory (RAM). To copy new settings into
non-volatile flash memory, use the SAVE command.
Turbo PMAC System Configuration and Auto Configuration
37
Turbo PMAC User Manual
38
Turbo PMAC System Configuration and Auto Configuration
Turbo PMAC User Manual
TALKING TO TURBO PMAC
This section covers the basic aspects of communicating with Turbo PMAC from a host computer. At this
level, we are assuming that there is a program for the host computer that processes these communications.
The PMAC Executive Program (Pewin32 Pro) is the most common of these programs.
If there will be a host computer in the final application, write custom communications routines for the
host computer as part of the front-end software for the application. That is a more advanced topic, and it
is covered in the Writing a Host Communications Program section. Delta Tau provides communications
libraries for use under Microsoft Windows operating systems.
At a basic level, Turbo PMAC can communicate to a host dumb terminal, over any of the ports. The
communications mostly consists of lines of ASCII characters sent back and forth. Of course, most of the
time the host will be a computer with considerably more intelligence, but at root it will talk to the card as
if it were a terminal. The PMAC Executive PC program has a terminal emulator mode to do this directly.
Communications Ports
Each version of Turbo PMAC has available multiple communications ports: one or two serial ports, a bus
port, and an optional dual-ported RAM shared-memory communications interface. Different
configurations of the Turbo PMAC will have different types of these interfaces.
Turbo PMAC can communicate simultaneously over multiple ports, without the communications on
different ports interfering with each other (the original PMAC family does not have this capability). This
gives the user useful capabilities such as employing both a host computer and an operator pendant, or
using one port for the applications communications and another for simultaneous monitoring and
debugging.
Serial Communications Ports
Turbo PMAC controllers have one or two serial communications ports that provide a simple
communications interface to host computers, terminals, or other devices. Depending on the particular
configuration and port, the port can use RS-232 or RS-422 signal levels.
Main Serial Port
Each configuration of Turbo PMAC comes standard with a serial communications port. Depending on
the configuration, this may be available at RS-232 levels (single-ended +/-10V) only, RS-422 levels
(differential 5V) only, or user-selectable by jumpers between the two levels. The RS-422 circuits on a
Turbo PMAC are capable of interfacing directly to an RS-232 port on a computer, but noise margins are
significantly reduced, so communications may be quite susceptible to electromagnetic noise. If both
levels are available, it is important to realize that this is only a single port, and that only one voltage level
may be used at any given time.
The baud rate for the main serial port is set by variable I54. At power-up reset, Turbo PMAC sets the
active baud-rate-control register based on the setting of I54 and the CPU speed as set by I52, as the baudrate frequency is divided down from the CPU’s operational frequency. The factory default baud rate is
38,400. This baud rate will be selected automatically on re-initialization of the Turbo PMAC, either
through use of the re-initialization jumper or the $$$*** command. To change the baud rate, change the
setting of I54, copy this new setting to non-volatile memory with the SAVE command, then reset the
Turbo PMAC. Re-establish communications at the new baud rate.
Talking to Turbo PMAC
39
Turbo PMAC User Manual
The values of I54 and the baud rates they produce are:
I54
Baud Rate
I54
Baud Rate
I54
Baud Rate
I54
Baud Rate
0
1
2
3
600
900
1200
1800
4
5
6
7
2400
3600
4800
7200
8
9
10
11
9600
14,400
19,200
28,800
12
13
14
15
38,400
57,600
76,800
115,200
The baud rates produced by odd-number settings of I54 are only exact if the CPU frequency is an exact
multiple of 30 MHz (technically, of 29.4912 MHz). This is because the baud rates are created by dividing
the CPU frequency by (256 * N), where N is an integer taken from a lookup table. The frequency is not
an exact match for odd settings of I54 and CPU frequencies that are not multiples of 30 MHz. For lower
baud rates of this type, the error is not significant. However, serial communications at 115,200 baud is
possible only if the CPU is running at an exact multiple of 30 MHz (actually an exact multiple of 29.4912
MHz). So to communicate at this rate, run an Option 5Cx 80 MHz CPU at 60 MHz, an Option 5Dx 100
MHz CPU at 90 MHz, and an Option 5Ex 160 MHz CPU at 150 MHz by setting I52 to a lower value than
the CPU is capable of.
It is possible to download new operational firmware through the main serial port (see Resetting Turbo
PMAC).
Auxiliary Serial Port
If Option 9T for the Turbo PMAC is ordered, directly or as part of an option package (as for the PMAC
Ladder programming environment), a second serial port is provided on the Turbo PMAC. This port is
required for programming and monitoring the PMAC Ladder IEC-1131 PLC programs.
Auxiliary Port Baud Rate
The baud rate for the auxiliary serial port is set by variable I53. At power-up reset, Turbo PMAC sets the
active baud-rate-control register based on the setting of I53 and the CPU speed as set by I52, as the baudrate frequency is divided down from the CPU’s operational frequency. The factory default baud rate is
38,400. If you wish to change the baud rate, you must change the setting of I53, copy this new setting to
non-volatile memory with the SAVE command, then reset the Turbo PMAC. Then you must re-establish
communications at the new baud rate. The values of I53 and the baud rates they produce are:
I53
Baud Rate
I53
Baud Rate
I53
Baud Rate
I53
Baud Rate
0
1
2
3
600
900
1200
1800
4
5
6
7
2400
3600
4800
7200
8
9
10
11
9600
14,400
19,200
28,800
12
13
14
15
38,400
57,600
76,800
115,200
The baud rates produced by odd-number settings of I53 are only exact if the CPU frequency is an exact
multiple of 30 MHz (technically, of 29.4912 MHz). This is because the baud rates are created by dividing
the CPU frequency by (256 * N), where N is an integer taken from a lookup table. The frequency is not
an exact match for odd settings of I53 and CPU frequencies that are not multiples of 30 MHz. For lower
baud rates of this type, the error is not significant. However, serial communications at 115,200 baud is
only possible if the CPU is running at an exact multiple of 30 MHz (actually an exact multiple of 29.4912
MHz). So to communicate at this rate, run an Option 5Cx 80 MHz CPU at 60 MHz, an Option 5Dx 100
MHz CPU at 90 MHz, and an Option 5Ex 160 MHz CPU at 150 MHz by setting I52 to a lower value than
the CPU is capable of.
40
Talking to Turbo PMAC
Turbo PMAC User Manual
Auxiliary Port Parser Disable
It is possible to turn off the automatic command parser on the auxiliary serial port by setting I43 to 1.
With I43 set to the default value of 0, Turbo PMAC will try to interpret any characters coming in the
auxiliary serial port as part of Turbo PMAC commands. However, with I43 set to 1, these characters are
just placed in the rotary command queue at X:$1C00 – X:$1CFF, where a user’s application program can
interpret them.
It is not possible to download new operational firmware through the auxiliary serial port.
Signal Format
Since serial interfaces vary from system to system, Turbo PMAC provides a simple but flexible interface.
In addition to the signal ground line, only four lines are required (eight if counting the complements in an
RS-422 interface): data-transmit, data-receive, clear-to-send, and ready-to-send. Turbo PMAC serial
ports simply short together the DSR and DTR handshake lines to provide an automatic return signal on
this strobe for those systems that require it.
The send and receive signals, for both data and handshake, are placed on the connector in the Data
Terminal Equipment (DTE) configuration (on the Turbo PMAC-PC, this can be changed using jumpers
E9 – E16). This permits a straight-across connection to the serial port configured as Data
Communications Equipment (DCE), which is the configuration found on virtually all modern computers.
However, if connecting the serial port to another device configured as DTE, such as a handheld
terminal/pendant, cross the data and handshake signals.
The pin-outs of the box-header connectors used for the serial ports on most Turbo PMACs are designed
so that a flat-cable connection to a D-sub connector on the other end will provide the proper connection if
the other end is in DCE configuration.
Data Format and Handshaking
The serial communications data format for each character is 8 bits, 1 start bit, 1 stop bit, no parity. (There
may be an option to permit parity in future firmware revisions.) No full-duplex echoing of characters is
supported.
RTS/CTS handshaking is enabled by default. Because of Turbo PMAC’s very high speed in handling
communications, it will never hold off a character unless it is locked in a reset or watchdog state. Turbo
PMAC will ignore the CTS handshake from the host computer on the main serial port if I1 is set to 1 or 3,
but in this case ensure high-priority monitoring of the serial port to make sure that no characters are lost
(standard Microsoft Windows serial drivers do not have high enough priority for this).
Most Turbo PMAC serial ports short together the DSR and DTR lines to provide an echo handshake on
those ports that require this. No XON/XOFF software handshaking is supported.
Multi-Drop Communications
With an RS-422 serial port, it is possible to communicate with multiple daisy-chained Turbo PMACs
over a single multi-drop cable. Up to 16 Turbo PMACs can be addressed on a single cable.
I-Variable Enabling
Turbo PMAC variable I0 determines the controller’s software address on a multi-drop serial cable. I0 has
a range of 0 to 15 (0 to F hex). Each controller on the cable must have a unique value of I0. Variable I1
on each Turbo PMAC must be set to 2 or 3 to enable the serial software addressing.
Talking to Turbo PMAC
41
Turbo PMAC User Manual
Serial Controller Addressing
A controller is addressed with the @n command, where n is a hex digit from 0 to F. When the @n
command is sent over the serial cable, the controller whose value of I0 is equal to n becomes the
addressed controller; all others become unaddressed. Only the addressed controller will respond to
commands and acknowledge them, driving its output data and handshake lines (which are tri-stated on the
unaddressed controllers to prevent signal contention).
If the @@ command is sent over the serial cable, all controllers on the cable will respond to action
commands. Commands requiring a data response are not permitted in this mode. The Turbo PMAC
whose I0 value is 0 will drive its output data and handshake lines (all others tri-state theirs) and
acknowledge host commands.
Optional Clock Sharing
It is possible to share servo and phase clock signals over additional lines on the RS-422 multi-drop cable.
This can prevent long-term drift in continuous motion between controllers due to tolerances in the clockcrystal frequency of each. Jumpers on each Turbo PMAC determine whether the controller will generate
its own servo and phase clock signals and output them on its serial port, or whether it will expect them as
inputs from its serial port. Only one controller on a multi-drop cable can be outputting its servo and
phase clocks. Only the frequency-control variables or jumpers on this controller affect the frequency of
the servo and phase clocks for all controllers on the cable.
If the servo and phase clock signals are not being shared between controllers, these signals should not be
connected between controllers through the cable.
Bus Communications Port
Each configuration of Turbo PMAC has a bus port that provides faster communications than the serial
port. There are both the traditional backplane buses (ISA, PCI, VME) and the newer wire buses (USB,
Ethernet).
Most of these bus ports use the host port on the Turbo PMAC’s CPU (the VME bus uses a set of mailbox
registers instead). Many of them can also use optional dual-ported RAM (DPRAM) as an alternate path
to communicate with the CPU, although this communications uses the same physical path between the
host computer and the Turbo PMAC.
It is possible to download new operational firmware through a bus communications port that uses the host
port on the Turbo PMAC’s CPU (see Resetting Turbo PMAC). This excludes the VME bus and any wire
bus interfaces that are using DPRAM communications only.
Backplane Buses
Several types of backplane buses are supported in various Turbo PMAC configurations. These buses
provide the highest-speed and lowest-latency communications, due to their high raw data rates (compared
to simple serial interfaces) and absence of intermediate processors (compared to the wire buses).
ISA (and PC/104) Bus
The ISA bus interface comes standard on board-level Turbo PMAC controllers with the –PC suffix (e.g.
Turbo PMAC-PC, Turbo PMAC2-PC). The PC/104 bus interface, a stack version of the ISA bus
(equivalent electrically and in software), comes standard on the 3U-format UMAC Turbo CPU board
(formerly Option 2 for this board, now always included).
The basic communications on the ISA bus goes through the host port on the Turbo PMAC’s CPU. Dualported RAM is an optional addition to this bus interface (see below). The address of the host port
communications in the I/O space of the ISA bus is selected by a bank of jumpers or DIP switches on the
Turbo PMAC board. The factory default setting is for base I/O address 528 (210 hex), and the interface
occupies 16 consecutive addresses in the PC’s I/O space.
42
Talking to Turbo PMAC
Turbo PMAC User Manual
The following table shows how the jumpers (Turbo PMAC boards) or DIP switches (Turbo PMAC2) set
the base address on the ISA or PC/104 port. A PMAC jumper that is OFF or a PMAC2 DIP-switch that is
OPEN adds the associated bit value to the base address; a jumper that is ON or a DIP-switch that is
CLOSED adds nothing to the base address value.
15
14
13
12
11
10
9
8
7
6
5
4
Address Bit #
x
x
x
x
E91
E92
E66
E67
E68 E69 E70
E71
PMAC1 Jumper
12
11
10
9
8
7
6
5
4
3
2
1
PMAC2 Switch
32768 16384 8192
4096 2048 1024 512
256
128
64
32
16
Bit Val (Dec)
$8000 $4000 $2000 $1000 $800 $400 $200 $100 $80
$40
$20
$10
Bit Val (Hex)
ON
ON
OFF
ON
ON
ON
ON OFF
Default Jumper
CLS
CLS
CLS
CLS
CLS CLS OPN CLS CLS CLS CLS OPN
Default Switch
PCI Bus
The PCI bus interface comes standard on board-level Turbo PMAC controllers with the –PCI suffix (e.g.
Turbo PMAC-PCI, Turbo PMAC2-PCI). The UMAC-CPCI Turbo CPU board has been designed to
accept a daughter board that implements the Compact PCI interface, a rack-mounted version of the PCI
bus (equivalent electrically and in software), but as of this writing the CPCI daughter board has not been
implemented.
The basic communications on the PCI bus goes through the host port on the Turbo PMAC’s CPU. Dualported RAM is an optional addition to this bus interface (see below). The address of the host port
communications in the I/O space of the PCI bus is selected by the host computer’s operating system.
Recent Microsoft Windows operating systems do this on a plug-and-play basis. The design of the PCI
bus interface on Turbo PMACs was implemented to be as much as possible like the older ISA bus
interface, to minimize any transition problems from users changing over from ISA. Other than the
automatic address setting, the software interface to the PCI bus on Turbo PMACs is identical to that for
the ISA bus.
VME Bus
The VME bus interface comes standard on board-level Turbo PMAC controllers with the –VME suffix
(e.g. Turbo PMAC-VME, Turbo PMAC2-VME). It is a slave interface on the bus, and can be used with
16, 24, and 32-bit addressing on the bus. The address, bus width, and other details of the interface are set
by the values of Turbo PMAC variables I90 – I99 at power-up/reset. The factory default setting is for 24bit addressing, with a base address of $7FA000 on the VME bus.
The most basic communications on the VME bus goes through a set of 16 mailbox registers in the VME
interface IC on the Turbo PMAC, each capable of holding one character. Note that unlike the other
buses, this communications does not go through the Turbo CPU’s host port, so it is not possible to
download new operational firmware in bootstrap mode to the Turbo PMAC over the VME bus. Dualported RAM is an optional addition to this bus interface (see below), but one that is virtually always used.
If DPRAM is present, usually it is easier to send text commands through the DPRAM, and the mailbox
registers are typically not used.
The following table shows how I90 – I99 should be set for 24-bit and 32-bit addressing (16-bit is almost
never used), with and without dual-ported RAM. The values in the table are for a mailbox base address
on the bus of $abcdef00, and a DPRAM base address of $abg00000 (a and b are not used in 24-bit
addressing).
I Variable
I90
I91
I92
I93
I94
I95
I96
I97
I98
I99
24-bit, no DPRAM
24-bit, w/ DPRAM
32-bit, no DPRAM
32-bit, w/ DPRAM
$29
$29
$39
$39
$04
$04
$04
$04
*
*
$ab
$ab
$cd
$cd
$cd
$cd
$ef
$ef
$ef
$ef
**
**
**
**
**
**
**
**
*
$0g
*
$0g
$60
$E0
$60
$E0
$10
$90
$10
$90
Talking to Turbo PMAC
43
Turbo PMAC User Manual
* Don’t care, not used
** System dependent setting
Delta Tau’s software packages for the PC, such as the Executive and Setup programs, and the PComm32
communications library, do not support VME bus communications, because there is no standard VME
communications method under the Microsoft Windows operating systems, even for PC-compatible VMEbus computers.
Wire Buses
Recently, high-speed wire links have become cost-effective and widespread enough to provide an
economical but high-bandwidth link between host computer and controller. Delta Tau presently supports
two of these links: USB and Ethernet. These links essentially match the speed of backplane buses for the
transfer of large data sets in either direction, but are only about half as fast in response time to short
commands, due to added delays from the processors at each end of the link.
USB
Several Turbo PMAC configurations support the use of Universal Serial Bus (USB) communications.
Initial implementations used 12Mbit/sec USB1.1 links; newer implementations use the 480Mbit/sec
USB2.0 links. Remember that USB1.1 devices can be used on 100 Mbit/sec buses.
Ethernet
Several Turbo PMAC configurations support the use of Ethernet communications. Initial
implementations used 10Mbit/sec Ethernet links; there are plans to upgrade to 100Mbit/sec. Remember
that 10Mbit/sec devices can be used on 100Mbit/sec networks.
Both TCP/IP and UDP/IP protocols are supported. The simpler UDP/IP protocol is strongly
recommended, as it is more efficient, and the additional packet information and handshaking
acknowledgments of TCP/IP are not required for the point-to-point communications used here. Use of
the TCP/IP protocol with a computer running a Microsoft Windows operating system is particularly
inefficient, due to an operating system time-out on every command/response event.
Dual-Ported RAM
The bus communications ports, both backplane and wire, provide optional support for communications
through a shared-memory interface using dual-ported RAM. ASCII text commands can be issued through
the DPRAM, but the main virtue of DPRAM is its binary data structures that permit direct parallel access
to many pieces of data.
These structures include data-reporting buffers, in which Turbo PMAC repeatedly updates fixed registers
in DPRAM with commonly desired position information, such as motor positions; data gathering buffers,
permitting the real-time uploading of synchronously gathered data; background variable data copying
buffers, which permit the repeated copying to and from user-specified Turbo PMAC registers; binary
rotary motion-program download buffers for the fastest possible downloading of motion program files;
and user-defined structures.
For the ISA and VME buses, Turbo PMAC variables I93 and I94 set the address of the DPRAM on the
bus at power-up/reset. For the PCI bus, the host computer’s operating system sets this address
automatically. For the wire buses, the DPRAM does not directly map into the host computer’s memory
space; software routines such as Delta Tau’s PComm32 library create a virtual image of the DPRAM in
the host computer’s memory, making the remoteness of the DPRAM transparent to the applications
program.
Several communications structures in DPRAM can individually be enabled with I-variables:
• DPRAM ASCII Communications: I58 and I56
• DPRAM Motor Data Foreground Reporting: I48 and I47
• DPRAM Motor Data Background Reporting: I57 and I50
44
Talking to Turbo PMAC
Turbo PMAC User Manual
•
•
•
DPRAM Coordinate System and Global Background Data Reporting: I49 and I50
DPRAM Background Variable Buffers: I55
DPRAM Binary Rotary Buffer Foreground Transfer: I45
It is possible to have multiple DPRAM ICs in some Turbo PMAC systems, especially in UMAC systems.
However, only one of these ICs can be used at any given time with any of the automatic data structures.
Turbo PMAC variable I24 determines which of the DPRAM ICs will support these automatic functions
by specifying the base address of this IC in Turbo PMAC’s own address space. On re-initialization of the
Turbo PMAC, I24 is set to the address of the first DPRAM IC found (the one with the lowest base
address). I4904 shows which DPRAM ICs have been found by the Turbo PMAC CPU.
The DPRAM IC associated with the main bus communications port has a base address of $060000 in
Turbo PMAC’s memory map. DPRAM ICs that are accessed by the Turbo PMAC CPU over the JEXP or
UBUS expansion port, such as those on the Acc-54E USB/Ethernet communications card for the UMAC,
can have base addresses of $06C000, $06D000, $06E000, $06F000, $074000, $075000, $076000, or
$077000.
Giving Commands to Turbo PMAC
Turbo PMAC is fundamentally a command-driven device, unlike other controllers that are register driven.
Turbo PMAC can do things if ASCII command text strings are issued and generally Turbo PMAC
provides information to the host in ASCII text strings.
Note:
If the Option 2 dual-ported RAM is present, Turbo PMAC can be commanded by
writing values to specific registers in the DPRAM and Turbo PMAC can provide
information by placing binary values in these registers, but the ASCII commands
must be sent to Turbo PMAC that cause it to take the proper action when these
values are received and to place the values in these registers.
Turbo PMAC Processing of Commands
When Turbo PMAC receives an alphanumeric text character over one of its ports, it does nothing but
place the character in its command queue. It requires a control character (ASCII value 1 to 31) to cause it
to take some actual action. The most common control character used is the carriage return (<CR>; ASCII
value 13), which tells Turbo PMAC to interpret the preceding set of alphanumeric characters as a
command and to take the appropriate action.
Other control characters cause Turbo PMAC to take an action independent of the alphanumeric characters
sent before it. These control characters can be sent in the middle of a line of alphanumeric command
characters without disturbing the flow of the command. Turbo PMAC will respond first to the controlcharacter command, storing the text string until the <CR> character is received.
Command Acknowledgement
The exact nature of Turbo PMAC’s acknowledgement of commands and its data response is controlled by
I-variables I3, I4, and I9, with I3 as the most important. If I3 is 1, PMAC acknowledges a valid
alphanumeric command by sending the line-feed (<LF>; ASCII value 10) character back to the host. If
I3 is 2 or 3, it uses the <ACK> character (ASCII value 6) instead. If I3 is 0, it does not provide any
acknowledging character. Regardless of the setting of I3, Turbo PMAC responds to an invalid command
by returning the <BELL> character (ASCII value 7).
When working interactively with Turbo PMAC in a dumb terminal mode, it is often nice to use the <LF>
as acknowledgement because it spaces commands and responses on the terminal screen automatically.
Data Response
When the command received requires a data response, Turbo PMAC will precede each line of the data
response with a line feed character if I3 is set to 1 or 3. It will not do so if I3 is set to 0 or 2. Turbo
PMAC will terminate each line of the data response with a carriage-return character regardless of the
Talking to Turbo PMAC
45
Turbo PMAC User Manual
setting of I3. For these commands, the command acknowledgement character – <LF> or <ACK> – is sent
after the data response, serving as an end-of-transmission character. For computer parsing of the
response, it is nice to have the <ACK> serve as a unique EOT character.
Data Integrity
Variable I4 determines some of the data integrity checks Turbo PMAC performs on the communications,
the most important of which is a line-by-line checksum. The Writing a Host Communications Program
section covers this feature in detail.
Data Response Format
Variable I9 controls some aspects of how Turbo PMAC sends data to the host. Its setting determines
whether Turbo PMAC lists program lines back to the host in long or short form, whether it reports Ivariable values and M-variable definitions as full command statements or not, and whether address Ivariable values are reported in decimal or hexadecimal form.
On-Line (Immediate) Commands
Many of the commands given to Turbo PMAC are on-line commands; that is, they are executed
immediately by Turbo PMAC, to cause some action, change some variable, or report some information
back to the host. The command itself is thrown away after executing (so cannot be listed back), although
its effects may stay in Turbo PMAC.
Some commands, such as P1=1, are executed immediately if there is no open program buffer, but are
stored in the buffer if one is open. Other commands, such as X1000 Y1000, cannot be on-line
commands; there must be an open buffer – even if it is a special buffer for immediate execution. These
commands will be rejected by Turbo PMAC (reporting an ERR005 if I6 is set to 1 or 3) if there is no
buffer open. Still other commands, such as J+, are on-line commands only, and cannot be entered into a
program buffer (unless in the form of CMD"J+", for instance).
Types of On-Line Commands
There are four basic classes of on-line commands:
• Port-specific commands, which only affect the action of subsequent commands on the same port;
• Motor-specific commands, which affect only the motor that is currently addressed by the host;
• Coordinate-system-specific commands, which affect only the coordinate system that is currently
addressed by the host;
• Global commands, which affect the card regardless of any addressing modes. In the reference section
of the manual, each command is classified into one of these types under the Scope descriptor.
Note:
Each program that can use the COMMAND statement to issue on-line commands
from within the card has its own motor and coordinate system addressing,
independent of which motor and coordinate system the host is addressing.
Changing a port’s addressing modes does not affect the program’s mode, or vice
versa.
46
Talking to Turbo PMAC
Turbo PMAC User Manual
Port-Specific Commands
To maintain truly independent communications among the multiple communications ports, it is necessary
for certain commands that affect subsequent commands to only affect commands on the same port. For
this reason, addressing commands – #n for motors and &n for coordinate systems – as well as buffer
OPEN and CLOSE commands, affect only subsequent commands on the same port.
Motor-Specific Commands
A motor is addressed for a port by a #n command, where n is the number of the motor, with a range of 1
to 32, inclusive. This motor stays the one addressed by this port until another #n is received by the card
over the same port. For instance, the command line #1J+#2J- tells Motor 1 to jog in the positive
direction, and Motor 2 to jog in the negative direction (like most commands, the jog command does not
take effect until the carriage return character is received, so both axes start acting on the command at
roughly the same time in this case).
There are only a few types of motor-specific commands. These include the jogging commands, a homing
command, an open loop command, and requests for motor position, velocity, following error, and status.
Note:
An on-line motor action command, such as jogging, homing, open-loop output, is
not permitted if the addressed motor is in a coordinate system that is running a
motion program, even if the motion program is not directly commanding any axis
assigned to that motor. Such a command will be rejected with an error.
Coordinate-System-Specific Commands
A coordinate system is addressed for a port by an &n command, where n is the number of the coordinate
system, with a range of 1 to 16, inclusive. This coordinate system remains the one addressed until
another &n command is received by the card over the same port. For instance, the command line
&1B6R&2B8R tells Coordinate System 1 to run Motion Program 6 and Coordinate System 2 to run
Motion Program 8.
There are a variety of types of coordinate-system-specific commands. Axis definition statements act on
the addressed coordinate system, because motors are matched to an axis in a particular coordinate system.
Since it is a coordinate system that runs a motion control program, all program control commands act on
the addressed coordinate system. Q-variable assignment and query commands are also coordinate system
commands, because the Q-variables themselves belong to a coordinate system.
Note that a command to a coordinate system can affect several motors if more than one motor is assigned
to that coordinate system. For instance, if motor 4 is assigned to coordinate system 1, a command to
coordinate system 1 to run a motion program can start motor 4 moving.
Global Commands
Some on-line commands do not depend on which motor or coordinate system is addressed. For instance,
the command P1=1 sets the value of P1 to 1 regardless of what is addressed. Among these global on-line
commands are the buffer management commands. Turbo PMAC has multiple buffers, one of which can
be open at a time. When a buffer is open, commands can be entered into the buffer for later execution.
Control character commands (those with ASCII values 0 - 31D) are always global commands. Those that
do not require a data response act on all cards on a serial daisy-chain. These characters include carriage
return <CR>, backspace <BS>, and several special-purpose characters. This allows, for instance,
commands to be given to several locations on the card in a single line, and have them take effect
simultaneously at the <CR> at the end of the line (&1R&2R<CR> causes both Coordinate Systems 1 and 2
to run).
Talking to Turbo PMAC
47
Turbo PMAC User Manual
Buffered (Program) Commands
As their name implies, buffered commands are not acted on immediately, but held for later execution.
Sending a buffered program command to Turbo PMAC merely cause the command to be loaded into the
open program buffer; the command will not actually be executed until that program is run.
Turbo PMAC has many program buffers – 224 regular motion program buffers, 16 rotary motion
program buffers (1 for each coordinate system), 16 forward-kinematic subroutine buffers (1 for each
coordinate system), 16 inverse-kinematic subroutine buffers (1 for each coordinate system), and 32
uncompiled PLC program buffers. (Buffers for compiled or assembled programs – compiled PLCs, userwritten servo algorithms, and user-written phase algorithms – do not take text commands; instead they
take DSP machine code.)
Before commands can be entered into a buffer, that buffer must be opened (e.g. OPEN PROG 3, OPEN
PLC 7, &1 OPEN FORWARD). Note that buffered commands can only be entered into the buffer
through the same port that issued the OPEN command. Only one port at a time may have an open
program buffer.
Each program command is added onto the end of the list of commands in the open buffer; if you wish to
replace the existing buffer, use the CLEAR command immediately after opening to erase the existing
contents before entering the new ones. After finishing entering the program statements, use the CLOSE
command to close the opened buffer.
The rotary motion program buffer for a coordinate system is a special program buffer that can execute
motion programs at the same time it is open for entry of program commands from the host computer. If
an open rotary program buffer is executing, but has already executed every command sent to it, it will
execute the next buffered program command sent to it almost immediately.
Note that certain commands will be treated as on-line commands if no buffer is open on the port over
which it is sent, or as buffered program commands if a buffer is open on this port. For example, the
command P1=1 will be executed immediately if no buffer is open on the port, but it will be entered into a
program buffer for later execution as part of the program if a buffer is open on the port. Some of these
commands will have completely different meanings if a buffer is open on the port or if it is not. The
command M10 is a query for the value of variable M10 if no buffer is open on the port; it is a call to an
M-code subroutine if a motion program buffer is open on the port.
48
Talking to Turbo PMAC
Turbo PMAC User Manual
SETTING UP FEEDBACK AND MASTER POSITION SENSORS
Turbo PMAC systems can interface to a wide variety of position sensors for both feedback and master
use, either on the main control boards or through a variety of accessory boards. This section summarizes
the basic hardware and software setup issues; more details can be found in the appropriate hardware
reference manuals and the Turbo PMAC Software Reference Manual.
Note that the initial setup for feedback or master sensors is independent of any motor or coordinate
system. A motor or coordinate system may use the numerical value resulting from the initial hardware
and/or software processing of a position signal, but this is not required. It is also possible for user
programs or commands to access these position values directly, without a motor or coordinate-system
automatic function using them at all.
Setting Up Quadrature Encoders
Digital quadrature encoders are the most common position sensors used with Turbo PMACs. Interface
circuitry for these encoders comes standard on board-level Turbo PMAC controllers, UMAC axisinterface boards, and QMAC control boxes.
Signal Format
Quadrature encoders provide two digital signals that are a function of the position of the encoder, each
nominally with 50% duty cycle, and nominally one-quarter cycle apart. This format provides four distinct
states per cycle of the signal, or per line of the encoder. The phase difference of the two signals permits
the decoding electronics to discern the direction of travel, which would not be possible with a single
signal.
Typically, these signals are at 5V TTL/CMOS levels, whether single-ended or differential. The input
circuits are powered by the main 5V supply for the controller, but they can accept up to +/-12V between
the signals of each differential pair, and +/-12V between a signal and the GND voltage reference.
Differential encoder signals can enhance noise immunity by providing common-mode noise rejection.
Modern design standards virtually mandate their use for industrial systems, especially in the presence of
PWM power amplifiers, which generate a great deal of electromagnetic interference.
Hardware Setup
This section describes Turbo PMAC encoder hardware interface in general terms. Consult the Hardware
Reference Manual for your particular configuration for details.
Turbo PMAC’s encoder interface circuitry employs differential line receivers, but is configured at the
factory to accept either single-ended or differential encoders. In this configuration, the main (+) line is
pulled up to 5V, and the complementary (-) line is tied to 2.5V with a voltage divider. With a singleended encoder, the single signal line for each channel is then compared to this reference voltage as it
changes between 0 and 5V.
Setting Up Feedback and Master Position Sensors
49
Turbo PMAC User Manual
Note:
When using single-ended TTL-level digital encoders, the complementary line input
should be left open, not grounded or tied high; this is required for PMAC's
differential line receivers to work properly.
It is possible to pull the complementary line as well to 5V. On most Turbo PMAC boards, this is done by
changing the setting of a 3-point jumper for the encoder channel. On Turbo PMAC2 boards and systems
(including UMAC), this is done by reversing a SIP resistor pack in its socket.
PMAC Encoder Input Circuitry
y
y
y
y
y
y
y
Connect pin 1 to 2 to tie differential line to +2.5V
Connect pin 2 to 3 to tie differential line to +5V (Reversible socketed SIP on PMAC2)
Tie to +2.5V when no connection
Tie to +2.5V for single-ended encoders
Do not care for differential line driver encoders
Tie to +5V for complementary open-collector encoders (obsolete)
Tie to +5V to support external XOR loss of encoder circuitry
Encoder-Loss Detection Circuitry
The main reason to pull up the complementary line is to enable the encoder-loss detection circuitry that is
present on many Turbo PMAC and interface boards. This circuitry uses exclusive-or (XOR) logic that
requires that the signals of each pair be in opposite logical states for it to consider the encoder present. If
the encoder fails or the cable becomes disconnected, both inputs of the pair pull high to a logical 1, and
the circuitry indicates encoder loss. Check the appropriate hardware reference manual to see if your
device has encoder-loss detection and how it is implemented.
Termination Resistors
When driving the encoder signals over a long cable (10 meters or more), it may be desirable to add
termination resistors between the main and complementary lines to reduce the ringing on transitions.
PMAC provides sockets for resistor packs for this purpose. The optimum value of the termination
resistor is system dependent; ideally matching the characteristic impedance of the cable, but 330 ohms is
a good starting point.
50
Setting Up Feedback and Master Position Sensors
Turbo PMAC User Manual
Power Supply and Isolation
In the basic configuration of a Turbo PMAC, the encoder circuitry is not isolated from Turbo PMAC’s
digital circuitry and the signals are referenced to Turbo PMAC’s digital common level GND. Typically,
the encoders in this case are powered from PMAC’s +5V lines with a return on GND. The total encoder
current draw must be considered in sizing the Turbo PMAC power supply.
It is also possible to use a separate supply for the encoders with non-isolated signals connected to Turbo
PMAC. In this case, the return of the supply should be connected to the digital common GND on Turbo
PMAC to give the signals a common reference. The +5V lines of separate supplies should never be tied
together, as they will fight each other to control the exact voltage level.
Isolated Encoder Signals
In some systems, it is desired to optically isolate the encoder circuitry from PMAC’s digital circuitry.
This is common in systems with long distances from the encoder to the controller (> 10m or 30 ft) and/or
systems with very high levels of electrical noise. Isolation can be achieved using the Acc-8D Opt 6 4channel encoder isolator board. With an isolated encoder, a separate power supply is required for the
encoders to maintain isolation, and the return on the supply must not be connected to the digital common
GND, or the isolation will be defeated.
Simulated Encoder Signals
Special consideration must be given to systems that have a simulated encoder signal provided from a
resolver-to-digital converter in a brushless motor amplifier. In these systems, the encoder signals are
referenced to the amplifier's signal return, which in turn is connected to Turbo PMAC’s analog common
AGND. The best setup in these cases is to isolate the simulated encoder signal from the PMAC digital
circuitry with the Acc-8D Opt 6 isolator board or similar module. This will keep full isolation between
the Turbo PMAC digital circuitry and the amplifier.
If isolation of the simulated encoder signals is not feasible, Turbo PMAC’s digital circuitry and the
amplifier signal circuitry (including Turbo PMAC’s analog circuitry) must be well tied together to
provide a common reference voltage. Do this by putting jumpers on the Turbo PMAC or interface board
(E-Points E85, E87, and E88 on many boards), tying the digital and analog circuits on Turbo PMAC
together, and therefore also the analog signal circuits. What must be avoided is having the simulated
encoder cables providing the only connection between the circuits. This can result in lost signals from
bad referencing, or even component damage from ground loops.
Wiring Techniques
There are several important techniques in the wiring of the encoders that are important for noise
mitigation. First, the encoder cable should be kept physically separate from the motor power cable if at
all possible. Second, both of these cables should be shielded, the motor cable to prevent noise from
getting out, and the encoder cable to prevent noise from getting in. These shields should be grounded at
the inward end only, that is, to the device that is itself tied to a ground.
A third important noise mitigation technique is to twist the leads of the complementary pairs around each
other. With these twisted pairs, what noise does get in tends to cancel itself out in opposite halves of the
twist.
Turbo PMAC Hardware-Control Parameter Setup
The Turbo PMAC ASICs are set up by default to accept quadrature feedback, but some settings may need
to be tweaked to optimize operation.
Setting Up Feedback and Master Position Sensors
51
Turbo PMAC User Manual
Encoder Sampling Clock Frequency: E34 – E38, I7m03, I6803, MI903, MI907
After the front-end processing through the differential line receivers, the quadrature encoder inputs are
sampled by digital logic in the Turbo PMAC Servo IC or MACRO IC at a rate determined by the
frequency of the SCLK encoder sample clock” which is user settable. The higher the SCLK frequency,
the higher the maximum permissible count rate; the lower the SCLK frequency, the more effective the
digital delay noise filter is.
Each encoder input channel has a digital delay filter consisting of three cascaded D-flip-flops on each
line, with a best two-of-three voting scheme on the outputs of the flip-flops. The flip-flops are clocked by
the SCLK signal. This filter does not pass through a state change that only lasts for one SCLK cycle; any
change this narrow should be a noise spike. In doing this, the filter delays actual transitions by two SCLK
cycles – a trivial delay in virtually all systems.
If both the A and B channels change state at the decode circuitry (post-filter) in the same SCLK cycle, an
unrecoverable error to the counter value will result. The ASIC hardware notes this problem by setting
and latching the “encoder count error” bit in the channel’s status word. The problem can also be detected
by capturing the count value each revolution on the index pulse and seeing whether the correct number of
counts have elapsed.
Encoder Digital Delay Filter
The SLCK frequency must be at least four times higher than the maximum encoder cycle (line) frequency
input, regardless of the quadrature decoding method used (with the most common times-4 decode, the
SCLK frequency must be at least as high as the count rate). In actual use, due to imperfections in the
input signals, a 20 – 25% safety margin should be used.
The default SCLK frequency of 9.83 MHz is acceptable for virtually all applications. It can accept
encoder signal cycle frequencies of up to about 2 MHz (8 MHz count rates) – with safety margin – and
still provide decent digital filtering. This frequency may be changed by factors of two, up to 19.66 or
39.32 MHz on most designs, or down to as low as 1.25 MHz or 306 kHz on most designs. On some
designs, an external SCLK signal can be provided. If very high encoder count rates are required, the
SCLK frequency may have to be raised; if better filtering is required to prevent count errors, the SCLK
frequency may have to be lowered.
52
Setting Up Feedback and Master Position Sensors
Turbo PMAC User Manual
Turbo PMAC Servo IC SCLK Frequency Control
On a Turbo PMAC board, or an Acc-24x with PMAC-style Servo ICs (Acc-24P or 24V), the SCLK
frequency is set by the configuration of jumpers E34 – E38. Only one of these jumpers may be on in any
given configuration. This SCLK frequency is common to all Servo ICs and channels on the board.
Turbo PMAC2 Servo IC SCLK Frequency Control
On a Turbo PMAC2 board, or an accessory board with PMAC2-style Servo ICs or MACRO ICs (e.g.
Acc-24P2, 24E2, 24E2A, 24E2S, 24C2, 24C2A, 5E), the SCLK frequency is set by an I-variable for the
Servo IC. This variable sets the frequency for all channels in the IC, but each IC has an independent
setting. Consult the description of the variable in the Software Reference manual for details.
For a PMAC2-style Servo IC m, IC variable I7m03 determines the SCLK frequency. This variable also
determines three other clock frequencies for the IC. If the IC is on a MACRO Station, MI903 or MI907
determines the SCLK frequency.
For a MACRO IC on a Turbo PMAC2 board or UMAC Acc-5E board (for the handwheel encoders),
I6803 determines the SCLK frequency in the same manner for both channels on the IC. If the IC is in a
MACRO Station with an Acc-1E stack 2-axis board, MI993 determines the SCLK frequency.
All of these variables work in the same manner. The other three clock frequencies that each controls are
virtually never changed, so the following table may be useful for setting the SCLK frequency with the
others left at the default frequency:
Variable Value
2256
2257
2258*
2259
*Default
SCLK Frequency
Variable Value
SCLK Frequency
39.32 MHz
19.66 MHz
9.83 MHz*
4.92 MHz
2260
2261
2262
2263
2.46 MHz
1.23 MHz
612 kHz
306 kHz
Encoder Decode Control: I7mn0, I68n0, MI910
The decoding of the encoder signal, both as to resolution and direction, is determined by a channelspecific I-variable. For Servo ICs of both PMAC-style and PMAC2-style, this is I7mn0 (for Servo IC m
Channel n). For MACRO IC 0 Channel n* on a Turbo PMAC (a “handwheel” port encoder), this is
I68n0. For encoder channels on a MACRO Station, this is node-specific variable MI910.
This variable is set for “times-4” decode, which derives four counts per signal cycle, one for each signal
edge. This requires a variable value of 3 or 7. The difference between these two values is the direction
sense – which direction of motion causes the counter to count up. Remember that for a feedback encoder,
the encoder’s direction sense must match the servo-loop output’s direction sense – a positive servo output
must cause the counter to count in the positive direction – otherwise a dangerous runaway condition will
occur when the servo loop is closed.
Conversion Table Processing Setup – Turbo PMAC Interface
Digital quadrature encoders are processed in the conversion table with the “1/T extension” method
(method digit $0), which uses timers associated with the counter to compute fractional count information
that enhances smoothness of motion. The source address specified is that of the base address of the
channel (e.g. $78200 for Servo IC 0 Channel 1); Turbo PMAC will use several registers of that channel to
assemble the enhanced position information automatically.
It is also possible to disable the 1/T extension (method digit $C) in the table processing. For details of
setting up the encoder conversion table to process quadrature encoders, consult the Setting up the Encoder
Conversion Table section and the specification for variables I8000 – I8191 in the Software Reference
Manual.
Setting Up Feedback and Master Position Sensors
53
Turbo PMAC User Manual
Conversion Table Processing Setup – MACRO Station Interface
If the quadrature encoder is connected to a remote MACRO Station, the conversion table processing
(usually 1/T extension) is done in the MACRO Station, and the resulting enhanced position information is
passed back to the Turbo PMAC, where it is processed by the Turbo PMAC’s encoder conversion table as
unshifted parallel data (usually method digit $2, mode switch bit = 1), because it already has fractional
count information, before use by the servo. For details of setting up the encoder conversion table to
process quadrature encoders, consult the MACRO Station manuals, the section Setting up the Encoder
Conversion Table in the Turbo PMAC User Manual, and the specification for variables I8000 – I8191 in
the Software Reference Manual.
Scaling the Feedback Units
The decoding scheme you select in the Servo IC or MACRO IC determines what a count is. For example,
if selecting times-4 decode, a count is defined as ¼ of an encoder line. If the encoder has 2000 lines per
revolution, a count is defined as 1/8000 of a revolution. All subsequent position, velocity, and
acceleration units are based on this definition of a count.
Setting Up Digital Hall Sensors
Three-phase digital hall-effect position sensors (or their equivalent) are popular for commutation
feedback. They can also be used with Turbo PMAC as low-resolution position/velocity sensors. As
commutation position sensors, typically they are just used by Turbo PMAC for approximate power-up
phase position; typically, ongoing phase position is derived from the same high-resolution encoder that is
used for servo feedback. (Many controllers and amplifiers use these hall sensors as their only
commutation position feedback, starting and ongoing, but that is a lower-performance technique.
Many optical encoders have hall tracks. These commutation tracks provide signal outputs equivalent to
those of magnetic hall commutation sensors, but use optical means to create the signals.
Note:
These digital hall-effect position sensors should not be confused with analog halleffect current sensors used in many amplifiers to provide current feedback data for
the current loop.
Signal Format
Digital hall sensors provide three digital signals that are a function of the position of the motor, each
nominally with 50% duty cycle, and nominally one-third cycle apart. (This format is often called 120o
spacing. Turbo PMAC has no automatic hardware or software features to work with 60o spacing.) This
format provides six distinct states per cycle of the signal. Typically, one cycle of the signal set
corresponds to one electrical cycle, or pole pair, of the motor. These sensors, then, can provide absolute
(if low resolution) information about where the motor is in its commutation cycle, and eliminate the need
to do a power-on phasing search operation.
54
Setting Up Feedback and Master Position Sensors
Turbo PMAC User Manual
Hardware Setup
If just used for power-up commutation position feedback, typically the hall sensors are wired into the U,
V, and W supplemental flags for a PMAC2-style interface channel. These are single-ended 5V digital
inputs on all existing hardware implementations. They are not optically isolated inputs; if isolation is
desired from the sensor, this must be done externally.
Note:
In the case of magnetic hall sensors, the feedback signals often come back to the
controller in the same cable as the motor power leads. In this case, the possibility
of a short to motor power must be considered; safety considerations and industrial
design codes may make it impermissible to connect the signals directly to the
Turbo PMAC TTL inputs without isolation.
If used for servo position and velocity feedback (PMAC2-style Servo ICs only), the three hall sensors are
connected to the A, B, and C encoder inputs, so that the signal edges can be counted. As with quadrature
encoders, these inputs can be single-ended or differential. They are not optically isolated inputs; if
isolation is desired from the sensor, this must be done externally. There may be applications in which the
signals are connected both to U, V, and W inputs (for power-on commutation position) and to A, B, and C
inputs (for servo feedback).
Turbo PMAC Hardware-Control Parameter Setup
Hall Sensor Demux Control: I7mn5, I68n5, MI915
If the hall sensors are connected to the channel’s U, V, and W inputs, you must make sure that bit 1 of the
channel’s Hall Sensor Demux Control variable is set to the default of 0. If this bit is set to 1, the
information in the U, V, and W bits of the channel’s status register is de-multiplexed from the C-channel
of the encoder input based on the states of the A and B inputs, as with Yaskawa incremental encoders.
This variable is I7mn5 for Servo ICs, I68n5, for MACRO ICs, and node-specific variable MI915 on a
MACRO Station.
Encoder Decode Control: I7mn0, I68n0, MI910
If the hall sensors are wired into the encoder inputs A, B, and C, the decoding of the signal is determined
by a channel-specific I-variable. For (PMAC2-style) Servo ICs, this is I7mn0 (for Servo IC m Channel
n). For MACRO IC 0 Channel n* on a Turbo PMAC (a handwheel port encoder), this is I68n0. For
encoder channels on a MACRO Station, this is node-specific variable MI910.
Setting Up Feedback and Master Position Sensors
55
Turbo PMAC User Manual
For the 3-phase hall sensors, the decode must be set to times-6 decode, which derives six counts per
signal cycle, one for each signal edge. This requires a variable value of 11 or 15. The difference between
these two values is the direction sense – which direction of motion causes the counter to count up.
Remember that for a feedback sensor, the sensor’s direction sense must match the servo-loop output’s
direction sense – a positive servo output must cause the counter to count in the positive direction –
otherwise a dangerous runaway condition will occur when the servo loop is closed.
Power-Up Phasing Usage
The most common use of these hall sensors with a Turbo PMAC2 is to establish an approximate absolute
position for the phase commutation algorithms. In this case, motor variable Ixx81 contains the address of
the flag register for these U, V, and W inputs. Ixx91 must be set to a value from $800000 to $FF0000;
the actual value is dependent on the direction and phasing of the sensors relative to the motor
commutation cycle.
Conversion Table Processing Setup – Turbo PMAC Interface
If the hall sensors are connected to the encoder inputs on Turbo PMAC and therefore to the encoder
counter, the count information should be processed in the conversion table with the 1/T-extension method
(method digit $0) as for a quadrature encoder. The sub-count estimation of this method is particularly
important here because of the low resolution of the sensors. For details of setting up the encoder
conversion table to process hall sensors used as encoders, consult the section Setting up the Encoder
Conversion Table section in this manual and the specification for variables I8000 – I8191 in the Software
Reference Manual.
Note:
If the hall sensors are used for ongoing commutation feedback, probably it will be
better to use the extended result data from the conversion table instead of the raw
counter information that usually is used with encoders. This will provide smoother
commutation. In this case, there will be 6 * 32 = 192 counts per commutation
cycle instead of just 6.
Conversion Table Processing Setup – MACRO Station Interface
If the hall sensors are connected to the encoder inputs on a MACRO Station and therefore to the encoder
counter, the count information should be processed in the Station’s conversion table with the 1/Textension method (method digit $0), just as for a quadrature encoder. The sub-count estimation of this
method is particularly important here because of the low resolution of the sensors. The resulting
enhanced position information is passed back to the Turbo PMAC, where it is processed by the Turbo
PMAC’s encoder conversion table as unshifted parallel data (usually method digit $2, mode switch bit =
1), because it already has fractional count information, before use by the Turbo PMAC servo and/or
commutation algorithms.
For details of setting up the encoder conversion table to process hall sensors used as encoders, consult the
MACRO Station manuals, the section Setting up the Encoder Conversion Table section in this manual
and the specification for variables I8000 – I8191 in the Software Reference Manual.
Scaling the Feedback Units
For purposes of the motor’s servo loop and associated functions, the hall sensors will provide six counts
per electrical cycle; a 4-pole motor would therefore have 12 counts per revolution. If using the 1/Textended count data for the commutation, the commutation cycle size (Ixx71/Ixx70) will be 192 counts.
56
Setting Up Feedback and Master Position Sensors
Turbo PMAC User Manual
Setting up Sinusoidal Encoders
Turbo PMAC systems can accept signals from sinusoidal (sine/cosine) encoders and generate position
data of very high resolution. This is done through special accessory boards and dedicated firmware
algorithms. Presently there are two classes of interpolators: low-resolution ones producing 128 or 256
states per line, and high-resolution ones producing 4096 states per line.
Note:
Many Turbo PMAC users will utilize sinusoidal encoders for position feedback,
but with the interpolation function performed in an external interpolator module,
which generates a high-frequency synthesized digital quadrature signal for the
controller that is connected to the Turbo PMAC. In this case, the Turbo PMAC
treats the feedback just as if a real digital quadrature encoder were used.
Hardware Setup
The sine and cosine channels of the encoder are connected either as single-ended or differential inputs
(differential is strongly recommended) into the interpolator accessory according to the instructions for the
interpolator. Optionally, the index channel may be connected as well. Consult the hardware reference
manual of the particular accessory for details.
If connecting a low-resolution interpolator into a Turbo PMAC or accessory with PMAC-style Servo ICs,
the fractional-count data is brought into the Servo IC at TTL levels on the flag signals for the channel
numbered one higher than that of the encoder signal itself. Normally, these signals are 12-24V isolated
flags to use the interpolator, the isolators must be removed and replaced with conducting shunts.
Turbo PMAC Hardware-Control Parameter Setup
Encoder Sampling Clock Frequency: E34 – E38, I7m03, MI903, MI907
For the low-resolution interpolator, the Turbo PMAC’s SCLK encoder-sampling clock drives the analog
conversion on the interpolator, as well as the encoder functions in the Servo IC. This clock must be set to
a frequency of 2.46 MHz. If the interpolator is connected to a Turbo PMAC or accessory with PMACstyle Servo ICs, this frequency is set by jumpers on the PMAC or accessory: of the E-points E34 – E38,
there should only be a jumper on E36.
If the low-resolution interpolator is connected to a Turbo PMAC2 or accessory with PMAC2-style Servo
ICs, Servo IC m’s variable I7m03 sets this frequency. If it is connected into a MACRO Station, Station
variable MI903 or MI907 sets this frequency. These variables also set other clock frequencies, but if the
other frequencies are left at their default (as they usually are), setting the SCLK frequency to 2.46 MHz
simply requires changing the variable value from its default of 2258 to 2260.
For an Acc-51x high-resolution interpolator, the SCLK signal just drives the sampling of the synthesized
quadrature in the interpolator’s own Servo IC. The default frequency of 9.83 MHz is virtually always
fine. On the Acc-51P, the frequency is set permanently to this value. On high-resolution interpolators
with PMAC2-style Servo ICs (e.g. Acc-51E, Acc-51C), IC variable I7m03 sets this frequency if the
interpolator is connected directly to a Turbo PMAC, or by MACRO Station variable MI903 or MI907 if
the interpolator is installed in a MACRO Station.
Encoder Decode Control: I7mn0, MI910
Both styles of interpolator create a digital quadrature signal from the sine/cosine inputs. This digital
signal goes into the encoder decoding and count circuitry to produce the whole-count data. To match the
whole-count and fractional count resolution properly, the decode-control variable must be set up to times4 quadrature decode. This variable is I7mn0 for Servo IC m Channel n in a Turbo PMAC system. On a
MACRO Station, it is node-specific variable MI910.
Setting Up Feedback and Master Position Sensors
57
Turbo PMAC User Manual
In addition, the direction sense of the whole-count data must match that of the fractional-count data. For
the low-resolution interpolators, this requires that the decode variable be set to 7. If this value does not
produce the direction sense needed, change the wiring of the encoder into the accessory (exchange sine
and cosine signals, or the plus and minus lines of one channel).
For the high-resolution interpolators, the conversion table will check the direction sense of the encoder
decode at power-up/reset and adjust the direction sense of the fractional data accordingly. This permits
the setting of the direction sense as needed by setting the decode variable to 3 or 7, but if this variable
setting is changed, save the setting and reset the card before the fractional direction sense matches.
Encoder Filter Bypass: I7mn1 (PMAC-style ICs)
If a low-resolution interpolator is connected to a PMAC-style Servo IC, the encoder’s digital delay filter
must be bypassed for both the channel with the encoder signal, and the channel with the flags, in order to
keep the whole-count and fractional-count data properly synchronized. The same is true if using the Acc51P high-resolution interpolator with its own PMAC-style Servo IC on board. Bypassing the delay filter
is selected by setting channel variable I7mn1 to 1.
Conversion Table Processing Setup – Turbo PMAC Interface
Sinusoidal encoders wired into a high-resolution (x4096) Acc-51x interpolator connected to a Turbo
PMAC are processed in the Turbo PMAC’s conversion table by a hi-res interpolator entry (method digit
$F), in which the fractional-count data is computed mathematically from the readings of the sine and
cosine A/D converters. The least significant bit (LSB) of the result represents 1/4096 of an encoder line.
Encoder Interpolation: High Resolution
Analog “SIN / COS” Encoders
A
Decoder/
Counter
+
A
Comparator
“1 bit A/D”
B
Address
Data
n-bit
A/D
+
-
n-bit
A/D
Differential
Amplifier
B
Analog
Photo
Current
SIN / COS
Signals
Encoder
Controller
Sinusoidal encoders wired into a low-resolution (x128/x256) interpolator connected to a Turbo PMAC are
processed in the Turbo PMACs conversion table by a parallel extension of incremental encoder entry
(method digit $C), in which the fractional-count data has been computed externally in the interpolator’s
look-up table and is just appended to the whole count data by the conversion table. The LSB of the result
represents 1/128 or 1/256 of an encoder line, depending on the setting of the resolution jumper on the
interpolator.
58
Setting Up Feedback and Master Position Sensors
Turbo PMAC User Manual
For details of setting up the encoder conversion table to process sinusoidal encoders, consult the section
Setting up the Encoder Conversion Table section in this manual and the specification for variables I8000
– I8191 in the Software Reference Manual.
Encoder Interpolation: Analog SIN/COS Encoders
A
n
+
n-bit
A/D
A
B
n
+
0
Differential
Amplifier
SIN / COS
Signals
Analog
Photo
Current
n-1, n bits
n-bit
A/D
B
2n - 1
0
2n - 1
Encoder
Flash
SIN / COS
A/D
Numbers
Converters
Arctan
Look-Up
Table
Angle
Number
Controller
Conversion Table Processing Setup – MACRO Station Interface
Sinusoidal encoders wired into a high-resolution (x4096) Acc-51x interpolator connected to a MACRO
Station CPU are processed in the MACRO Station CPU’s conversion table by a hi-res interpolator entry
(method digit $F), in which the fractional-count data is computed mathematically from the readings of the
sine and cosine A/D converters. The least significant bit (LSB) of the result represents 1/4096 of an
encoder line. This data is then passed over the MACRO ring to the Turbo PMAC, where it is processed
as unshifted parallel data (method digit $2).
Sinusoidal encoders wired into a “low-resolution (x128/x256) interpolator connected to a MACRO
Station CPU are processed in the MACRO Station CPU’s conversion table by a parallel extension of
incremental encoder entry (method digit $C), in which the fractional-count data has been computed
externally in the interpolator’s look-up table and is just appended to the whole count data by the
conversion table. The LSB of the result represents 1/128 or 1/256 of an encoder line, depending on the
setting of the resolution jumper on the interpolator. This data is then passed over the MACRO ring to the
Turbo PMAC, where it is processed as unshifted parallel data (method digit $2).
For details of setting up the encoder conversion table to process sinusoidal encoders, consult the MACRO
Station manuals and the Setting up the Encoder Conversion Table section in this manual and the
specification for variables I8000 – I8191 in the Software Reference Manual.
Scaling the Feedback Units
Turbo PMAC automatically considers the data it reads for the servo with Ixx03, Ixx04, and Ixx05 to be in
units of 1/32 of a count (i.e. to have 5 bits of fractional count resolution). With data processed in the
conversion table from a sinusoidal-encoder interpolator, be very careful to understand what a count is.
Setting Up Feedback and Master Position Sensors
59
Turbo PMAC User Manual
High-Resolution (x4096) Interpolators
For the high-resolution interpolators, the interpolation algorithm produces data with the LSB representing
1/4096 of an encoder line. Since the Turbo PMAC motor software considers this data to be in units of
1/32 of a count, this means that the software considers there to be 128 counts per encoder line. Other
motor and axis position, velocity, and acceleration values are based on this definition of a count which we
will call a software count.
Note that the hardware counter in the Servo IC of the high-resolution interpolator only contains four
counts per line of the encoder. If using the encoder for Turbo PMAC-based commutation, use the
hardware counter register as the commutation position source, not the interpolated result, and this
hardware count resolution should be known.
Low-Resolution (x128/x256) Interpolators
For the low-resolution interpolators, the interpolation algorithm produces data with the LSB representing
1/128 or 1/256 of an encoder line, and the hardware counter for the channel has data with the LSB
representing 1/4 or 1/8 of an encoder line, depending on the setting of the resolution jumper for the
interpolator. Since the Turbo PMAC motor software considers this data to be in units of 1/32 of a count,
the software considers there to be four or eight counts per encoder line, matching the resolution of the
encoder counter. However, compared to 1/T timer-based sub-count estimation, this method produces real
fractional-count position data (i.e. command and hold final position to resolution less than a count).
Setting up Resolvers
A Turbo PMAC system can interface directly to resolvers through an Acc-8D Option 7 resolver-to-digital
(R/D) converter board. This board can create the AC excitation signal for up to four resolvers, accept the
modulated sine and cosine signals back from these resolvers, demodulate the signals and derive the
position of the resolver from the resulting information, in an absolute sense if necessary.
Note:
Many Turbo PMAC users will utilize resolvers for position feedback, but with the
resolver-to-digital conversion performed in the servo amplifier, which uses the
resulting position information for its commutation functions. This style of
amplifier generates a synthesized digital quadrature signal for the controller, which
is connected to the Turbo PMAC. In this case, the Turbo PMAC treats the
feedback just as if a real digital quadrature encoder were used.
Hardware Setup
The details of the hardware setup are covered in the hardware reference manual for the R/D converter
board. Fundamentally, the R/D board connects three differential analog signal pairs to each resolver: a
single excitation signal pair, and two feedback signal pairs. It has two different digital connections to the
Turbo PMAC: one for serial absolute position to the JTHW Multiplexer port (common for all channels on
the board, optional in use), and one with synthesized digital quadrature for each channel, connected to a
normal quadrature encoder interface.
Turbo PMAC Hardware-Control Parameter Setup
Encoder Decode Control: I7mn0, I68n0, MI910
To match the resolution and direction sense of the ongoing position information derived from the
synthesized quadrature with absolute position read through the Multiplexer port, set the channel’s encoder
decode variable to 7 (times-4 decode, counterclockwise). If not using absolute position, set the decode
resolution and direction as wanted. For Servo ICs of both PMAC-style and PMAC2-style, this is I7mn0
(for Servo IC m Channel n). For MACRO IC 0 Channel n* on a Turbo PMAC (a handwheel port
encoder), this is I68n0. For encoder channels on a MACRO Station, this is node-specific variable MI910.
60
Setting Up Feedback and Master Position Sensors
Turbo PMAC User Manual
Conversion Table Processing Setup – Turbo PMAC Interface
Usually, digital quadrature signals, even if synthesized, are processed in the conversion table with the 1/T
extension method (method digit $0), which uses timers associated with the counter to compute fractional
count information that enhances smoothness of motion. The source address specified is that of the base
address of the channel (e.g. $78200 for Servo IC 0 Channel 1); Turbo PMAC will use several registers of
that channel to assemble the enhanced position information automatically.
For details of setting up the encoder conversion table to process synthesized quadrature encoder signals,
consult the section Setting up the Encoder Conversion Table section in this manual and the specification
for variables I8000 – I8191 in the Software Reference Manual.
Conversion Table Processing Setup – MACRO Station Interface
If the synthesized quadrature signal is connected to a remote MACRO Station, the conversion table
processing (usually 1/T extension) is done in the MACRO Station, and the resulting enhanced position
information is passed back to the Turbo PMAC, where it is processed by the Turbo PMAC’s encoder
conversion table as unshifted parallel data (usually method digit $2, mode switch bit = 1), because it
already has fractional count information, before use by the servo. For details of setting up the encoder
conversion table to process synthesized quadrature encoder signals, consult the MACRO Station manuals,
the Setting up the Encoder Conversion Table section in this manual, and the specification for variables
I8000 – I8191 in the Software Reference Manual.
Setting Up for Power-On Absolute Position
It is possible to get absolute position directly from the Acc-8D Option 7 through the multiplexer port.
Most commonly, this is just the absolute position within one motor revolution or even one commutation
cycle to establish the commutation phase reference position without any motion. This section
summarizes the variable settings for this technique; refer to the Software Reference Manual for details.
Absolute Phase Power-On Position Address and Format: Ixx81, Ixx91, MI11x
To read an R/D converter for absolute phase position from a board directly connected to the Turbo
PMAC, Ixx81 must be set to the address of the converter board on the multiplexer port (not directly in
Turbo PMAC’s address space). Bits 16 – 23 of Ixx91 must be set to the location of the converter on the
board at that address.
If the R/D board is connected to a MACRO Station, Ixx81 is set to the MACRO node number, and Ixx91
is set to $730000. On the MACRO Station, MI11x is set to specify the address of the converter board on
the multiplexer port in bits 0 – 7, and the location of the converter on the board in bits 16 – 23.
Motor phase offset variable Ixx75 contains the difference between the absolute resolver position and the
resulting phase angle position (if any).
Absolute Servo Power-On Position Address and Format: Ixx10, Ixx95, MI11x
To read one, two, or three R/D converters for absolute phase position from a board directly connected to
the Turbo PMAC, Ixx81 must be set to the address of the first converter board on the multiplexer port
(not directly in Turbo PMAC’s address space). Bits 16 – 23 of Ixx91 must be set to the location of the
first (fine) converter used on the board at that address.
If the R/D board is connected to a MACRO Station, Ixx81 is set to the MACRO node number, and Ixx91
is set to $730000. On the MACRO Station, MI11x is set to specify the address of the converter board on
the multiplexer port in bits 0 – 7, and the location of the converter on the board in bits 16 – 23.
If a second resolver, geared down from the first by an integer factor of n, is used to get multi-turn absolute
position, motor variable Ixx99 must be set to this value of n. If a third resolver, geared down from the
second by an integer factor of m, is used to extend the range of the multi-turn absolute position, motor
variable Ixx98 must be set to this value of m.
Setting Up Feedback and Master Position Sensors
61
Turbo PMAC User Manual
Motor offset variable Ixx26 contains the difference between the absolute resolver position and the
resulting motor position (if any).
Scaling the Feedback Units
The Acc-8D Option 7 R/D converter is a 12-bit converter. It reports 4096 separate states per electrical
cycle of the resolver (per mechanical revolution for a typical 2-pole resolver, per half revolution for a 4pole resolver). It reports this as a 12-bit binary number if queried for absolute position over the
multiplexer port. For ongoing position, it generates 1024 digital quadrature cycles per electrical cycle,
which creates 4096 counts per cycle after times-4 decode.
Setting Up MLDTs
A Turbo PMAC with PMAC2-style ICs can provide direct interface to magnetostrictive linear
displacement transducers (MLDTs), such as MTS’s Temposonics brand. MLDTs can provide absolute
position information in rugged environments; they are particularly well suited to hydraulic applications.
In this interface Turbo PMAC provides a periodic excitation pulse output to the MLDT, receives the echo
pulse that returns at the speed of sound in the transducer, and very accurately measures the time between
these pulses, which is directly proportional to the distance of the moving member from the stationary base
of the transducer. The timer therefore contains a position measurement.
MLDT Interface Type
MLDTs are available with several different interface formats; for this interface, a format with external
excitation is required, because Turbo PMAC provides the excitation pulse. Usually, this format has an
RS-422 interface, because the excitation and echo pulses are at RS-422 levels. The Turbo PMAC MLDT
interface inputs and outputs are at RS-422 levels.
Some MLDTs come with internal excitation and computation of position, providing a position value in a
format such as Synchronous Serial Interface (SSI) or an analog voltage. In these cases, setting up the
Turbo PMAC interface is dependent only on the data format, not on the underlying principle of the
sensor. Refer to the appropriate feedback-format section for details.
Signal Formats
There are two common signal formats of the external excitation type; MTS calls them RPM and DPM. In
the RPM format there are two short pulses returned from the MLDT: an immediate start pulse, and a
delayed stop pulse.
RPM Signal Format
Pulse Out
Return
Start
Stop
Since Turbo PMAC uses the first rising signal edge returned after the falling edge of the output pulse to
latch the timer, the key setup issue in this format is to make sure that the output pulse width is large
enough so that the falling edge of the output pulse occurs after the rising edge of the return line’s start
pulse (see PFM Pulse Width below).
In the DPM format, there is only one long pulse returned from the MLDT.
DPM Signal Format
Pulse Out
Return
Start
62
Stop
Setting Up Feedback and Master Position Sensors
Turbo PMAC User Manual
The rising edge of the return pulse in the DPM format is the equivalent of the rising edge of the start pulse
in the RPM format. The falling edge of the return pulse in the DPM format is the equivalent to the rising
edge of the stop pulse in the RPM format. Because Turbo PMAC is expecting a rising signal edge to
latch the timer, in this signal format the return signals should be inverted so that the ‘+’ output of the
MLDT is wired into Turbo PMAC’s ‘-’ input, and vice versa.
Hardware Setup
The PULSEn output that is used commonly to command stepper drives is used as the excitation signal for
the MLDT; typically the CHAn input that is part of encoder feedback is used to accept the response. The
PULSEn output is an RS-422 style differential line-drive pair. The CHAn input is an RS-422 style
differential line receiver pair. The use of differential pairs for both inputs and outputs is strongly
encouraged for the common-mode noise rejection it provides.
On some interface boards (e.g. Acc-24E2A, Acc-24E2S), the PULSEn+/- signals are output on lines that
otherwise would be supplemental flag inputs, and jumper(s) must be installed to enable the outputs on
these lines. Consult the user manual for the board for details.
Remember that in the DPM signal format or equivalent (see above), the ‘+’ output of the MLDT should
be wired into the CHAn- input, and the ‘-’ output of the MLDT should be wired into the CHAn+ input.
Turbo PMAC Hardware-Control Parameter Setup
PFM Clock Frequency: I7m03, I6803, MI903, MI907, MI993
The pulse output uses Turbo PMAC’s pulse frequency modulation (PFM) feature. The PFM circuitry
generates periodic output pulses by repeatedly adding a command value into an accumulator. When the
accumulator overflows, an output pulse is generated.
The addition of the command value into the accumulator is performed once per PFM clock (PFMCLK)
cycle. The PFMCLK frequency is governed by I7m03 for the channels on Servo IC m, I6803 for the
supplemental channels on MACRO IC 0, and MI903, MI907, or MI993 for channels on a MACRO
Station. The default frequency of the PFMCLK for all channels is 9.83 MHz; this frequency should be
suitable for all MLDT applications.
PFM Output Frequency: Mxx07, MI926
The pulse output frequency for a channel is controlled by both the PFMCLK frequency and the PFM
command value for the channel, which is the C-output register for that channel. When used for stepper
motor applications, the PFM command value is determined by the instantaneous command velocity and
the gains of the simulated servo loop on Turbo PMAC; Ixx02 tells Turbo PMAC to write this to the PFM
command register.
For MLDT use, we will write to the PFM command register once on power-up/reset with an M-variable.
(If the interface is on a MACRO Station, the Station’s firmware will do this automatically, using the
saved value of node-specific variable MI926.) The suggested M-variable definition for this register is
Mxx07. The following table shows the registers for these suggested definitions:
Setting Up Feedback and Master Position Sensors
63
Turbo PMAC User Manual
PFM Pulse Output Addresses (Y-registers)
IC# - Chan#
0-1
0-2
0-3
0-4
1-1
1-2
1-3
1-4
Mxx07->Y:
$078004
$07800C
$078014
$07801C
$078104
$07810C
$078114
$07811C
IC# - Chan#
2-1
2-2
2-3
2-4
3-1
3-2
3-3
3-4
Mxx07->Y:
$078204
$07820C
$078214
$07821C
$078304
$07830C
$078314
$07831C
IC# - Chan#
4-1
4-2
4-3
4-4
5-1
5-2
5-3
5-4
Mxx07->Y:
$079204
$07920C
$079214
$07921C
$079304
$07930C
$079314
$07931C
IC# - Chan#
6-1
6-2
6-3
6-4
7-1
7-2
7-3
7-4
Mxx07->Y:
$07A204
$07A20C
$07A214
$07A21C
$07A304
$07A30C
$07A314
$07A31C
IC# - Chan#
8-1
8-2
8-3
8-4
9-1
9-2
9-3
9-4
Mxx07->Y:
$07B204
$07B20C
$07B214
$07B21C
$07B304
$07B30C
$07B314
$07B31C
IC# - Chan#
M0 – 1*
M0 – 2*
Mxx07->Y:
$078414
$07841C
Servo ICs 0 and 1 are on the Turbo PMAC2 itself or on Acc-2E 3U-format stack boards.
Servo ICs 2 – 9 are on Acc-24x2 boards.
On dual-IC boards, IC channels 1 – 4 for odd-numbered Servo ICs are board channels 5 – 8.
MACRO IC channels are on the handwheel port
A sample M-variable definition is:
M107->Y:$078004,8,16,S
; Servo IC 0 Chan 1 PFM (C) cmd val
Note:
The PFM command registers are actually 24-bit registers; the suggested definitions
just use the high 16 bits, providing enough resolution for our purposes here. If you
use the register as a 24-bit value, e.g. M107->Y:$078004,0,24,S, it takes
numerical values 256 times higher than the equations shown here. When using the
MACRO Station, MI926 is a 24-bit, not 16-bit value.
The frequency of the pulse output should produce a period just slightly longer than the longest expected
response time for the echo pulse. For MLDTs, the response time is approximately 0.35 µsec/mm (9
µsec/inch). On an MLDT 1500 mm (~60 in) long, the longest response time is approximately 540 µsec; a
recommended period between pulse outputs for this device is 600 µsec, for a frequency of 1667 Hz.
To produce the desired pulse output frequency from a Turbo PMAC IC, the following formula can be
used (assuming a 16-bit M-variable definition):
OutputFreq( kHz ) =
Mxx07
PFMCLK _ Freq( kHz )
65 ,536
or:
Mxx07 = 65 ,536 *
OutputFreq( kHz )
PFMCLK _ Freq( kHz )
To produce a pulse output frequency of 1.667 kHz with the default PFMCLK frequency of 9.83 MHz, we
calculate:
Mxx07 = 65 ,536 *
1.667
9 ,830
≅ 11
To write this value to the register, a power-on PLC routine is suggested; this can also be done with online commands from the host computer. Turbo PMAC firmware will not write this value to the register
automatically.
To produce the desired pulse output frequency from a channel on the MACRO Station, the following
formula can be used for node-specific 24-bit variable MI926:
OutputFreq( kHz ) =
64
MI 926
16 ,777 ,216
PFMCLK _ Freq( kHz )
Setting Up Feedback and Master Position Sensors
Turbo PMAC User Manual
or:
MI 926 = 16 ,777 ,216 *
OutputFreq( kHz )
PFMCLK _ Freq( kHz )
To produce a pulse output frequency of 1.667 kHz with the default PFMCLK frequency of 9.83 MHz, we
calculate:
MI 926 = 16 ,777 ,216 *
1.667
9 ,830
= 2 ,982
Note:
The servo update time for the motor using the MLDT should be at least as high as
the output time set here (the servo frequency should be as low as or lower than the
output frequency).
PFM Pulse Width: I7m04, I6804, MI904, MI908, MI994
The width of the output pulse is controlled by the PFMCLK frequency with I7m04 for the channels on
Servo IC m, I6804 for the channels on MACRO IC 0, or MI904, MI908, or MI994 for the ICs on a
MACRO Station. This I-variable specifies the pulse width as the number of PFMCLK cycles. At the
default PFMCLK frequency of 9.83 MHz, the default value of 15 produces a 1.5-µsec output pulse width.
This should be satisfactory for most MLDT devices. When using the RPM format or equivalent (see
Signal Format, above), the pulse width must be large enough to enclose the rising edge of the returned
start pulse – that is, it must be longer than the delay between the output pulse and the returned start pulse.
PFM Format Select: I7mn6, I68n6, MI916
The output format of channel signals is controlled by variable I7mn6 for Servo IC m Channel n, by I68n6
for MACRO IC 0 Channel n, or by node-specific variable MI916 on a MACRO Station. In order for the
C-register circuitry of Channel n to output a PFM pulse train rather than a PWM pulse train, this variable
must be set to 2 or 3. Most commonly, it will be set to 3, so that the A and B registers for Channel n
output DAC signals rather than PWM.
Note:
One channel of Turbo PMAC cannot be used simultaneously for direct PWM
control of a motor and for MLDT pulse generation. Direct PWM control of a
motor writes to the channel’s A, B, and C registers every phase cycle
automatically.
MLDT Feedback Select: I7mn0, I68n0, MI910
The decoding of the signals on the encoder inputs is controlled by I7mn0 for Servo IC m Channel n, I68n0
for MACRO IC 0 Channel n, or by node-specific variable MI910 for a channel on a MACRO Station. For
proper decoding of the MLDT signal, this variable must be set to 12. With this setting, the pulse timer is
cleared to zero at the falling edge of the output pulse. It then counts up at 117.96 MHz until a rising edge on
the return pulse is received, at which time the timer’s value is latched into a memory-mapped register that
the processor can read. This register is the X-register at the base address of each channel.
Note:
The MLDT feedback uses the same circuitry that would be used for quadrature
encoder feedback on that channel, so an encoder and MLDT cannot be connected
simultaneously to the same channel’s feedback on Turbo PMAC. In this mode, it
is the pulse timer that is used as a position measurement for feedback, not the pulse
counter that is used with encoders. The counter still registers the number of pulses
returned, but does not represent a position measurement here.
Setting Up Feedback and Master Position Sensors
65
Turbo PMAC User Manual
Conversion Table Processing Setup – Turbo PMAC Interface
The timer registers are processed in the conversion table as parallel feedback representing the position,
just as an absolute encoder would be. The filtered parallel read (method digit $3) should be used to reject
errors due to extra or missing echo pulses. Consult the Setting up the Encoder Conversion Table section
in this manual and the specification for variables I8000 – I8191 in the Software Reference Manual.
Conversion Table Processing Setup – MACRO Station Interface
If the MLDT is connected to the MACRO Station, the data must be processed through conversion tables
on both the Station and the Turbo PMAC (although the table on the Turbo PMAC really just copies the
data off the ring). The conversion table on the MACRO Station is contained in setup variables MI120 –
MI151. This table should do a filtered parallel read of the timer register, the filtering able to reject errors
due to extra or missing echo pulses. The result is passed back over the ring to Turbo PMAC, where it is
treated just as any other type of feedback from the ring would be. Consult the section Setting Up the
Encoder Conversion Table in this manual, the specification for variables I8000 – I8191 in the Software
Reference Manual, and the manuals for the MACRO Station.
Setting Up for Power-On Absolute Position
Absolute Power-On Position Address and Format: Ixx10, Ixx95, MI11x
Because MLDTs are absolute position sensors, no homing search move is required if they are used for
position feedback. Turbo PMAC variables Ixx10 (Motor xx Power-On Servo Position Address) and
Ixx95 (Motor xx Power-On Position Format) can be used to establish this absolute position. The Ixx10
description in the Turbo PMAC Software Reference Manual has a complete table of the possible address
settings for the MLDT timer registers, and also for getting absolute position from a MACRO Station.
Ixx95 should be set to $170000 for an on-board MLDT timer register. It should be set to $740000 to
specify power-on position from a MACRO Station; station variable MI11x should be set to $17nnnn,
where “nnnn” represents the Y-address of the timer register on the MACRO Station.
Position Reference Offset: Ixx26
When used with an absolute position read, Ixx26 specifies the difference between the sensor’s zero
position and the motor zero position. With an MLDT, the sensor’s zero position is at the
transmitter/receiver module, outside the possible range of motion of the sensor. If it is desired that the
motor’s zero position be within the range of travel, Ixx26 should be used to define the offset between
sensor and motor zero. Ixx26 has units of 1/16 count, where a count here is an LSB of the timer.
Power-On Mode: Ixx80
Ixx80 controls whether the absolute position read for Motor xx is performed immediately at poweron/reset, or whether the absolute position read is delayed until a $* or $$* command is issued. Because
the output frequency for the MLDT is not established yet when a read would be done immediately at
power-on, it should be delayed with an MLDT. This requires that Bit 2 of Ixx80 be set to 1, making the
value of Ixx80 equal to 4 if the motor is not to be enabled immediately on power-on/reset, or equal to 5 if
the motor is to be enabled immediately.
Scaling the Feedback Units
Motor Units
For a motor set up with MLDT feedback, a count is one increment (LSB) of the timer. The physical
length of this increment, which is the resolution of the measurement, is dependent on the speed of the
return pulse in the MLDT, and the frequency of the timer in the Turbo PMAC. The speed of the return
pulse (the speed of sound in the metal) varies from device to device, but is always approximately the
inverse of 0.35 µsec/mm, or 9.0 µsec/inch. The frequency of the timer is 117.96 MHz. The resolution
can be calculated as:
66
Setting Up Feedback and Master Position Sensors
Turbo PMAC User Manual
1
 length 
 µ sec 
 length 

 * Re turnSpeed 

=
 increment  TimerFreq  increment 
 µ sec 
Re solution
increments 

 41.3

117.96 increment 0.35 µ sec
increment 
mm

1
1 inches
inches 
increments 
µ sec
≅
*
≅ 0.00094

 1060
117.96 increment 9.0 µ sec
increment 
inch

Axis User Units
≅
1
µ sec
*
1
mm
≅ 0.024
mm
Typically the axis assigned to the motor using the MLDT for feedback will be given engineering units of
millimeters or inches. The scaling is done in the axis definition statement. With the above example, an
axis definition statement to create axis user units of millimeters would be:
#1->41.3X
An axis definition statement to create axis user units of inches would be:
#1->1060X
Setting Up LVDTs, RVDTs, & Other Analog
Turbo PMAC systems can accept analog voltages from devices such as linear variable displacement
transducers (LVDTs), rotary variable displacement transducers (RVDTs), capacitive gauges, and
potentiometers, as feedback through optional analog-to-digital converters. These voltages must be DC;
that is, a constant position must be represented by a constant DC voltage (not an AC magnitude). This
means if the sensor is fundamentally AC with a carrier frequency (e.g. LVDTs, RVDTs), the sensor signal
must be demodulated to remove the AC carrier signal before it is connected to the Turbo PMAC system.
Typically, the vendors of these sensors can provide this kind of signal conditioning.
Hardware Setup
The analog signal should be wired into the connector for the analog accessory or option according to the
directions in the associated manual. If there is provision for differential inputs, and a single-ended input
is being used, physically wire the complementary input to the return voltage to get a proper reading
through the differential receiver.
Turbo PMAC Hardware-Control Parameter Setup
Most of the 12-bit A/D converters – as on Acc-36P, Acc-36V, Acc-36E, Acc-6E, and Option 12 – use
multiplexed converters. The first step in accessing these for servo use is to de-multiplex the data at a
fixed high rate. In Turbo PMAC, this de-multiplexing step is controlled by variables I5060 – I5096.
Each phase interrupt, one pair of ADCs is read and copied into memory registers and the next pair is
selected for the next conversion. Up to 16 pairs of ADCs can be in this ring. The de-multiplexed data is
left in registers Y:$003400 – Y:$00341F, where it can be read by the conversion table.
If the 12-bit A/D converters are in the MACRO Station, this de-multiplexing process is controlled by
Station variables MI987 and MI989. It is a fixed de-multiplexing of eight pairs of ADCs from one
device. MI989 specifies the address of the ADCs, and MI987 is set to 1 to enable the process. The
resulting data is placed in Station registers Y:$0200 – Y:$0207, with each pair of 12-bit values in a 24-bit
register.
Setting Up Feedback and Master Position Sensors
67
Turbo PMAC User Manual
Conversion Table Processing Setup – Turbo PMAC Interface
16-bit data from an Acc-28x converter connected to a Turbo PMAC is processed for servo feedback using
an Acc-28 A/D entry (method digit $1) in the Turbo PMAC’s conversion table. Fundamentally, this just
copies the data from a Y-register to an X-register (where the servo loop can access it) and shifts the data
so that the LSB out of the converter will be treated by the servo as a count. The conversion table sets all
of the fractional-count bits to zero. Unlike the parallel data conversion, it will not roll over and softwareextend the data. The integrated Acc-28 A/D entry (method digit $5) does the transfer and shift too, but
also adds the previous result to the present cycle’s data. In either case, if the A/D data is unsigned, as
from an Acc-28B or 28E, the bit-19 mode switch must be set to 1 to handle the data properly.
Data from 12-bit A/D converters is processed using the parallel data conversion (method digit $2 without
filtering or $3 with filtering). The source address is the de-multiplexed memory register, not the address
of the ADC itself (unless you are only using a single channel of the ADC). The width/offset setup word
in the entry is $00C00C (width 12, offset 12) to use data in the high 12 bits of the register. This
conversion transfers the data to an X-register and shifts the data so that the LSB out of the converter will
be treated by the servo as a “count”. The conversion table sets all of the “fractional-count” bits to zero.
Note that this conversion does permit rollover and software extension, so make sure there cannot be an
11-bit change (half of the input range) in a single servo cycle, or Turbo PMAC will roll over the result.
For details of setting up the encoder conversion table to process analog-to-digital converters, consult the
section Setting up the Encoder Conversion Table in this manual and the specification for variables I8000
– I8191 in the Software Reference Manual.
Conversion Table Processing Setup – MACRO Station Interface
16-bit data from an Acc-28x converter connected to a MACRO Station is processed for servo feedback
using an “Acc-28 A/D” entry (method digit $1) in the MACRO Station’s conversion table.
Fundamentally, this just copies the data from a Y-register to an X-register (where the servo loop can
access it) and shifts the data so that the LSB out of the converter will be treated by the servo as a count.
The conversion table sets all of the fractional-count bits to zero. Unlike the parallel data conversion, it
will not roll over and software-extend the data. The integrated Acc-28 A/D entry (method digit $5) does
the transfer and shift too, but also adds the previous result to the present cycle’s data. Because the only
16-bit ADCs that can be connected to a MACRO Station are unsigned, the bit-19 mode switch must be set
to 1 (making the second hex digit 8) to handle the data properly. The resulting data is copied to the
MACRO ring using MI10x, and then treated in the Turbo PMAC as any MACRO feedback data would
be, using the unshifted parallel conversion, with the MACRO node address as the source.
Data from 12-bit A/D converters is processed in the MACRO Station’s using the parallel data conversion
(method digit $2 without filtering or $3 with filtering). The source address is the de-multiplexed memory
register, not the address of the ADC itself (unless you are only using a single channel of the ADC). The
bits-used mask setup word in the entry is either $000FFF to specify the low 12 bits, or $FFF000 to use the
high 12 bits. This conversion transfers the data to an X-register and shifts the data so that the LSB out of
the converter will be treated by the servo as a count. The conversion table sets all of the fractional-count
bits to zero.
This conversion does permit rollover and software extension, so you must make sure there cannot be an
11-bit change (half of the input range) in a single servo cycle or Turbo PMAC will roll over the result.
For details of setting up the encoder conversion table to process analog-to-digital converters, consult the
MACRO Station manuals, the section Setting up the Encoder Conversion Table in this manual, and the
specification for variables I8000 – I8191 in the Software Reference Manual.
68
Setting Up Feedback and Master Position Sensors
Turbo PMAC User Manual
Setting Up for Power-On Absolute Position
Absolute Power-On Position Address and Format: Ixx10, Ixx95, MI11x
Generally, position data that comes to a Turbo PMAC system as an analog voltage is absolute in nature,
so no homing search move is required if it is used for position feedback. Turbo PMAC variables Ixx10
(Motor xx Power-On Servo Position Address) and Ixx95 (Motor xx Power-On Position Format) can be
used to establish this absolute position. The Ixx10 description in the Turbo PMAC Software Reference
Manual has a complete table of the possible address settings for the ADC registers, and also for getting
absolute position from a MACRO Station.
Ixx95 should be set to $B10000 for the signed 16-bit data from an Acc-28A; it should be set to $310000
for the unsigned 16-bit data from an Acc-28B. For 12-bit unsigned data, it should be set to $0C0000; for
12-bit signed data, it should be set to $8C0000. It should be set to $740000 to specify power-on position
from a MACRO Station; station variable MI11x should be set to $17nnnn, where “nnnn” represents the
Y-address of the ADC register (or de-multiplexed register in memory) on the MACRO Station.
Scaling the Feedback Units
If the above instructions are followed, the increment of the least significant bit (LSB) of the ADC is
considered a count by the Turbo PMAC motor functions. All subsequent position, velocity, and
acceleration units are based on this definition of a count.
Setting Up Absolute Encoders
Turbo PMAC can accept feedback data from a variety of absolute encoders through various interface
boards. The instructions for setting up the hardware and software for these interfaces is detailed in the
manuals for the particular interface boards. The following table briefly summarizes these setups:
Encoder
Interface Board
Conversion Table
Absolute Power On*
Parallel Format
Acc-14D
Filtered Parallel ($3)
Ixx95=$nn0000
Acc-14V
Filtered Parallel ($3)
Ixx95=$nn0000
Acc-14E
Filtered Parallel Byte ($F/$3)
Ixx95=$nn000b
Yaskawa Sigma I
Acc-8D Opt 9
1/T Encoder ($0)
Ixx95=$710000
SSI
Acc-53E
Filtered Parallel ($3)
Ixx95=$nn0000
Hiperface
Acc-51x Opt 2
Hi-Res Interpolator ($F/$0)
Ixx95=$nn0000
Heidenhain EnDat
Acc-51x Opt 3
Hi-Res Interpolator ($F/$0)
Ixx95=$nn0000
Mitsubishi
Acc-57E Opt M
Filtered Parallel ($3)
Ixx95=$nn0000
Sanyo
Acc-49P
Filtered Parallel ($3)
Ixx95=$nn0000
Tamagawa
Acc-70P
Filtered Parallel ($3)
Ixx95=$nn0000
*The nn in $nn0000 expresses the number of bits to use as a hexadecimal value. The b specifies which
byte of the 24-bit word to use: 4 for low, 5 for middle, 6 for high.
Details on the actual hardware and software setup for these data formats can be found in the user manual
for the appropriate accessory.
Setting Up Feedback and Master Position Sensors
69
Turbo PMAC User Manual
70
Setting Up Feedback and Master Position Sensors
Turbo PMAC User Manual
BASIC MOTOR SETUP
Turbo PMAC has many modes for controlling motors. A major part of the initial setup of a Turbo PMAC
is the hardware and software configuration to specify a specific mode of operation. The commonly used
modes of operation are:
• Analog command of velocity-mode drives
• Analog command of torque-mode drives
• Analog command of sine-wave input drives
• Direct PWM control of power-block drives
• Pulse-and-direction command of stepper or stepper-replacementservo drives
Any of these modes can be employed directly in a Turbo PMAC system, or over the MACRO ring.
When using the MACRO ring, the command and feedback values are passed across the ring as binary
numerical values; the actual generation of command signals and processing of feedback signals is done at
the remote MACRO node. The motor algorithms in the Turbo PMAC are the same regardless of whether
the MACRO ring is used or not. The choice of mode of operation is independent for each motor.
Hardware Setup
The details of the hardware setup of the machine interface ports are dependent on the actual type of the
Turbo PMAC family used, and often the style of the interface. Broadly speaking, there are four classes of
interface:
• Turbo PMAC boards
• Turbo PMAC2 boards
• 3U-format Stack boards (Turbo Stack and MACRO Stack)
• 3U-format UMAC (pack) boards (UMAC Turbo and UMAC MACRO)
This section summarizes the connection strategies for these classes of boards. Details of the connections
and other hardware setup issues are covered in the Hardware Reference manuals for the individual boards.
Parameters to Set up Basic Motor Operation
Each Motor xx has setup I-variables to permit specific software configuration of that motor’s control
algorithms. The hundreds’ and thousands’ digit of the I-variable specifies which motor is being
configured; for example, I1200 activates or de-activates Motor 12. The generic reference for the variable
uses the letters xx for the motor digits; Ixx00 refers generally to the activation variable for Motor xx,
where xx = 1 to 32.
Most of the software configuration of a motor involves setting proper values for these variables, as
explained in the following sections. This section has a quick survey of the key variables, and the
variables that are common to all modes. Depending on how the motor is set up, branch to other sections
of the manual. The following flow chart summarizes the motor setup possibilities:
Basic Motor Setup
71
Turbo PMAC User Manual
Beginning
Motor
Setup
No
Use PMAC to
commutate and/or close
current loop?
Yes
Single Output
Setup
Basic
Commutation
Setup
Use PMAC to close
current loop?
Analog
Output?
No
Yes
Pulse and
Direction
Setup
Analog
Output
Setup
No
Yes
Sine-Wave
Output
Setup
Direct PWM
Output
Setup
Current Loop
Setup/Tuning
Synchronous
Motor?
No
Yes
Slip and
Magnetization
Current Setup
Phase
Referencing
Position/Velocity
Servo Loop
Setup
Turbo PMAC Motor Setup Flowchart
Initial Setup Parameters
Activating the Motor: Ixx00
The Ixx00 motor activation parameter must be set to 1 for any motor that is to be used on Turbo PMAC. It
should be set to 0 for any motor that is not used, so Turbo PMAC will not waste computation time on that
motor. An activated motor can be enabled or disabled; a de-activated motor is not monitored in any way.
Activating PMAC Motor Commutation: Ixx01
Bit 0 of the Ixx01 commutation-enable parameter must be set to 1 if Turbo PMAC is performing the
commutation and/or digital current-loop closure for the motor. It must be set to 0 for any motor if Turbo
PMAC is performing neither the commutation nor digital current loop for the motor. If commutation is
enabled, variables Ixx70 – Ixx84 for the motor must be set to specify how the commutation is done. The
instructions for setting these variables are given in the Commutation section of this manual.
72
Basic Motor Setup
Turbo PMAC User Manual
Motor Address Setup Parameters
Each Turbo PMAC motor has several address I-variables that tell the motor what registers to use for its
inputs and outputs. Each of these variables contains the Turbo PMAC address of the register for the
particular function. This provides a mapping between the motor calculation registers and the different
types of servo I/O registers (encoders, D/A converters, flags, etc.) used for the physical interface. By
providing a user-settable mapping, Turbo PMAC makes it very easy to utilize different types of feedback
and output signals for each motor.
The following diagram shows how this mapping works for a motor directly accessing servo I/O registers
(without MACRO):
PMAC Direct Register Mapping
Output A
Output B
DAC, PWM
Output C
PFM, PWM
Control
Ix02
DAC, PWM
Amp Enable
Ix25
Motor x
Calculation
Registers
Ix03
Ix04
Ix83
Ix82
Ix25
Status
ADC A
ADC B
Encoder
Conversion
Table
Encoder
Flags
Current
Feedback
Signals
Position
Feedback Signal
This next diagram shows the three-stage process to map between the motor calculation registers and the
servo I/O registers in a MACRO Station over a MACRO ring. The motor address I-variables specify the
first stage only – the mapping between the motor calculation registers and the Turbo PMAC’s own
MACRO node registers. The second step – the mapping between the Turbo PMAC’s MACRO node
registers and the MACRO Station’s MACRO node registers – is covered in the Basic Setup section (and
in the MACRO Station manuals). The third step – the mapping between the MACRO Station’s MACRO
node registers – is covered in the appropriate section of this manual (and in the MACRO Station
manuals). Once this mapping is set up, the motor will operate just as it would with a direct connection.
PMAC / MACRO Station Register Mapping
Motor x
Calculation
Registers
MACRO Station
Machine Interface
Channel x Registers
Servo &
Commutation
Address
I-variables
Channel
Interface Signals
Station SW1 Setting,
Conversion Table,
M110x
PMAC
MACRO Node
n Registers
MACRO Ring
Connection
MACRO Station
Node n
Registers
I996 bits 20-23 = Station SW2 Setting
Turbo: I6841, I6891, I6941, I6991 bits 20-23 = Station SW2 Setting
PMAC Node n = Station Node n
PMAC
Basic Motor Setup
MACRO Station
73
Turbo PMAC User Manual
Command Output Address: Ixx02
Ixx02 instructs Turbo PMAC where to place its output command values for Motor xx by specifying the
address of the register (or the first register if multiple outputs are used). The default values of Ixx02 use
the output registers for the machine interface channel usually assigned to the motor, or command registers
(starting with Register 0) for the MACRO node usually assigned to the motor. The exact setting is
dependent on the mode used, and is covered in the section for each mode of control.
Position-Loop Feedback Address: Ixx03
Ixx03 specifies the address of the register Turbo PMAC reads for the ongoing position-loop feedback for
the motor. This is the address of a processed feedback data register in the encoder conversion table. With
the default conversion table, the default values of Ixx03 specify the processed data for the encoder whose
channel matches the motor; these values are suitable for most applications, regardless of the command
output mode. Refer to the Encoder Conversion Table section in this manual for more information.
Velocity-Loop Feedback Address: Ixx04
Ixx04 specifies the address of the register Turbo PMAC reads for the velocity-loop feedback for the
motor. It works just like Ixx03, and contains the same address as Ixx03 unless dual feedback is used for
the motor.
Flag Addresses: Ixx25, Ixx42, Ixx43
Ixx25, and possibly Ixx42 and Ixx43, specify the addresses of the registers Turbo PMAC uses for its flag
information for Motor xx. The flags come in three sets:
1. The capture input flags (including the encoder index), used for trigger moves such as homing search
moves
2. The amplifier flags (enable output and fault input), used to handshake with the drive
3. The overtravel limit input flags, used to enforce the position range of the motor
If Ixx42 and Ixx43 are set to the default value of 0, Ixx25 sets the address of all three sets of flags. This is
the typical case, and the only one permitted in firmware revisions 1.939 and earlier (through 2002).
However, starting with firmware revision 1.940, new variables Ixx42 and Ixx43 permit the splitting of the
sets of flags. This is particularly useful when triggering from an Acc-51 encoder interpolator board, or
when using the MACRO ring for feedback and/or triggering but not the amplifier interface.
If Ixx42 is set to a non-zero value, it specifies the address of the amplifier flags alone. If Ixx43 is set to a
non-zero value, it specifies the address of the overtravel limit flags alone. Ixx25 always specifies the
address of the capture flags.
The following table shows the possible addresses for these variables when the flags are accessed through
PMAC-style Servo ICs.
PMAC-Style Servo IC Flag Addresses
IC# - Chan#
Ixx25/42/43
IC# - Chan#
Ixx25/42/43
IC# - Chan#
Ixx25/42/43
IC# - Chan#
Ixx25/42/43
IC# - Chan#
Ixx25/42/43
0-1
$078000
2-1
$078200
4-1
$079200
6-1
$07A200
8-1
$07B200
0-2
$078004
2-2
$078204
4-2
$079204
6-2
$07A204
8-2
$07B204
0-3
$078008
2-3
$078208
4-3
$079208
6-3
$07A208
8-3
$07B208
0-4
$07800C
2-4
$07820C
4-4
$07920C
6-4
$07A20C
8-4
$07B20C
1-1
$078100
3-1
$078300
5-1
$079300
7-1
$07A300
9-1
$07B300
1-2
$078104
3-2
$078304
5-2
$079304
7-2
$07A304
9-2
$07B304
1-3
$078108
3-3
$078308
5-3
$079308
7-3
$07A308
9-3
$07B308
1-4
$07810C
3-4
$07830C
5-4
$07930C
7-4
$07A30C
9-4
$07B30C
Servo ICs 0 and 1 are on the Turbo PMAC itself.
Servo ICs 2 – 9 are on Acc-24P/V or Acc-51P boards.
Channels 1 – 4 on odd-numbered Servo ICs are Channels 5 – 8 on the boards.
74
Basic Motor Setup
Turbo PMAC User Manual
The following table shows the possible addresses for these variables when the flags are accessed through
PMAC2-style Servo ICs.
PMAC2-Style Servo IC Flag Addresses
IC# - Chan#
Ixx25/42/43
IC# - Chan#
Ixx25/42/43
IC# - Chan#
Ixx25/42/43
IC# - Chan#
Ixx25/42/43
IC# - Chan#
Ixx25/42/43
0-1
$078000
2-1
$078200
4-1
$079200
6-1
$07A200
8-1
$07B200
0-2
$078008
2-2
$078208
4-2
$079208
6-2
$07A208
8-2
$07B208
0-3
$078010
2-3
$078210
4-3
$079210
6-3
$07A210
8-3
$07B210
0-4
$078018
2-4
$078218
4-4
$079218
6-4
$07A218
8-4
$07B218
1-1
$078100
3-1
$078300
5-1
$079300
7-1
$07A300
9-1
$07B300
1-2
$078108
3-2
$078308
5-2
$079308
7-2
$07A308
9-2
$07B308
1-3
$078110
3-3
$078310
5-3
$079310
7-3
$07A310
9-3
$07B310
1-4
$078118
3-4
$078318
5-4
$079318
7-4
$07A318
9-4
$07B318
Servo ICs 0 and 1 are on the Turbo PMAC2 itself or on Acc-2E 3U-format stack boards.
Servo ICs 2 – 9 are on Acc-24x2 or Acc-51E boards.
Channels 1 – 4 on odd-numbered Servo ICs are Channels 5 – 8 on the dual-IC boards.
If the flags are accessed through the MACRO ring, the address specified is that of a dedicated image
register in memory for the MACRO node, not the MACRO node flag register itself. The following table
shows the possible addresses for these variables when the flags are accessed through the MACRO ring.
MACRO Ring Flag Addresses
IC# - Node#
0-0
0-1
0-4
0-5
0-8
0-9
0 - 12
0 - 13
Ixx25/42/43
$003440
$003441
$003444
$003445
$003448
$003449
$00344C
$00344D
IC# - Node#
1-0
1-1
1-4
1-5
1-8
1-9
1 - 12
1 - 13
Ixx25/42/43
$003450
$003451
$003454
$003455
$003458
$003459
$00345C
$00345D
IC# - Node#
2-0
2-1
2-4
2-5
2-8
2-9
2 - 12
2 - 13
Ixx25/42/43
$003460
$003461
$003464
$003465
$003468
$003469
$00346C
$00346D
IC# - Node#
3-0
3-1
3-4
3-5
3-8
3-9
3 - 12
3 - 13
Ixx25/42/43
$003470
$003471
$003474
$003475
$003478
$003479
$00344C
$00347D
Flag Modes: Ixx24
Variable Ixx24 is a collection of individual control bits that determines how the flags specified in Ixx25,
Ixx42, and Ixx43 are used (or not used). Each bit is described in detail in the Software Reference
Manual; a quick summary is given here.
Bit 0:
Set to 0 for flags in a PMAC-style Servo IC, to 1 for flags in a PMAC2-style Servo IC.
Bit 11: Set to 0 to capture with quadrature encoders, to 1 to capture with sinusoidal encoders through
Acc-51 with high-resolution interpolation.
Bit 12: Set to 0 to use whole-count capture, to 1 to use fractional-count capture.
Bit 14: Set to 0 to stop on desired-position limit, to 1 to saturate on desired-position limit.
Bit 15: Set to 0 to disable desired-position limits, to 1 to enable.
Bit 16: Set to 0 to use amplifier-enable output flag, to 1 not to use.
Bit 17: Set to 0 to use overtravel limit flags, to 1 not to use.
Bit 18: Set to 0 to use capture flags from Turbo PMAC, to 1 for capture flags through MACRO.
Bit 20: Set to 0 to use amplifier-fault input flag, to 1 not to use.
Bits 21 and 22: Set to 00 to kill all motors on a fault; to 01 to kill motors in same coordinate system, to 10
to kill only faulted motor.
Bit 23: Set to 0 for low-true amplifier fault (0 means fault), to 1 for high-true amplifier fault.
Basic Motor Setup
75
Turbo PMAC User Manual
Absolute Power-Up Position Address and Format: Ixx10 and Ixx95
If you have a position sensor for the motor that is absolute over the entire range of travel for the motor,
you can use motor variables Ixx10 and Ixx95 to tell Turbo PMAC where to read this absolute position
data, and how to interpret the format, respectively. This does not have to be the same register or even the
same sensor that is selected for ongoing position feedback with Ixx03. If Ixx10 is left at its default value
of 0, no absolute position read will be performed for the motor, and the motor’s power-up/reset position is
set automatically to zero. In this case, a homing-search move must be used to establish the position
reference for the motor.
(If the position sensor is absolute only over one motor revolution or commutation cycle, as with a singleturn absolute encoder or resolver, and the motor is commutating with Turbo PMAC, use similar variables
Ixx81 and Ixx91 to establish the absolute phase position, but in this case Ixx10 will be set to 0, and a
homing-search move will still be required. The setup for absolute phase position is covered in the
Commutation section.)
The absolute position read defined by Ixx10 and Ixx95 will be performed automatically at power-up/reset
if bit 2 of Ixx80 is set to its default value of 0. This absolute read can be done subsequently on the $*
command, whether or not it was done on power-up/reset.
The sensor used for absolute power-on position does not have to be the same sensor as that used for
ongoing position feedback, and it does not have to use the same data format or channel. However, in
order to use the automatic absolute power-on position read with Ixx10 and Ixx95, it must have the same
resolution as the ongoing position feedback. If the absolute power-on position has a different resolution,
you must write directly to the motor’s absolute position feedback register (suggested M-variable Mxx62),
usually from a power-on PLC program. If the absolute position sensor for Motor 1 has 0.4 counts per
count of the ongoing position feedback and this sensor were read through M150, the power-on PLC
statement to set the absolute position could be:
M162=M150*0.4*I108*32
This must be done with the motor’s loop open. Typically, power-up mode variable Ixx80 (see just below)
will be set to 0 or 2 and the loop will be closed after this statement is executed.
Power-Up Mode: Ixx80
Ixx80 specifies whether PMAC will try to control the motor immediately on power-up/reset (bit 0 = 1),
performing a phasing search if necessary; or whether it will put the motor in a killed state on powerup/reset (bit 0 = 0) and await a command to enable the motor. If not sure that everything is instantly
ready for control when Turbo PMAC is powered up, set bit 0 of Ixx80 to 0 (making Ixx80 an even
number). In general, the more complex and high-powered the system is, the more likely to delay enabling
of the motor to ensure that everything is ready before attempting to enable the motor.
Ixx80 also controls what kind of phasing-search move is done if one is required for a PMAC-commutated
motor (this is covered in the Commutation section), and whether an absolute position read is done
automatically on power-up/reset.
76
Basic Motor Setup
Turbo PMAC User Manual
Is Turbo PMAC Commutating or Closing the Current Loop for This
Motor?
All motors of significant travel require commutation (reversal of current) in the motor phases in order to
generate consistent torque/force as the motor moves. The only question is where and how this
commutation is done. In a brush DC motor the commutation is performed mechanically inside the motor.
With brushless motors, the commutation is often performed electronically inside the drive. In these cases,
Turbo PMAC is not performing the commutation.
Virtually all modern servomotor control employs current-loop closure for high response, tolerance of
parameter variation, and protection against overcurrent conditions. While this has traditionally performed
in the servo drive, Turbo PMACs with PMAC2-style Servo ICs are capable of closing the current loop
digitally for motors.
If Turbo PMAC is either performing the phase commutation, closing the current loop, or both, for the
motor, refer to the Setting Up Turbo PMAC-Based Commutation and/or Current Loop sections for further
instructions. If Turbo PMAC is doing neither task for this motor, continue below in this section.
There are two subsequent sections in this section. The first deals with using the traditional analog
velocity-mode or torque-mode interface, still the most common servo-amplifier interface. The second
deals with the pulse-and-direction interface, the traditional and still most common stepper-amplifier
interface, also used with “stepper-replacement” servo amplifiers.
Setting up Turbo PMAC for Velocity or Torque Control
If Turbo PMAC is not performing the commutation or current loop for a motor, it provides a single output
command value for the motor. Usually this output represents either a velocity command or a torque
(force, or current magnitude) command, and typically this output is encoded as an analog signal voltage
level. While the servo-loop tuning for velocity and torque commands is different, the setup until that
point is identical for both modes.
When driving a hydraulic cylinder through either a proportional valve or a servo valve, the dynamics
appear to the Turbo PMAC to be those of a velocity command to a motor.
Hardware Setup
Each axis-interface channel on a Turbo PMAC board or axis-expansion board with PMAC-style Servo
ICs (e.g. Acc-24P or Acc-24V) has a single 16-bit analog output. Breakout boards simply direct this
signal to an appropriate connector.
Board-level Turbo PMAC2 controllers do not have on-board analog outputs for their servo channels; a
breakout board with D/A converters (DACs), such as the Acc-8A or the Acc-8E must be used. With
UMAC systems, the Acc-24E2A analog axis-interface board has DACs for this purpose. With UMACCPCI systems, the Acc-24C2A has DACs for this purpose. With the QMAC box-level controller, if the
analog option is ordered, the DACs are installed inside. In all of these systems, two DACs per channel
are either standard or optional, but only a single DAC is required for velocity or torque-mode control.
The DACs provided by Delta Tau for Turbo PMAC2 systems have 18-bit resolution.
Consult the appropriate hardware reference or accessory manual for the details of the hardware setup and
connection.
Turbo PMAC Parameter Setup
Hardware Setup for PMAC2-Style ICs
If the analog output is created through a PMAC2-style Servo IC, which supports other output types as
well, a few parameters in the IC must be set up to achieve the analog output. (This is not necessary if the
analog output is created through a PMAC-style Servo IC.)
Basic Motor Setup
77
Turbo PMAC User Manual
DAC Clock Frequency Control: I7m03, MI903, MI907, MI993
An I-variable specifies the frequency of the DACCLK signal that controls the rate at which data is
clocked into the serial DACs on all channels of the Servo IC. This variable is I7m03 for a Turbo
PMAC’s Servo IC m. If the IC is part of a MACRO Station, the variable is MI903, MI907, or MI993.
The default DACCLK frequency of 4.92 MHz is appropriate for all DACs used by Delta Tau. This
variable also controls other clock frequencies for the Servo IC; if you change the value of this variable,
make sure to keep the DACCLK frequency the same.
DAC Strobe Control: I7m05, MI905, MI909, MI999
A PMAC2-style Servo IC generates a common DAC strobe word for all four channels on the IC. It does
this by shifting out a 24-bit word each phase cycle, one bit per DAC clock cycle, most significant bit first.
I7m05 contains this word for the channels on a Turbo PMAC’s Servo IC m. If the IC is part of a
MACRO Station, the variable is MI905, MI909, or MI999. The default value of $7FFFC0 is suitable for
use with the 18-bit DACs used by Delta Tau.
Output Mode Control: I7mn6, MI916
I7mn6 must be set to 1 or 3 to specify that outputs A and B for Servo IC m Channel n are in DAC mode,
not PWM. (On a MACRO Station, this is controlled by node-specific variable MI916.) A setting of 1
puts output C (not used for servo or commutation tasks in this mode) in PWM mode; a setting of 3 puts
output C in PFM mode.
Output Inversion Control: I7mn7, MI917
I7mn7 controls whether or not the serial data streams to the DACs on Servo IC m Channel n are inverted.
(On a MACRO Station, this is controlled by node-specific variable MI917.) The default value of 0 (noninverted) is suitable for use with any of the Delta Tau analog outputs. Inverting the bits of the serial data
stream has the effect of negating the DAC voltage. In a servo algorithm this changes the polarity match
between output and input, which would produce a dangerous runaway condition if the system were
working properly before the inversion.
Motor Parameter Setup
Commutation Enable/Disable: Ixx01
Bit 0 of Ixx01 must be set to 0 to disable the commutation algorithms for Motor xx. Commutation is
performed in the drive or the motor in this mode of operation. Bit 1 of Ixx01 is set to 0 to specify that the
output register whose address is specified in Ixx02 is a Y-register. Virtually all of the output devices
usable in this mode are located in Y-registers, so usually Ixx01 is left at the default value of 0 here.
Command Output Address: Ixx02
Ixx02 instructs Turbo PMAC where to place its output commands for Motor xx by specifying the address.
The default values of Ixx02 use the DAC register for the Machine Interface Channel matching the motor.
Ixx02 seldom needs to be changed from the default value for DAC applications. Typically, the values
used are:
PMAC-Style Servo IC Command Output Addresses (Y-registers)
IC# - Chan#
0-1
0-2
0-3
0-4
1-1
1-2
1-3
1-4
Ixx02
$078003
$078002
$07800B
$07800A
$078103
$078102
$07810B
$07810A
IC# - Chan#
2-1
2-2
2-3
2-4
3-1
3-2
3-3
3-4
Ixx02
$078203
$078202
$07820B
$07820A
$078303
$078302
$07830B
$07830A
IC# - Chan#
4-1
4-2
4-3
4-4
5-1
5-2
5-3
5-4
Ixx02
$079203
$079202
$07920B
$07920A
$079303
$079302
$07930B
$07930A
IC# - Chan#
6-1
6-2
6-3
6-4
7-1
7-2
7-3
7-4
Ixx02
$07A203
$07A202
$07A20B
$07A20A
$07A303
$07A302
$07A30B
$07A30A
IC# - Chan#
8-1
8-2
8-3
8-4
9-1
9-2
9-3
9-4
Ixx02
$07B203
$07B202
$07B20B
$07B20A
$07B303
$07B302
$07B30B
$07B30A
Servo ICs 0 and 1 are on the Turbo PMAC itself.
Servo ICs 2 – 9 are on Acc-24P/V or Acc-51P boards.
Channels 1 – 4 on odd-numbered Servo ICs are Channels 5 – 8 on the boards.
78
Basic Motor Setup
Turbo PMAC User Manual
PMAC2-Style Servo IC Command Output Addresses (Y-registers)
IC# - Chan#
Ixx02
IC# - Chan#
Ixx02
IC# - Chan#
Ixx02
IC# - Chan#
Ixx02
IC# - Chan#
Ixx02
0-1
0-2
0-3
0-4
1-1
1-2
1-3
1-4
$078002
$07800A
$078012
$07801A
$078102
$07810A
$078112
$07811A
2-1
2-2
2-3
2-4
3-1
3-2
3-3
3-4
$078202
$07820A
$078212
$07821A
$078302
$07830A
$078312
$07831A
4-1
4-2
4-3
4-4
5-1
5-2
5-3
5-4
$079202
$07920A
$079212
$07921A
$079302
$07930A
$079312
$07931A
6-1
6-2
6-3
6-4
7-1
7-2
7-3
7-4
$07A202
$07A20A
$07A212
$07A21A
$07A302
$07A30A
$07A312
$07A31A
8-1
8-2
8-3
8-4
9-1
9-2
9-3
9-4
$07B202
$07B20A
$07B212
$07B21A
$07B302
$07B30A
$07B312
$07B31A
Servo ICs 0 and 1 are on the Turbo PMAC2 itself or on Acc-2E 3U-format stack boards.
Servo ICs 2 – 9 are on Acc-24x2 or Acc-51E boards.
Channels 1 – 4 on odd-numbered Servo ICs are Channels 5 – 8 on the dual-IC boards.
When the Turbo PMAC is operating Motor xx in velocity mode or torque mode commanded over a
MACRO ring, Ixx02 will contain the address of a MACRO output register for one of the MACRO nodes.
When using the Type 1 MACRO protocol commonly found on multi-axis MACRO components such as
the Delta Tau MACRO Station, the following Ixx02 values should be used. Typically, these will be the
default values on a Turbo PMAC2 Ultralite.
MACRO Node Command Output Addresses (Y-registers)
IC# - Node#
0-0
0-1
0-4
0-5
0-8
0-9
0 - 12
0 - 13
Ixx02
$078420
$078424
$078428
$07842C
$078430
$078434
$078438
$07843C
IC# - Node#
1-0
1-1
1-4
1-5
1-8
1-9
1 - 12
1 - 13
Ixx02
$079420
$079424
$079428
$07942C
$079430
$079434
$079438
$07943C
IC# - Node#
2-0
2-1
2-4
2-5
2-8
2-9
2 - 12
2 - 13
Ixx02
$07A420
$07A424
$07A428
$07A42C
$07A430
$07A434
$07A438
$07A43C
IC# - Node#
3-0
3-1
3-4
3-5
3-8
3-9
3 - 12
3 - 13
Ixx02
$07B420
$07B424
$07B428
$07B42C
$07B430
$07B434
$07B438
$07B43C
If using the older Type 0 MACRO protocol, add 1 to the value shown in the above table (e.g. $078420
becomes $078421).
Setting Up Turbo PMAC2 for Pulse-and-Direction Control
A Turbo PMAC with PMAC2-style Servo ICs is capable of commanding stepper-motor drives that
require pulse-and-direction format input, or “stepper-replacement” servo drives that require this format.
Turbo PMAC can command these drives either in open-loop fashion, in which case it internally routes the
pulse train into its own encoder counters to create a pseudo-closed loop, or in closed-loop fashion, in
which case an external feedback device is wired to the Turbo PMAC to create a true feedback loop.
A PMAC2-style Servo IC creates its pulse-and-direction output signal with an on-board fully digital
pulse-frequency-modulation (PFM) circuit. This circuit repeatedly adds the latest command frequency
value into an accumulator at a programmable rate of up to 40 MHz. When the accumulator overflows, an
output pulse is generated with a positive direction signal; when the accumulator underflows, an output
pulse is generated with a negative direction signal. This creates a pulse train whose frequency is directly
proportional to the command value, with virtually no harmonic distortion, and none of the offset problems
that affect analog pulse generation schemes.
Basic Motor Setup
79
Turbo PMAC User Manual
PMAC2 Pulse and Direction Stepper System
PMAC2
Drive and Motor
Over/Under flow
Pulse
Master
Position
Stepper
Drive
Accumulator
+
Trajectory +
Generation
PID
-
Adder
Direction
PFM Circuit
Decoder/
Counter
E
Note:
The analog output of a PMAC-style Servo IC, used in sign-and-magnitude mode
(Ixx96=1), and passed through a voltage-to-frequency converter, can be used for
the same type of operation. However, it is strongly recommended that the purely
digital method of the PMAC2-style Servo IC be used instead.
Hardware Setup
PMAC2-style Servo ICs, and the DSPGATE2 MACRO IC, have pulse-and-direction outputs for each
channel on the IC. In most configurations of interface and breakout hardware, these signals are accessible
as RS-422-level differential line-driver output pairs. These signals are driven by the value in output
register C for the channel, with the pulse frequency proportional to the value in this register. For this
reason, these outputs are technically known as pulse-frequency-modulated (PFM) outputs.
Alternately, the signals from the C output register can be used as pulse-width-modulated (PWM) outputs,
and commonly form the third-phase command signals for direct-PWM output of brushless motors. Note
that if using the channel for direct-PWM control of a motor, the PFM outputs will not be seen on the same
channel.
For board-level Turbo PMAC2 controllers and 3U-format stack boards (Acc-1E and 2E), the Acc-8S
breakout board usually is the most effective way of bringing out the pulse-and-direction signals. These
signals are available also on the Acc-8A and Acc-8E analog breakout boards and the Acc-8F PWM
breakout boards (as the third-phase PWM).
The pulse-and-direction signals are available on the encoder connectors of the QMAC boxed controller
and of the UMAC Acc-24E2S, Acc-24E2A, and Acc-24E2 axis-interface boards. The outputs use the
same pins as the T, U, V, and W flag inputs for the channel; a jumper must be installed to enable the
outputs on these pins.
Signal Timing
The PULSEn and DIRn signals are driven from the internal PFMCLK signal, whose frequency is controlled
by I7m03 (see below). The width of the pulse is controlled by the PFMCLK frequency and I7m04 (see
below). The output on PULSEn can be high-true (high during pulse, low otherwise) or low-true, as
controlled by I7mn7; the default is high-true. The polarity of the DIRn signal is controlled by I7mn8.
PULSEn and DIRn signals can change only on the rising edge of PFMCLK. If DIRn changes on a pulse,
it will change simultaneously with the front end of PULSEn. Some stepper drives require a setup time on
the DIRn line before the rising edge of PULSEn; these systems can be accommodated by inverting the
PULSEn signal with I7mn7.
The DIRn signal is latched in this state at least until the front end of the next pulse. The PULSEn signal
stays true for the number of PFMCLK cycles set by I7m04. It then goes false and stays false for a
minimum of this same time. This guarantees that the pulse duty cycle never exceeds 50%; the pulse
signal can be inverted with I7mn7 without violating minimum pulse width specifications.
80
Basic Motor Setup
Turbo PMAC User Manual
Turbo PMAC Parameter Setup
Hardware Setup for PMAC2-Style ICs
PFM Clock Frequency: I7m03, I6803, MI903, MI907, MI993
An I-variable controls the frequency of addition of the command value into the accumulator by setting the
frequency of a clock signal called PFMCLK. One addition is performed during each PFMCLK cycle, so
the addition frequency is equal to the PFMCLK frequency. The pulse frequency for a given command
value is directly proportional to this addition frequency. While the default frequency is suitable for
almost all applications, those requiring very high or very low pulse frequencies may need to change this
clock frequency.
This PFMCLK/addition frequency puts an upper limit on the pulse frequency that can be generated – with
an absolute limit of 1/4 of the PFMCLK/addition frequency. Depending on the worst-case frequency
distortion that can be tolerated at high speeds, most people will limit their maximum pulse frequency to
1/10 of the PFMCLK/addition frequency, therefore selecting a PFMCLK/addition frequency 10 to 20
times greater than their maximum desired pulse frequency.
The PFMCLK/addition frequency sets a lower limit on the pulse frequency as well – an absolute limit of
one eight-millionth of the addition frequency (without dithering). The default frequency of approximately
10 MHz can provide a useful range of about 1 Hz to 1 million Hz, and is suitable for a wide variety of
applications, especially with microstepping drives. For full or half step drives, the PFMCLK/addition
frequency probably will be set considerably lower – to the approximately 1.2 MHz or 600 kHz settings.
I7m03 controls the PFMCLK frequency for all of the axis-interface channels on Servo IC m; I6803 does
the same for the two supplementary channels on the handwheel port. On a MACRO Station, MI903,
MI907, and MI997 control this frequency. The input to the clock control circuitry is a 39.3216 MHz
signal; this can be divided by 1, 2, 4, 8, 16, 32, 64, or 128 to create the PFMCLK signal. Therefore, the
possible PFMCLK frequencies and the I-variable values that set them are:
Divide by:
Divider N (1/2^N)
PFMCLK Freq
I-Variable Value*
1
0
39.3216 MHz
2240
2
1
19.6608 MHz
2249
4
2
9.8304 MHz
2258
8
3
4.9152 MHz
2267
16
4
2.4576 MHz
2276
32
5
1.2288 MHz
2285
64
6
611.44 kHz
2294
128
7
305.72 kHz
2303
*SCLK frequency = PFMCLK frequency; ADCCLK and DACCLK frequencies at default
The divider N is used in these I-variables to determine the frequency.
These variables also independently control the frequencies of the encoder sample clock SCLK, plus the
clocks for the serial D/A and A/D converters, DACCLK and ADCCLK, which are also divided down in
the same way from the same 39.3216 MHz signal. The SCLK frequency should be the same as the
PFMCLK frequency if the pulse train is fed into the encoder counters. The above table shows the value
of the I-variable for each possible frequency of the PFMCLCK, assuming the SCLK frequency is set
equal to the PFMCLK frequency, and the DACCLK and ADCCLK frequencies are left at their default
settings.
Basic Motor Setup
81
Turbo PMAC User Manual
PFM Pulse Width: I7m04, I6804, MI904, MI908
I7m04 controls the pulse width for the axis-interface channels on Turbo PMAC PMAC2-style Servo IC
m; I6804 does so for supplementary channels on the handwheel port. MI904, MI908, and MI994 set this
for channels on a MACRO Station. The pulse width is specified in PFMCLK cycles; the range is 1 to 255
cycles.
The minimum gap between pulses is equal to the pulse width, so the minimum pulse cycle period is twice
the pulse width set here. This sets a maximum frequency of the PFM output. If the algorithm asks for a
higher frequency, Turbo PMAC will not produce the requested frequency, and pulses will be skipped.
Output Mode Control: I7mn6, MI916
I7mn6 controls what types of signals are brought out from Servo IC m Channel n’s A, B, & C command
registers; it must be set to 2 or 3 to use the PFM signals from the C register. Node-specific variable
MI916 controls this for a MACRO Station channel.
Output Inversion Control: I7mn7, MI917
I7mn7 controls whether the pulse signals are inverted or not. A value of 0 or 1 means the C pulse is hightrue; a value of 2 or 3 means that it is low true. Node-specific variable MI917 controls this for a MACRO
Station channel.
PFM Direction Inversion Control: I7mn8, MI918
I7mn8 controls the polarity of the PFM direction signal alone (it does not affect the pulse signal). A
value of 0 means positive direction is low; a value of 1 means the negative direction is low. Nodespecific variable MI918 controls this for a MACRO Station channel.
Encoder Decode Control: I7mn0, MI910
I7mn0 controls the source of the position feedback signal and how it is decoded. Node-specific variable
MI910 controls this for a MACRO Station channel. Values of 0 to 7 set up for an external signal wired
into PMAC, with the different values in this range determining how the signal is decoded. One of these
values should be used if there is a real feedback sensor; typically 3 or 7 for times-4 decode of a quadrature
encoder signal.
A value of 8 selects the internally generated PFM signal, and automatically selects the pulse-and-direction
decode for the signal. Note that no external cable is required to feed back the PFM signal. If the pulse
train is modified externally to the Turbo PMAC to meet the limitations of a particular stepper drive,
I7mn0 must be set to 0 or 4 to accept this external pulse-and-direction signal. If there has been no net
inversion of the direction signal, I7mn0 will be set to 0.
For an external feedback signal, the correct setting of I7mn0 should cause the encoder counter to count up
in the direction you desire. It also must match the direction sense of the output; a positive command
value (for instance, with the O10 command) must cause the counter to count up, and a negative command
value (e.g. O-10) must cause the counter to count down. Invert the direction sense of the output with
I7mn8, or by changing the wiring.
Parameters to Set up Basic Motor Operation
Several motor I-variables must be set up properly to use the PFM signals properly. Most of these are
address registers. Typically motor #1 will use the circuits for axis interface 1, and so on, but this is not
absolutely required.
Activation and Mode: Ixx00, Ixx01
Ixx00 must be 1 to activate the software for the motor, and Ixx01 must be 0 to tell Turbo PMAC not to do
the commutation. (The drive does the commutation in this mode.)
82
Basic Motor Setup
Turbo PMAC User Manual
Command Output Address: Ixx02
Ixx02 tells Turbo PMAC where to write the output command value for Motor xx. To use the PFM, the
output must be written to the C command register for the axis interface circuit of the proper number. (The
default is to the A command register.) The addresses of the C command registers for PMAC2-style Servo
ICs are:
PFM Pulse Output Addresses (Y-registers)
IC# - Chan#
0-1
0-2
0-3
0-4
1-1
1-2
1-3
1-4
Ixx02
$078004
$07800C
$078014
$07801C
$078104
$07810C
$078114
$07811C
IC# - Chan#
2-1
2-2
2-3
2-4
3-1
3-2
3-3
3-4
Ixx02
$078204
$07820C
$078214
$07821C
$078304
$07830C
$078314
$07831C
IC# - Chan#
4-1
4-2
4-3
4-4
5-1
5-2
5-3
5-4
Ixx02
$079204
$07920C
$079214
$07921C
$079304
$07930C
$079314
$07931C
IC# - Chan#
6-1
6-2
6-3
6-4
7-1
7-2
7-3
7-4
Ixx02
$07A204
$07A20C
$07A214
$07A21C
$07A304
$07A30C
$07A314
$07A31C
IC# - Chan#
8-1
8-2
8-3
8-4
9-1
9-2
9-3
9-4
Ixx02
$07B204
$07B20C
$07B214
$07B21C
$07B304
$07B30C
$07B314
$07B31C
IC# - Chan#
M0 - 1
M0 - 2
Ixx02
$078414
$07841C
Servo ICs 0 and 1 are on the Turbo PMAC2 itself or on Acc-2E 3U-format stack boards.
Servo ICs 2 – 9 are on Acc-24x2 or Acc-51E boards.
Channels 1 – 4 on odd-numbered Servo ICs are Channels 5 – 8 on the dual-IC boards.
MACRO IC 0 (M0) provides the two pulse channels on the handwheel port.
When using this mode through a MACRO Station, the output must be written to Register 2 of the
appropriate MACRO node, from which it will be copied to the C command register for the matching
channel on the MACRO Station. The addresses of Register 2 for each MACRO node are:
MACRO IC PFM Pulse Output Addresses (Y-registers)
IC# - Node#
0-0
0-1
0-4
0-5
0-8
0-9
0 - 12
0 - 13
Ixx02
$078422
$078426
$07842A
$07842E
$078432
$078436
$07843A
$07843E
IC# - Node#
1-0
1-1
1-4
1-5
1-8
1-9
1 - 12
1 - 13
Ixx02
$079422
$079426
$07942A
$07942E
$079432
$079436
$07943A
$07943E
IC# - Node#
2-0
2-1
2-4
2-5
2-8
2-9
2 - 12
2 - 13
Ixx02
$07A422
$07A426
$07A42A
$07A42E
$07A432
$07A436
$07A43A
$07A43E
IC# - Node#
3-0
3-1
3-4
3-5
3-8
3-9
3 - 12
3 - 13
Ixx02
$07B422
$07B426
$07B42A
$07B42E
$07B432
$07B436
$07B43A
$07B43E
Encoder Conversion Table: The encoder conversion table implemented with variables I8000 – I8191
(MI120 – MI151 on a MACRO Station) does the initial processing of the real or simulated feedback
registers. The default conversion table in Turbo PMAC processes encoder channels with timer-based 1/T
sub-count interpolation, in its first ten entries. These are the ideal settings if real incremental encoder
feedback is used.
However, if the output pulse train is used for simulated feedback, it is best to process the counts without
any sub-count interpolation. This will prevent the PMAC from trying to position between pulses and
create dithering. In the PMAC Executive program Configure Encoder Table window, change the entry
from Incremental with 1/T Interpolation to Incremental with No Interpolation. If writing directly to the
memory location of the table, change the format digit from $0 to $C. The entries for this type of
conversion for the encoder registers of a Turbo PMAC2 system are:
Basic Motor Setup
83
Turbo PMAC User Manual
Encoder Conversion Table Entries, No Interpolation
IC# - Chan#
0-1
0-2
0-3
0-4
1-1
1-2
1-3
1-4
I8xxx
$C78000
$C78008
$C78010
$C78018
$C78100
$C78108
$C78110
$C78118
IC# - Chan#
2-1
2-2
2-3
2-4
3-1
3-2
3-3
3-4
I8xxx
$C78200
$C78208
$C78210
$C78218
$C78300
$C78308
$C78310
$C78318
IC# - Chan#
4-1
4-2
4-3
4-4
5-1
5-2
5-3
5-4
I8xxx
$C79200
$C79208
$C79210
$C79218
$C79300
$C79308
$C79310
$C79318
IC# - Chan#
6-1
6-2
6-3
6-4
7-1
7-2
7-3
7-4
I8xxx
$C7A200
$C7A208
$C7A210
$C7A218
$C7A300
$C7A308
$C7A310
$C7A318
IC# - Chan#
8-1
8-2
8-3
8-4
9-1
9-2
9-3
9-4
I8xxx
$C7B200
$C7B208
$C7B210
$C7B218
$C7B300
$C7B308
$C7B310
$C7B318
IC# - Chan#
M0 - 1
M0 - 2
Ixx02
$C78410
$C78418
Servo ICs 0 and 1 are on the Turbo PMAC2 itself or on Acc-2E 3U-format stack boards.
Servo ICs 2 – 9 are on Acc-24x2 or Acc-51E boards.
Channels 1 – 4 on odd-numbered Servo ICs are Channels 5 – 8 on the dual-IC boards.
MACRO IC 0 (M0) provides the two pulse channels on the “handwheel port.”
Feedback Addresses: Ixx03, Ixx04
Ixx03 tells Turbo PMAC where to look to get its position-loop feedback. Whether real or simulated
feedback is used, this location should be that of a processed encoder counter in Turbo PMAC’s encoder
conversion table. In a typical setup, Motor xx on Turbo PMAC will use the “xth” entry of the conversion
table, with addresses $003501 to $0035C0. These are the default values for Ixx03.
Ixx04 tells Turbo PMAC where to look to get its velocity-loop feedback. For the simulated feedback, this
should be set to the same address as Ixx03 for each motor.
Output (Frequency) Limit: Ixx69
Ixx69 controls the maximum pulse frequency for a given PFMCLK frequency. At the maximum Ixx69
value of 32,767, the maximum pulse frequency would be one-half of PFMCLK. However, because the
shortest pulse cycle is two clock cycles on and two clock cycles off, the maximum pulse frequency can be
25% of the PFMCLK frequency. The equation for Ixx69 is:
Ixx69 =
MaxFreq( kHz , MHz )
PFMCLK ( kHz , MHz )
* 65 ,536
The maximum value of Ixx69 is 25% of 65,536, or 16,384. If Ixx69 is set higher than this value, and the
value written to the PFM register exceeds 16,384, the pulse-generating circuitry will drop pulses and the
frequency will fall.
For example, with PFMCLK at the default of 9.83 MHz, and a desired maximum frequency of 500 kHz,
Ixx69 = (0.5 MHz / 9.83 MHz) * 65,536 = 3,333.
Parameters to Set Up Motor Servo Gains
If using real feedback sensors, the motor should be tuned as a normal servo motor would be tuned. If
feeding the pulse train back into the encoder counter to create a fully electronic loop, the loop response is
predictable, and the tuning gains can be set by formula as explained here. In either case, the control loop
appears to Turbo PMAC to act as a velocity-mode servo drive.
To create a closed-loop position response with a natural frequency of 25 Hz and a damping ratio of 1
(suitable for almost all systems), use the gain settings as calculated in the following sections.
84
Basic Motor Setup
Turbo PMAC User Manual
Proportional Gain: Ixx30
The proportional gain term Ixx30 is set according to the equation:
Ixx30 =
660 ,000
Ixx08 * PFMCLK ( MHz )
For example, with PFMCLK at the default of 9.83 MHz, and Ixx08 at the default of 96, Ixx30 = 660,000 /
(96 * 9.83) = 700.
Derivative Gain: Ixx31
The derivative gain term Ixx31 is set to zero, because the loop behaves like a velocity-loop servo drive,
and there is no need to have the Turbo PMAC add damping.
Velocity Feedforward Gain: Ixx32
The velocity feedforward gain term Ixx32 is set according to the equation:
Ixx32 = 6660 * ServoFreq( kHz )
where ServoFreq is the frequency of the servo interrupt as established by I7m00, I7m01, and I7m02 (or
I6800, I6801, and I6802 on a Turbo PMAC2 Ultralite). For example, with ServoFreq at the default of
2.26 kHz (I7m00=6527, I7m01=0, I7m02=3), Ixx32 = 6660 * 2.26 = 15,050.
If you set Ixx30 differently from what its above formula dictates, Ixx32 should be changed from the value
its above formula dictates in inverse proportion to the change in Ixx30. For instance, if Ixx30 is half of
what is calculated above, Ixx32 should be twice what is calculated above.
Integral Gain and Mode: Ixx33, Ixx34
The integral gain term Ixx33 typically is set to zero because there are no offsets or disturbances to the
digital electronic loop. Some people will use integral gain to force zero steady-state errors, even when
other gains are not well set.
Integration mode Ixx34 is irrelevant if Ixx33 is set to zero. If Ixx33 is used, setting Ixx34 to 1 turns on
the integral gain only when the commanded velocity is zero; setting Ixx34 to 0 turns it on all of the time.
Acceleration Feedforward Gain: Ixx35
The acceleration feedforward gain term Ixx35 typically is set to zero, because the electronic loop has no
inertia to overcome. However, some users will want to use Ixx35 to compensate for the small time delays
created by the addition process in pulse generation that will cause small following errors when the
velocity is changing. For a given servo update time, an optimum Ixx35 value can be found that virtually
eliminates errors at all accelerations.
Notch Filter Parameters: Ixx36 – Ixx39
Notch filter parameters Ixx36 to Ixx39 are set to zero in the open-loop case, because the electronic loop
itself has no resonance, even if the mechanical system does.
Testing the Setup
Preparing for the Test
With an open-loop system, you can test much of the initial operation of the setup without attaching
anything to the Turbo PMAC, because the loop operates entirely within Turbo PMAC. To do this, you
will have to disable your position limits by setting bit 17 of Ixx24 to 1. If no other bits of Ixx24 were set
to 1, this would mean setting Ixx24 to $20000.
The initial test should be simple monitoring of motor position and velocity as commands are given to the
motor. The position-reporting window in the PMAC Executive program provides a very convenient way
to monitor this. Select the window option that displays position, velocity, and following error for the
motor you are testing. Specify the units of the reported position and velocity (counts per second for
velocity equals Hertz).
Basic Motor Setup
85
Turbo PMAC User Manual
Executing the Open-Loop Test
First, use the “O” (open-loop output) command to verify the operation of the frequency generator. This
command simply places a value proportional to the magnitude of the command into the output register.
This should generate a constant frequency output from the pulse generator that shows up as a constantly
changing position and a steady non-zero (for non-zero commands) velocity in the reporting window. Do
not worry about exact values now; just see if a change can be created. An O10 command should create a
positive velocity and a position counting in the positive direction; an O-10 command should create a
negative velocity and a position counting in the negative direction.
Troubleshooting the Open-Loop Test
If this result is not received, look at the following:
1. Use the Why Am I Not Moving? screen in the Executive program to see if there is some reason Turbo
PMAC firmware is not permitting the command. For instance, the overtravel limit inputs may need
to be held low, or the limit function disabled with Ixx24.
2. If there is an oscilloscope or frequency counter, see if a pulse train is being received on the output pin.
This will help to isolate the problem. If the pulse train is not received with a non-zero O command,
and no pulse train with an O0 command, pulses are being generated but not feeding back properly. If
there is no pulse train, pulses are not being generated properly.
3. Assign an M-variable to the output command register pointed to by Ixx02, double-checking to be sure
that that this address is one of the Output Command C registers that can drive a PFM circuit. For
example, define M102->Y:$078004,8,16,S to look at the Output Command Register 1C. Now
monitor the M-variable in the Executive program’s Watch window or in the Terminal window as O
commands are being issued. The value should change. The reported value should be equal to Ixx69
* (O-command magnitude) / 100.
Note:
The command register is really a 24-bit register. By assigning the M-variable to
the top 16 bits only, the reported value is in the units of the Ixx69 limit.
If changes in the reported M-variable are not being received, recheck Ixx02 and the Why Am I Not
Moving? window. If changes are being received here, next check variable I7mn6 for the machine
interface channel n that is being used to make sure the pulse and direction signals are being permitted
to be output.
4. To check the feedback path, first look at I7mn0 for the machine interface channel n being used. It
should be set to 8 to take the internal pulse and direction signal into encoder counter n, or to 0 or 4 for
external pulse and direction. Next, look at the setup of the Encoder Conversion Table under the
Configure menu of the Executive program. It should contain an entry specifying the quadrature
conversion of the address of the encoder channel (listed above) that you are using, possibly with 1/T
interpolation, but preferably without. Check the actual hexadecimal address value (e.g. $78000 for
Servo IC 0 Encoder 1, $78008 for Servo IC 0 Encoder 2). Note the address of the result register in
the table (e.g. $003501 for the first entry).
5. Now check Ixx03 and Ixx04. They should contain the address of the result register from the
conversion table that was just noted above. This tells the motor to use the processed value in the
conversion table as its feedback.
Once the presence of a pulse train and its feedback into the encoder counter and the motor position
registers have been verified, check to see if the correct frequency range has been reached. Issue an O100
command. Check the frequency by looking at the reported velocity in the window, and/or by examining
the output pulse train on an oscilloscope or frequency counter. Check the frequency at several other
positive and negative command values (e.g. O50, O-50, O-100) so that the frequency is proportional to
the command and covers the range needed.
86
Basic Motor Setup
Turbo PMAC User Manual
If not getting the proper frequency range, double check the setting of I7m03 that sets the PFMCLK
frequency. Also check the value of Ixx69 that determines the maximum frequency (O100 frequency) at
this PFMCLK frequency.
Executing the Closed-Loop Test
Next, close the loop with a J/ command. The reported position should hold steady, and the reported
velocity should be zero. Set up yotheur jogging I-variables Ixx19 to Ixx22 to get the speeds and
accelerations needed. Issue a J+ command; count up at the rate specified by Ixx22 (watch the units).
Issue a J- command; count down at this same rate.
Now that what PMAC considers to be closed-loop moves are being executed, the servo loop gain
parameters are important. The easiest way to monitor performance is with the position window in the
PMAC Executive Program, configured to display position, velocity, and following error for the current
motor. More detailed analysis can be done with the data gathering plots.
Troubleshooting the Closed-Loop Test
When troubleshooting these jogging moves, it is important to note what PMAC thinks the motor is doing
based on the pulse feedback compared to what the motor is actually doing. For example, if the motor has
stopped, but you see through the position window that PMAC keeps counting position, the motor has
probably stalled. Possibilities here include excessive velocity command, excessive load, and resonance
problems. However, if PMAC is also reporting a stop, something in the PMAC simulated loop has failed,
probably causing a shutdown on excessive following error. The main possibilities here are values set too
low for Ixx30 proportional gain, Ixx32 velocity feedforward, or Ixx69 output limit.
Basic Motor Setup
87
Turbo PMAC User Manual
88
Basic Motor Setup
Turbo PMAC User Manual
SETTING UP TURBO PMAC-BASED COMMUTATION AND/OR
CURRENT LOOP
This section provides detailed instructions for the step-by-step manual setup of motor phase commutation
and/or digital current-loop closure within the Turbo PMAC. Few users will do these steps manually;
most will use the automated procedures of the Turbo Setup program on the PC, even for the setup of the
first unit. The instructions in this section are for the user who wants a full understanding of the Turbo
PMAC algorithms and how they are set up for a particular application.
Beginning Setup of Commutation
If Turbo PMAC is to perform the commutation of a motor, it must do more than simply close the
position/velocity-loop servo for the motor. Several parameters must be set up correctly to configure the
commutation.
The first steps in setting up the commutation are common whether Turbo PMAC is performing the current
loop closure (direct PWM output mode) or not (sine-wave output mode). These steps are described in this
section. The next steps differ based on which mode is used; these are described in the next two sections,
only one of which is used for a particular motor. Finally, the last steps in the setup of commutation are
again common to the two modes of operation; these are described in the following section.
Commutation Enable: Ixx01
If Turbo PMAC is performing the commutation for Motor xx using a directly addressed (not over
MACRO) Servo IC encoder register, Ixx01 must be set to 1 to enable the commutation and use an Xregister for commutation feedback (as addressed by Ixx83). If Turbo PMAC is performing the
commutation for Motor xx using a Servo IC on a MACRO Station, Ixx03 must be set to 3 to enable the
commutation and use a Y-register for commutation feedback. (If Turbo PMAC is performing the digital
current loop closure, it must also perform the phase commutation for the motor.)
Note:
Direct PWM control of brush motors with digital current loop utilizes Turbo
PMAC’s commutation algorithms even though the motor does not require
electronic commutation; Ixx01 must be set to 1 or 3 for this case.
Commutation Cycle Size: Ixx70 & Ixx71
Ixx70 and Ixx71 define the size of the commutation cycle (electrical cycle). The cycle is equal to Ixx71
divided by Ixx70, expressed in encoder counts (after decode). Ixx70 and Ixx71 must both be integers, but
the ratio Ixx71/Ixx70 does not have to be an integer. On a rotary motor, typically Ixx71 is set to the
number of counts per mechanical revolution, and Ixx70 is set to the number of pole-pairs (half of the
number of poles) for the motor, which is equal to the number of commutation cycles per mechanical
revolution.
When commutating across the MACRO ring with the Type 1 protocol used by Delta Tau MACRO
products, the standard position feedback reported in Register 0 is in units of 1/32 count. If this is used as
the commutation position feedback, as specified by Ixx83, the units of Ixx71/Ixx70 must also be in 1/32
of a count. For example, if a 4-pole (2-electrical-cycle) motor has a 2000-line (8000-count) encoder,
instead of setting Ixx70 to 2 and Ixx71 to 8000, Ixx70 should be set to 2 and Ixx71 should be set to
256,000.
When performing direct PWM control of brush motors, the commutation algorithm must be fooled to
create DC output instead of its usual AC output. This is best done by setting Ixx70 and Ixx71 to 0.
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
89
Turbo PMAC User Manual
Commutation Feedback Address: Ixx83
Ixx83 specifies the address of the register used for the commutation position feedback. The default
values of Ixx83 use the encoder register for the machine interface channel usually assigned to the motor.
Ixx83 seldom needs to be changed from the default value for PWM applications. When using on-board
encoder counter, these are X-registers, so Ixx01 must be set to 1, not 3, to use these. Typically, the values
used are:
PMAC-Style Servo IC Commutation Position Feedback Addresses (X-registers)
IC# - Chan#
Ixx83
IC# - Chan#
Ixx83
IC# - Chan#
Ixx83
IC# - Chan#
Ixx83
IC# - Chan#
Ixx83
0-1
$078001
2-1
$078201
4-1
$079201
6-1
$07A201
8-1
$07B201
0-2
$078005
2-2
$078205
4-2
$079205
6-2
$07A205
8-2
$07B205
0-3
$078009
2-3
$078209
4-3
$079209
6-3
$07A209
8-3
$07B209
0-4
$07800D
2-4
$07820D
4-4
$07920D
6-4
$07A20D
8-4
$07B20D
1-1
$078101
3-1
$078301
5-1
$079301
7-1
$07A301
9-1
$07B301
1-2
$078105
3-2
$078305
5-2
$079305
7-2
$07A305
9-2
$07B305
1-3
$078109
3-3
$078309
5-3
$079309
7-3
$07A309
9-3
$07B309
1-4
$07810D
3-4
$07830D
5-4
$07930D
7-4
$07A30D
9-4
$07B30D
Servo ICs 0 and 1 are on the Turbo PMAC itself.
Servo ICs 2 – 9 are on Acc-24P/V or Acc-51P boards.
Channels 1 – 4 on odd-numbered Servo ICs are Channels 5 – 8 on the boards.
PMAC2-Style Servo IC Commutation Position Feedback Addresses (X-registers)
IC# - Chan#
0-1
0-2
0-3
0-4
1-1
1-2
1-3
1-4
Ixx83
$078001
$078009
$078011
$078019
$078101
$078109
$078111
$078119
IC# - Chan#
2-1
2-2
2-3
2-4
3-1
3-2
3-3
3-4
Ixx83
$078201
$078209
$078211
$078219
$078301
$078309
$078311
$078319
IC# - Chan#
4-1
4-2
4-3
4-4
5-1
5-2
5-3
5-4
Ixx83
$079201
$079209
$079211
$079219
$079301
$079309
$079311
$079319
IC# - Chan#
6-1
6-2
6-3
6-4
7-1
7-2
7-3
7-4
Ixx83
$07A201
$07A209
$07A211
$07A219
$07A301
$07A309
$07A311
$07A319
IC# - Chan#
8-1
8-2
8-3
8-4
9-1
9-2
9-3
9-4
Ixx83
$07B201
$07B209
$07B211
$07B219
$07B301
$07B309
$07B311
$07B319
Servo ICs 0 and 1 are on the Turbo PMAC2 itself or on Acc-2E 3U-format stack boards.
Servo ICs 2 – 9 are on Acc-24x2 or Acc-51E boards.
Channels 1 – 4 on odd-numbered Servo ICs are Channels 5 – 8 on dual-Servo-IC boards.
When performing commutation over the MACRO ring, Ixx83 points to the position feedback register
(Register 0) for the node used. These are Y-registers, so Ixx01 must be set to 3, not 1, to use these. The
values used in the MACRO Type 1 protocol expected by Delta Tau MACRO products are:
MACRO Node Commutation Position Feedback Addresses (Y-registers)
IC# - Node#
0-0
0-1
0-4
0-5
0-8
0-9
0 - 12
0 - 13
Ixx83
$078420
$078424
$078428
$07842C
$078430
$078434
$078438
$07843C
IC# - Node#
1-0
1-1
1-4
1-5
1-8
1-9
1 - 12
1 - 13
Ixx83
$079420
$079424
$079428
$07942C
$079430
$079434
$079438
$07943C
IC# - Node#
2-0
2-1
2-4
2-5
2-8
2-9
2 - 12
2 - 13
Ixx83
$07A420
$07A424
$07A428
$07A42C
$07A430
$07A434
$07A438
$07A43C
IC# - Node#
3-0
3-1
3-4
3-5
3-8
3-9
3 - 12
3 - 13
Ixx83
$07B420
$07B424
$07B428
$07B42C
$07B430
$07B434
$07B438
$07B43C
If using the older Type 0 MACRO protocol, add 3 to the value shown in the above table (e.g. $078420
becomes $078423).
90
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
Turbo PMAC User Manual
Current Loop in Turbo PMAC or Not
Turbo PMAC can perform commutation for a motor with or without closing the current loop for the
motor phases. If the current loops are closed in the Turbo PMAC, the outputs from Turbo PMAC are
phase voltage commands, usually represented as pulse-width-modulated (PWM) digital outputs. This
technique is called direct PWM control. If the current loops are not closed in the Turbo PMAC, the
outputs from Turbo PMAC are phase current commands, usually represented as +/-10V analog voltage
signals. This technique is called sine-wave output control.
Ixx82 controls which mode of operation is used. If Ixx82 is greater than zero, Turbo PMAC will close
the current loops for Motor xx, and Ixx82 specifies the address of the current feedback for the motor.
Setup for this mode is covered in the next section, Setting Up for Direct PWM Control.
If Ixx82 is set to 0, Turbo PMAC will not close the current loops for Motor xx. Setup for this mode is
covered in the following section, Setting Up for Sine-Wave Output Control.
Setting Up for Direct PWM Control
This section explains how to set up Turbo PMAC for direct PWM control of amplifiers. In this mode,
Turbo PMAC performs all of the control tasks for the motor, including commutation and digital current
loop closure. The amplifier performs only the power conversion task. In this mode, Turbo PMAC
outputs PWM voltage commands for each phase of the motor. If not using Turbo PMAC to perform this
task for any of the motors, skip this section. Simply make sure that Ixx82 is set to 0 for all of the
activated motors, so Turbo PMAC will not try to close the current loop for them.
Direct PWM control can be performed only using PMAC2-style Servo ICs, either directly commanded
from the CPU, or over the MACRO ring. It is possible, but rare, to do this control through accessory
boards when the base controller is a Turbo PMAC.
Introduction
In this mode, the current control loop is closed by using digital computation operating on numerical
values in registers rather than by using analog processing operating on voltage levels with op-amps. The
digital techniques bring many advantages: there is no need for pot tweaking or personality modules; there
is no drift over temperature or time; computer analysis and auto-tuning are possible; gain values are easily
stored for backup and copying onto other systems; and adaptive techniques are possible.
When performing digital current loop closure on the Turbo PMAC, several hardware and software
features must be set up properly to utilize the digital current loop and direct PWM outputs correctly. The
following section details how to perform this setup manually. However, typically, these steps are
automated through the use of the “Turbo Setup” program running on a PC and communicating with the
Turbo PMAC.
Note:
The instructions given in this section are for the first-time setup with an otherwise
unknown interface. For a given interface to the drive, motor and feedback device,
many parts of the setup will be taken from a list, and will not have to be tested, or
tested as thoroughly as described in this section. A list of configuration-specific
settings should come with the manual for each particular interface or drive.
Subsequent versions of the same setup should be even easier.
Digital Current Loop Principle of Operation
Traditionally, motor phase current loops have been closed in analog fashion, with op-amp circuits
creating phase voltage commands from the difference between commanded and actual phase current
signal levels. These analog phase voltage commands are converted to PWM format through analog
comparison to a saw tooth waveform.
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
91
Turbo PMAC User Manual
Turbo PMAC permits digital closure of the motor current loops, mathematically creating phase voltage
commands from numerical registers representing commanded and actual current values. These numerical
phase voltage commands are converted to PWM format through digital comparison to an up/down
counter that creates a digital saw tooth waveform. The analog current measurements must be converted
to digital form with ADCs before the loop can be closed.
PMAC2 Digital PWM Generation (per phase)
New command loaded
at counter turnaround
Ix66
I900 + 1
[I7m00] + 1
PWM
Command
Range
PWM
Counter
Range
PWM Command
Time
4 x I900 [I7m00] + 6
PWM_CLK Cycles
I900 - 2
[I7m00] - 2
Counter Clock is 120 MHz
PWM
Up/Down
Counter
Servo
Interrupt
generated
here
- Ix66
PWM
Top Signal
PWM
Bottom Signal
Deadtime
(I904 [I7m04])
By directly commanding the on-off states of the power transistors in this manner, Turbo PMAC
minimizes the calculation and transport delays in the servo loop. This permits the use of higher gains,
which in turn permit greater stiffness, acceleration, and disturbance rejection. Also, digital techniques
permit the use of mathematical transformations of the current-loop data, turning measured AC quantities
into DC quantities for loop closure. This technique, explained in the next section, significantly improves
high-speed performance by minimizing high-frequency problems.
Frames of Reference
A very important advantage of the digital current loop is its ability to close the current loops in the field
frame. To understand this advantage, some basic theoretical background is required.
In a motor, there are three frames of reference that are important. The first is the stator frame” which is
fixed on the non-moving part of the motor, called the stator. In a brushless motor, the motor armature
windings are on the stator, so they are fixed in the stator frame.
The second frame is the rotor frame, which is referenced to the mechanics of the moving part of the
motor, called the rotor. This frame, of course, rotates with respect to the stator. For linear brushless
motors, this is actually a translation, but because it is cyclic, we can think of it as a rotation.
The third frame is the field frame which is referenced to the magnetic field orientation of the rotor. In a
synchronous motor such as a permanent-magnet brushless motor, the field is fixed on the rotor, so the
field frame is the same as the rotor frame. In an asynchronous motor such as an induction motor, the field
slips with respect to the rotor, so the field frame and rotor frame are separate.
Working in the Field Frame
The physics of motor operation are best understood in the field frame. A current vector in the stator that
is perpendicular to the rotor field (that is, current in the stator that produces a magnetic field perpendicular
to the rotor magnetic field) produces torque. This component of the stator current is known as quadrature
current. The output of the position/velocity loop servo algorithm is the magnitude of the commanded
quadrature current. For diagnostic purposes on a Turbo PMAC, an “O” command can be used to set a
fixed quadrature current command.
92
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
Turbo PMAC User Manual
A current vector in the stator that is parallel to the rotor field induces current in the rotor that changes the
magnetic field strength of the rotor (when the stator and rotor field are rotating relative to each other).
This component of the stator current is known as direct current. For an induction motor, this is required
to create a rotor magnetic field. For a permanent-magnet brushless motor, the rotor magnets always
produce a field, so direct current is not required, although it can be used to modify the magnetic field
strength. On Turbo PMAC, parameter Ixx77 for Motor xx determines the magnitude of the direct current.
Analog Loops in the Stator Frame
In an amplifier with an analog current loop, the closure of the loops on the stator windings must be closed
in the stator frame, because the current measurements are in the stator frame, and analog circuitry has no
practical way to transform these. In such a system, the current commands must be transformed from the
field frame in which they are calculated to the stator frame, and converted to voltage levels representing
the individual stator phase current commands. These are compared to other voltage levels representing
the actual stator phase current measurements.
As the motor is rotating, and/or the field is slipping, these current values, command and actual, are AC
quantities. Overall loop gain, and therefore system performance, is reduced at high frequencies (high
speeds). The back EMF phase voltage, which acts as a disturbance to the current loop, is also an AC
quantity. The current loop integral gain or lag filter, which is supposed to overcome disturbances, falls
well behind at high frequencies.
PMAC / PMAC2 Commutation with Analog Current Loop
DC
Magnetization
Current
Ix77
Torque Command
from Position/
Velocity Servo
AC
Id’
Ia’
+
DAC
Direct Current
Command
-
Ia
dq
to
abc
Ib’
+
DAC
-
Ib
Vb
PI
Vb
PWM
E
Θe
Phase B Current Loop
Ix78
Va
PWM
Phase A Current Loop
Iq’
Quadrature
Current Command
Va
PI
SG
-
-
+
Vc
-
PWM
Vc
Balance Loop
1/s
Θr
+
+
PMAC
Amplifier
Digital Loops in the Field Frame
In a system with a digital current loop, it is possible to close the current loops in the field frame (not all
such systems do, however). Instead of the current commands being transformed from field frame to stator
frame before the loop is closed, the actual current measurements are transformed from stator frame to
field frame. In the field frame, a direct-current loop is closed, and a quadrature current loop is closed.
This produces a direct-voltage command and a quadrature-voltage command; these are transformed back
into the stator frame to become phase-voltage commands, which are implemented as PWM values.
The direct and quadrature current values are DC quantities, even as the motor is rotating and the field is
slipping. Therefore, the high-frequency limitations of the servo loop are irrelevant. This provides
significantly improved high-speed performance from identical motors and power electronics.
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
93
Turbo PMAC User Manual
Turbo PMAC has a PI (proportional-integral) digital current loop. There is only one set of gains, which
serves for both the direct current loop and the quadrature current loop. Tuning is best done on the direct
current loop, because this will generate no torque, and therefore no movement. The current-loop autotuner in the PMAC Executive program uses the direct current loop to tune. This is valid even for
permanent-magnet brushless motors in which no direct current will be used in the actual application.
Current loop performance is not load-dependent, so the motor does not need to be attached to the load
during the tuning process. For position/velocity loop tuning, the load does need to be attached.
PMAC / PMAC2 Commutation with Digital Current Loop
DC
Magnetization
Current
Ix77
Id +
Direct Current
Command
PI
AC
Vd
+
-
-
dq
to
abc
P
Vq
Va
PWM
Va
Vb
PWM
Vb
Vc
PWM
Vc
E
Direct Current Loop
Torque Command
Iq +
from Position/
Velocity Servo
Quadrature
Current
Command
PI
Θ
+
Id
-
-
Ia
abc
to
dq
P
Iq
Θe
Quadrature Current Loop
SG
Ix78
1/s
+
Ib
ADC
ia
ADC
ib
Θr
+
PMAC
Amplifier
Hardware Setup
The connection between Turbo PMAC and the direct PWM digital amplifier is made through a Mini-D 36pin connector with a standard pin-out defined by Delta Tau. This connector contains the PWM command
signals, the ADC feedback signals, and the amplifier-enable and fault handshake signals. Usually, this
connector is found on an Acc-8F breakout board for board-level Turbo PMAC2s, on Acc-24E2 PWM axisinterface boards for the UMAC and on the front panel of a QMAC with the PWM option ordered.
Position feedback, from an encoder or possibly a resolver, typically comes directly back to the Turbo
PMAC; there is no need for position feedback in the drive in this mode of operation.
For a three-phase motor, whether delta-wound or (more commonly) Y-wound, each of the three phase
leads of the motor is connected to one of the phase outputs of the amplifier. Exchanging any two of the
three leads switches the direction sense of the outputs; proper matching of this to feedback will be
evaluated in a setup test. Rotating the three motor leads with respect to the amplifier outputs changes the
commutation cycle’s zero point with respect to any absolute sensor’s zero point.
For a DC brush motor driven from a 3-phase power block amplifier, the two motor leads must be
connected between the first and third amplifier outputs, corresponding to the A and C outputs from the
Turbo PMAC.
Turbo PMAC Parameter Setup
Much of the Turbo PMAC interface hardware is software-configurable through I-variables. This section
provides basic information on each of the I-variables that is important in this type of application. There is
a detailed description of each I-variable in the Software Reference Manual.
94
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
Turbo PMAC User Manual
Parameters to Set up Global and Multi-Channel Hardware Signals
PWM Frequency Control: I7m00, MI900, MI906
If you are driving the axes directly (not over the MACRO ring), set Turbo PMAC I-variable I7m00 to
define the PWM frequency for the 4 channels on PMAC2-style Servo IC m according to the equation:

I 7 m00 = int 
117 ,964.8 kHz
 4 * PWMFreq( kHz )

− 1

If the axes are being driven from a MACRO Station, MI900 on the Station controls the PWM frequency
of the first four channels on the Station according to the same equation; MI906 does the same for the
second four channels on the Station.
Set the frequency within the specified range for the drives controlled by Turbo PMAC. Too high a
frequency can lead to excessive drive heating due to switching losses; too low a frequency can lead to
lack of responsiveness, excess acoustic noise, possible physical vibration, and excessive motor heating
due to high current ripple.
The PWM frequency for any set of channels must have a definite relationship to the phase clock
frequency. If the channels are driven by the same Servo IC that is generating the phase and servo clock,
this relationship is set automatically. If the PWM frequency for channels on another Servo IC is the
same, this relationship also will hold automatically.
The PWM frequency on other channels does not have to be the same as the frequency of those channels
on the Servo IC generating the phase clock, but they do have to have a synchronous relationship with the
phase clock. The following relationship must hold for proper direct-PWM operation of other channels:
2*
PWMFreq
PhaseFreq
= { Integer }
If a Servo IC is used to generate the Phase and Servo clocks, I7m00 for that IC also sets the frequency of
the MaxPhase clock to twice the PWM frequency for the channels on that IC. (On a Turbo PMAC2
Ultralite board, I6800 for MACRO IC 0 determines the MaxPhase frequency.) The MaxPhase clock is
the highest frequency at which Turbo PMAC’s phase update tasks, which include phase commutation and
digital current loop closure, can operate. Note that any change to this I7m00 automatically changes the
Phase and Servo clock frequencies.
Hardware Clock Frequency Control: I7m03, MI903, MI907
I7m03 determines the frequency of four hardware clock signals used for the machine interface channels
on Servo IC m. Probably these can be left at the default values. The four hardware clock signals are
SCLK (encoder sample clock), PFMCLK (pulse frequency modulator clock, DAC_CLK (digital-toanalog converter clock), and ADCCLK (analog-to-digital converter clock).
If the axes are being driven from a MACRO Station, MI903 on the Station controls the hardware clock
frequencies of the first four channels on the Station according to the same equations; MI907 does the
same for the second four channels on the Station.
Only the ADCCLK signal is used directly with the digital current loop, to control the frequency of the
serial data stream from the current-loop ADCs. The ADC clock frequency must be at least 96 times
higher than the PWM frequency, but it must be within the capability of the serial ADCs. Refer to the
I7m03 description for detailed information on setting these variables.
The encoder SCLK frequency should be at least 20% greater than the maximum count (edge) rate that is
possible for the encoder on any axis. Higher SCLK frequencies than this minimum may be used, but
these make the digital delay anti-noise filter less effective.
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
95
Turbo PMAC User Manual
PWM Deadtime Control: I7m04, MI904, MI908
I7m04 determines the PWM deadtime between top and bottom signals for the machine interface channels
on Servo IC m. I7m04 has a range of 0 to 255, and the deadtime is 0.135 µsec times the I-variable value.
The deadtime should not be set smaller than the recommended minimum for the drive, or excessive drive
heating could occur. Too large a deadtime value can cause unresponsive performance. The default value of
15, which produces a deadtime of 2.0 µsec, is large enough to protect most drives, and small enough not to
create unresponsive performance unless PWM frequencies are extremely high. Some high-power drives
operating from a 480VAC supply will require about 3 µsec of deadtime, for an I7m04 setting of about 23.
If the axes are being driven from a MACRO Station, MI904 on the Station controls the hardware clock
frequencies of the first four channels on the Station according to the same equations; MI908 does the
same for the second four channels on the Station.
ADC Strobe Word: I7m06, MI940, MI941
I7m06 defines the 24-bit strobe word for all of the A/D converters interfaced to the DSPGATE1 PMAC2style Servo IC. This word is shifted out, MSB first, each phase cycle to start the conversion of the A/D
converters. The default value of $FFFFFE is suitable for the ADCs in most direct-PWM amplifiers.
The D revision of the DSPGATE1 IC, introduced in 2002, uses bit 0 (the LSB) of this word as a modecontrol bit for formatting the serial data from the ADCs. If this bit is set to 1, the IC can accept up to four
bits of header data on the data stream and roll it over to the lowest bits of the ADC register where
numerical data is not expected. The ADCs in Delta Tau’s Geo Direct PWM drives, introduced late in
2002, have this header data, and so require this bit to be set.
When using the Turbo PMAC with the Geo Direct PWM amplifiers, the I7m06 ADC strobe word in the
Servo IC must be set to $1FFFFF, setting this mode-control bit to 1, and expecting only 1 header bit.
If the axes are being driven from a MACRO Station, MI940 on the Station controls the ADC strobe word
for the first four channels on the Station in this same way; MI941 does the same for the second four
channels on the Station.
Parameters to Set Up Per-Channel Hardware Signals
For each machine interface channel n (n = 1 to 4) on Servo IC m used for PWM outputs, a few channelspecific I-variables must be set up properly. On a Turbo PMAC system itself, these variables are named
as I7mnp, where m specifies the Servo IC number, n specifies the channel number, and p specifies the
parameter number. On a MACRO Station these variables are named as “node-specific” MI-variables
MI91p (addressed from the Turbo PMAC as MS{node},MI91p), where p is the parameter number.
Output Mode Control: I7mn6, MI916
I7mn6 (node-specific variable MI916 on a MACRO Station) must be set to 0 to specify that all 3 outputs
A, B, and C be in PWM format for a 3-phase motor.
Parameters to Set Up Motor Operation
Several I-variables must be set up for each Motor xx to enable and configure the digital current loop for
that motor. Of course, Ixx00 must be set to 1 for any active motor, regardless of whether digital current
loop is used for that motor or not.
Command Output Address: Ixx02
Ixx02 instructs Turbo PMAC where to place its output commands for Motor xx by specifying the address.
The default values of Ixx02 for Turbo PMACs with on-board PMAC2-style Servo ICs use the PWM
registers A, B, and C for the machine interface channel usually assigned to the motor. (PMAC-style
Servo ICs cannot be used for direct-PWM control.) Ixx02 seldom needs to be changed from the default
value for PWM applications. The actual address specified is that of the PWM A register; Turbo PMAC
then writes to the B and C registers automatically as well. Typically, the values used are:
96
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
Turbo PMAC User Manual
PWM Local Command Output Addresses (Y-registers)
IC# - Chan#
0-1
0-2
0-3
0-4
1-1
1-2
1-3
1-4
Ixx02
$078002
$07800A
$078012
$07801A
$078102
$07810A
$078112
$07811A
IC# - Chan#
2-1
2-2
2-3
2-4
3-1
3-2
3-3
3-4
Ixx02
$078202
$07820A
$078212
$07821A
$078302
$07830A
$078312
$07831A
IC# - Chan#
4-1
4-2
4-3
4-4
5-1
5-2
5-3
5-4
Ixx02
$079202
$07920A
$079212
$07921A
$079302
$07930A
$079312
$07931A
IC# - Chan#
6-1
6-2
6-3
6-4
7-1
7-2
7-3
7-4
Ixx02
$07A202
$07A20A
$07A212
$07A21A
$07A302
$07A30A
$07A312
$07A31A
IC# - Chan#
8-1
8-2
8-3
8-4
9-1
9-2
9-3
9-4
Ixx02
$07B202
$07B20A
$07B212
$07B21A
$07B302
$07B30A
$07B312
$07B31A
Servo ICs 0 and 1 are on the Turbo PMAC2 itself or on Acc-2E 3U-format stack boards.
Servo ICs 2 – 9 are on Acc-24x2 or Acc-51E boards.
Channels 1 – 4 on odd-numbered Servo ICs are Channels 5 – 8 on dual-Servo-IC boards.
When performing direct PWM control over the MACRO ring, Ixx02 points to the first (Register 0) of a
set of three MACRO output registers (0, 1, and 2) for the MACRO IC and the node used. The values
used in the MACRO Type 1 protocol expected by Delta Tau MACRO products are:
PWM MACRO Command Output Addresses (Y-registers)
IC# - Node#
0-0
0-1
0-4
0-5
0-8
0-9
0 - 12
0 - 13
Ixx02
$078420
$078424
$078428
$07842C
$078430
$078434
$078438
$07843C
IC# - Node#
1-0
1-1
1-4
1-5
1-8
1-9
1 - 12
1 - 13
Ixx02
$079420
$079424
$079428
$07942C
$079430
$079434
$079438
$07943C
IC# - Node#
2-0
2-1
2-4
2-5
2-8
2-9
2 - 12
2 - 13
Ixx02
$07A420
$07A424
$07A428
$07A42C
$07A430
$07A434
$07A438
$07A43C
IC# - Node#
3-0
3-1
3-4
3-5
3-8
3-9
3 - 12
3 - 13
Ixx02
$07B420
$07B424
$07B428
$07B42C
$07B430
$07B434
$07B438
$07B43C
If using the older Type 0 MACRO protocol, add 1 to the value shown in the above table (e.g. $078420
becomes $078421).
Current Feedback Address: Ixx82
Ixx82 instructs Turbo PMAC where to look for its current feedback values for Motor xx. It also acts as
the variable that tells Turbo PMAC whether or not to perform current-loop closure itself. If Ixx82=0 (the
default), Turbo PMAC will not execute the current loop for Motor xx; any current loop must be executed
in the amplifier. If Ixx82>0, Turbo PMAC will look at the Y-register whose address is specified by
Ixx82, and for a multi-phase motor, the next lower addressed register, to get the current feedback
information for its current loop.
Usually, the registers specified are the serial ADC shift registers in Turbo PMAC’s Servo IC or the
matching registers in a MACRO IC that have brought the current information over the ring. The actual
address specified is that of the ADC B register; Turbo PMAC then automatically reads from the A
register as well. Typically, the values used when commanding the axes directly (not over the MACRO
ring) are:
On-Board ADC Current-Feedback Addresses (Y-registers)
IC# - Chan#
0-1
0-2
0-3
0-4
1-1
1-2
1-3
1-4
Ixx82
$078006
$07800E
$078016
$07801E
$078106
$07810E
$078116
$07811E
IC# - Chan#
2-1
2-2
2-3
2-4
3-1
3-2
3-3
3-4
Ixx82
$078206
$07820E
$078216
$07821E
$078306
$07830E
$078316
$07831E
IC# - Chan#
4-1
4-2
4-3
4-4
5-1
5-2
5-3
5-4
Ixx82
$079206
$07920E
$079216
$07921E
$079306
$07930E
$079316
$07931E
IC# - Chan#
6-1
6-2
6-3
6-4
7-1
7-2
7-3
7-4
Ixx82
$07A206
$07A20E
$07A216
$07A21E
$07A306
$07A30E
$07A316
$07A31E
IC# - Chan#
8-1
8-2
8-3
8-4
9-1
9-2
9-3
9-4
Ixx82
$07B206
$07B20E
$07B216
$07B21E
$07B306
$07B30E
$07B316
$07B31E
Servo ICs 0 and 1 are on the Turbo PMAC2 itself or on Acc-2E 3U-format stack boards.
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
97
Turbo PMAC User Manual
Servo ICs 2 – 9 are on Acc-24x2 or Acc-51E boards.
Channels 1 – 4 on odd-numbered Servo ICs are Channels 5 – 8 on dual-Servo-IC boards.
When performing direct PWM control over the MACRO ring, Ixx82 points to the second of a set of two
MACRO input registers for the MACRO IC and the node used. Typically, the values used are the same
whether MACRO Type 0 or Type 1 protocol is employed for the node:
MACRO Node ADC Current-Feedback Addresses (Y-registers)
IC# - Node#
0-0
0-1
0-4
0-5
0-8
0-9
0 - 12
0 - 13
Ixx82
$078422
$078426
$07842A
$07842E
$078432
$078436
$07843A
$07843E
IC# - Node#
1-0
1-1
1-4
1-5
1-8
1-9
1 - 12
1 - 13
Ixx82
$079422
$079426
$07942A
$07942E
$079432
$079436
$07943A
$07943E
IC# - Node#
2-0
2-1
2-4
2-5
2-8
2-9
2 - 12
2 - 13
Ixx82
$07A422
$07A426
$07A42A
$07A42E
$07A432
$07A436
$07A43A
$07A43E
IC# - Node#
3-0
3-1
3-4
3-5
3-8
3-9
3 - 12
3 - 13
Ixx82
$07B422
$07B426
$07B42A
$07B42E
$07B432
$07B436
$07B43A
$07B43E
If using the older Type 0 MACRO protocol, add 1 to the value shown in the above table (e.g. $078422
becomes $078423).
Current Feedback Mask Word: Ixx84
Ixx84 specifies a mask word to tell Turbo PMAC what bits of the registers specified by Ixx82 are to be
used in the current-loop algorithm. This permits the use of ADCs of various resolutions. It also permits
use of the rest of the 18-bit ADC shift register for other information, such as fault codes. Ixx84 is a 24-bit
value that is combined with the feedback word in a bit-by-bit AND operation. The default value of
$FFF000 specifies that the top 12 bits of the feedback words are to be used. Most direct-PWM amplifiers
use 12-bit ADCs, so $FFF000 is the appropriate value for these amplifiers.
PWM Scale Factor: Ixx66
Ixx66, the PWM Scale Factor, scales the output command values so that they use the PWM circuitry
effectively. The result of the current-loop calculations is a fractional value between -1.0 and +1.0. This
value is multiplied by Ixx66 before being written to a PWM command register, where it is digitally
compared to a PWM up/down counter moving between I7m00+1 and -I7m00-2. To utilize the dynamic
range of the PWM circuitry well, Ixx66 should be set slightly greater than I7m00. Typically a value 10%
greater is used, permitting full-on conditions at maximum command values over about 1/6 of the
commutation cycle.
Ixx66 acts as a voltage limit for the motor. If the amplifier is oversized for the motor, exceeding the
maximum permitted voltage for the motor, Ixx66 should be set proportionately less than I7m00 to limit
the maximum possible voltage for the motor. Since Ixx66 is a gain, the current loop must be tuned or
retuned after any change.
Servo Loop Output Limit: Ixx69
Ixx69 is the limit of the output of the position/velocity servo loop, which is the torque (quadrature)
current command input to the digital current loop. As such, it acts as an instantaneous current limit for
the motor. Open-loop O-commands are expressed as a percent of Ixx69.
In most other modes, a value of 32,767 (215-1) for Ixx69 for this parameter is full range. In 3-phase
direct-PWM mode, however, the value of 32,767 corresponds to the full-range readings of the Phase A
and Phase B A/D converters. The mathematics involved in the transformation from the phase currents to
the direct and quadrature currents effectively multiplies the phase values by cos(30o), or 0.866. This
means that the Ixx69 value corresponding to the full-range ADC reading is 32,767*0.866 = 28,377.
Therefore, Ixx69 should never be set to a value greater than 28,377 in direct-PWM mode.
98
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
Turbo PMAC User Manual
The amplifier manual should specify the level of current that provides full-range feedback from the
ADCs. Then take the instantaneous current limit for the drive or for the motor, whichever is less, and set
Ixx69 according to the following relationship:

InstCurrentLimit

FullRangeCurrent
Ixx69 = min 28 ,377 ,

* 28 ,377 

If the drive outputs analog current readings and the ADCs are on the interface board, the full-range
current value must be calculated from the volts-per-amp gain of the current sensing in the drive and the
full-range voltage into the interface board.
If a non-zero value of Ixx77 magnetization current will be used, for induction motor control or for field
weakening of a permanent-magnet brushless motor, then Ixx69 should be replaced in the above equation
by
Ixx69 2 + Ixx77 2 .
In early testing, it may be desirable to set Ixx69 to an artificially low value to prevent accidental
overcurrent commands into the motor.
Continuous Current Limit: Ixx57
Ixx57 specifies the magnitude of the continuous current limit for the motor/drive system for integratedcurrent algorithms for thermal protection. If Ixx57 is greater than 0, Turbo PMAC uses I2T protection,
squaring the value of the current before integrating it. This is appropriate when the most temperaturesensitive component is resistive in nature (such as the motor windings). If Ixx57 is less than 0, Turbo
PMAC uses |I|T protection, just taking the absolute value of the current before integrating it. This is
appropriate when the most temperature-sensitive component has an essentially fixed voltage across it, as
with a fully-on bipolar transistor.
Usually, it is the continuous current rating of the drive that is used for a motor’s Ixx57. Even if its
continuous current rating is somewhat higher than that of the motor, usually its thermal time constant is
so much shorter that the integrated current parameters are set to protect it first. Ixx57 is calculated in a
manner similar to Ixx69:
Ixx57 =
ContCurrentLimit
FullRangeCurrent
* 28 ,377
Integrated Current Limit: Ixx58
Ixx58 sets the permitted limit of the time-integrated current over the continuous current value. If the
time-integrated current exceeds this threshold, Turbo PMAC will kill this axis as it would for an amplifier
fault. Typically, this parameter is set by noting the drive specification for time permitted at the
instantaneous current limit. If I2T protection is used, this specification is used in the following equation:
Ixx58 =
Ixx69 2 + Ixx77 2 − Ixx57 2
* ServoUpdateRate( Hz ) * PermittedTime(sec)
32 ,768 2
If |I|T protection is used, this specification is used in the following equation:
Ixx58 =
Ixx69 2 + Ixx77 2 − Ixx57 2
32 ,768
* ServoUpdateRate( Hz ) * PermittedTime(sec)
Refer to the Making an Application Safe section for a more detailed explanation of I2T protection.
Commutation Phase Angle: Ixx72
Ixx72 controls the angular relationship between the phases of a multiphase motor. When Turbo PMAC is
closing the current loop digitally for Motor xx, the proper setting of this variable is dependent on the
polarity of the current measurements.
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
99
Turbo PMAC User Manual
If the phase current sensors and ADCs in the amplifier are set up so that a positive PWM voltage
command for a phase yields a negative current measurement value, Ixx72 must be set to a value less than
1024: 683 for a 3-phase motor, or 512 for a DC brush motor. If these are set up so that a positive PWM
voltage command yields a positive current measurement value, Ixx72 must be set to a value greater than
1024: 1365 for a 3-phase motor, or 1536 for a DC brush motor. The testing described below will show
how to determine the proper polarity.
The direct-PWM algorithms in the Turbo PMAC are optimized for 3-phase motors and will cause
significant torque ripple when used with 2- or 4-phase motors. Delta Tau has created user-written phase
algorithms for these motors; contact the factory if interested in obtaining these.
CAUTION
It is very important to set the value of Ixx72 properly for your system; otherwise
the current loop will have unstable positive feedback and want to saturate. This
could cause damage to the motor, the drive, or both, if overcurrent shutdown
features do not work properly. If unsure of the current measurement polarity in
your drive, consult the Testing PWM and Current Feedback Operation section.
For commutation with digital current loops, the proper setting of Ixx72 is unrelated to the polarity of the
encoder counter. This is different from commutation with an analog current loops (sine-wave control), in
which the polarity of Ixx72 (less than or greater than 1024) must match the encoder counter polarity.
With the digital current loop, the polarity of the encoder counter must be set for proper servo operation.
With the analog current loop, once the Ixx72 polarity match has been made for commutation, the servo
loop polarity match is guaranteed.
Special Instructions for Direct-PWM Control of Brush Motors
A few special settings must be made to use the direct-PWM algorithms for DC brush motors. The basic
idea is to trick the commutation algorithm into thinking that the commutation angle is always stuck at 0
degrees, so current into the A phase is always quadrature (torque-producing) current. This section
summarizes what must be done in terms of variable setup; some of these settings have been discussed
elsewhere as well.
These instructions assume:
• The brush motor’s rotor field comes from permanent magnets or a wound field excited by a separate
means; the field is not controlled by one of the phases of this channel.
• The two leads of the brush motor’s armature are connected to amplifier phases (half-bridges) that are
driven by the A and C-phase PWM commands from Turbo PMAC. The amplifier may have an
unused B-phase half-bridge, but this does not need to be present.
Settings that are the same as for permanent-magnet brushless servo motors with an absolute phase reference:
• Ixx01 = 1 (commutation directly on Turbo PMAC) or Ixx01=3 (commutation over the MACRO ring).
• Ixx02 should contain the address of the PWM A register for the output channel used or the MACRO
Node register 0 (these are the defaults), just as for brushless motors.
• Ixx29 and Ixx79 phase offset parameters should be set to minimize measurement offsets from the A
and B-phase current feedback circuits, respectively.
• Ixx61, Ixx62, and Ixx76 current loop gains are set just as for brushless motors.
• Ixx73 = 0, Ixx74 = 0: These default settings ensure that Turbo PMAC will not try to do a phasing
search move for the motor. A failed search could keep Turbo PMAC from enabling this motor.
• Ixx77 = 0 to command zero direct (field) current.
• Ixx78 = 0 for zero slip in the commutation calculations.
100
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
Turbo PMAC User Manual
•
Ixx82 should contain the address of ADC B register for the feedback channel used (just as for
brushless motors) when the ADC A register is used for the rotor (armature) current feedback. The B
register itself should always contain a zero or near-zero value.
• Ixx81 > 0: Any non-zero setting here makes Turbo PMAC do a “phasing read” instead of a search
move for the motor. This is a dummy read, because whatever is read is forced to zero degrees by the
settings of Ixx70 and Ixx71, but Turbo PMAC demands that some sort of phase reference be done.
(Ixx81=1 is fine.)
• Ixx84 is set just as for brushless motors, specifying which bits the current ADC feedback uses.
Usually, this is $FFF000 to specify the high 12 bits.
Special settings for brush motor direct PWM control:
• Ixx70 = 0: This causes all values for the commutation cycle to be multiplied by 0 to defeat the
rotation of the commutation vector.
• Ixx72 = 512 (90oe) if voltage and current numerical polarities are opposite, or1536 (270oe) if they are
the same. If the amplifier would use 683 (120oe) for a 3-phase motor, use 512 here; if it would use
1365 (240oe) for a 3-phase motor, use 1536 here.
• Ixx96 = 1: This causes Turbo PMAC to clear the integrator periodically for the (non-existent) direct
current loop, which could slowly charge up due to noise or numerical errors and eventually interfere
with the real quadrature current loop.
Settings that do not matter:
• Ixx71 (commutation cycle size) does not matter because Ixx70 setting of 0 defeats the commutation
cycle
• Ixx75 (Offset in the power-on phase reference) does not matter because commutation cycle has been
defeated. Leaving this at the default of 0 is fine.
• Ixx83 (ongoing commutation position feedback address) doesn’t matter, since the commutation has
been defeated. Leaving this at the default value is fine.
• Ixx91 (power-on phase position format) does not matter, because whatever is read for the power-on
phase position is reduced to zero.
Testing PWM and Current Feedback Operation
CAUTION:
On many motor and drive systems, potentially deadly voltage and current levels
are present. Do not attempt to work directly with these high voltage and current
levels unless you are fully trained on all necessary safety procedures. Low-level
signals on Turbo PMAC and interface boards can be accessed much more safely.
Introduction
Most of the time in setting up a direct PWM interface, do not execute all of the steps listed in these
sections (or the Turbo Setup program will do them automatically), but the first time this type of interface
is set up, or if there are problems, these steps will be of great assistance.
All of these tests should be done with the motor disconnected from any loads for safety reasons. All
settings made as a result of these tests are independent of load properties, so will still be valid when the
load is connected.
Before testing any of Turbo PMAC’s software features for digital current loop and direct PWM interface,
it is important to know whether the hardware interface is working properly. We will use PMAC’s Mvariables to access the input and output registers directly. The examples shown here use the suggested Mvariable definitions for Motor 1; for other motors there are equivalent suggested definitions shown in the
Examples section of the manual.
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
101
Turbo PMAC User Manual
Purpose
The purpose of this set of tests is to confirm the basic operation of the hardware circuits on PMAC, in the
drive, and in the motor and to check the proper interrelationships. Specifically:
• Confirm operation of encoder inputs and decode
• Confirm operation of PWM outputs
• Confirm operation of ADC inputs
• Confirm correlation between PWM outputs and ADC inputs
• Determine proper current loop polarity
• Confirm commutation cycle size
• Determine proper commutation polarity
Preparation
First define the M-variables for the encoder counter, the 3 PWM output registers, the amplifier-enable
output bit, and the two ADC input registers. Using the suggested definitions for Motor 1, utilizing Servo
IC 0, Channel 1, we have:
M101->X:$078001,0,24,S
M102->Y:$078002,8,16,S
M104->Y:$078003,8,16,S
M107->Y:$078004,8,16,S
M105->Y:$078005,8,16,S
M106->Y:$078006,8,16,S
M114->X:$078005,14
;
;
;
;
;
;
;
Channel
Channel
Channel
Channel
Channel
Channel
Channel
1
1
1
1
1
1
1
Encoder position register
PWM Phase A command value
PWM Phase B command value
PWM Phase C command value
Phase A ADC input value
Phase B ADC input value
Amp Enable command bit
Note that the ADC values are declared as 16-bit variables even though typically 12-bit ADCs are used;
this puts the scaling of the variable in the same units as Ixx69, Ixx57, Ixx29, and Ixx79.
It will be useful to monitor these values in the Watch window of the Executive program, so add the
variable names to the Watch window, causing the program to repeatedly query Turbo PMAC for the
values which it will display. The hardware can then be exercised with on-line commands issued through
the Terminal window.
To prepare Turbo PMAC for these tests:
• Set I100 to 0 to deactivate the motor.
• Set I101 to 0 to disable commutation (This allows for manual use of these registers.)
• Make sure that I7000, I7004, I7016, and I7017 are set up properly to provide the PWM signals needed.
• If the Amplifier Enable bit is 1, set it to zero with the command M114=0.
• Set Ixx00 and Ixx01 for all other motors to zero.
Position Feedback and Polarity Test
If the PWM command values observed in the Watch window are not zero, set them to zero with the
command
M102=0 M104=0 M107=0
The motor should move freely by hand now. As the motor is turned, monitor the M101 value in the
Watch window. Look for the following:
• It should change as the motor is moved.
• It should count up in one direction and count down in the other direction.
• It should provide the expected number of counts in one revolution or linear distance increment.
• As the motor is returned repeatedly to a reference position, it should report (approximately) the same
position value each time.
If these things do not happen, check the encoder/resolver operation, its connection to Turbo PMAC, and
the Turbo PMAC decode variable I7mn0. Double-check that the sensor is powered. Look at the encoder
waveforms with an oscilloscope.
102
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
Turbo PMAC User Manual
If the positive direction of motion is known, check this here. If the direction is incorrect, invert it by
changing I7mn0, usually from 7 to 3, or from 3 to 7. If it is not known yet which direction sense is
needed, change it later, but make another change at that time to maintain the proper commutation polarity
match, usually by exchanging two of the motor phase leads at the drive.
Note:
Because I100 has been set to 0, and I103 may not yet have been set properly, any
change of position will not be reflected in the motor position window.
PWM Output & ADC Input Connection
Make sure before applying any PWM commands to the drive and motor in this fashion that the resulting
current levels are within the continuous current rating of both drive and motor.
First enable the amp, then apply a very small positive command value to Phase A and a very small
negative command value to Phase B with the on-line commands
M114=1
; Enable amplifier
M102=I7000/50 M104=-I7000/50 M107=0 ; A pos, B neg, C zero
This provides a command at 2% of full voltage into the motor, which should be well within the
continuous current rating of both drive and motor. It is always a good idea to make the sum of these
commands equal to zero so as not to put a net DC voltage on the motor; putting all three commands on
one line causes the changes to happen virtually instantaneously.
With power applied to the drive and the amplifier enabled (M114=1), current readings should be received
in the ADC registers as shown by their M-variables M105 and M106 in the Watch window.
As the M-variables have been defined, +/-32,768 is full current range, which should correspond
approximately to the instantaneous current limit. Make sure that the value read does not exceed the
continuous current limit, usually which is about 1/3 of the instantaneous limit. If the value is well below
the continuous current limit, increase the voltage command to 5% to 10% of maximum. For example:
M102=I7000/10 M104=-I7000/10 M107=0
; 10% of maximum
PWM/ADC Phase Match
Command values from Turbo PMAC’s Phase A PWM outputs should cause a roughly proportionate
response of one sign or the other on Turbo PMAC’s Phase A ADC input (whatever the phase is named in
the motor and drive). The same is true for Phase B.
If there is no response on either phase, re-check the entire setup, including:
• Is the drive properly wired to Turbo PMAC, either directly or through an interface board?
• Is the motor properly connected to the drive?
• Is the drive properly powered, both the power stage, and the input stage?
• Is the interface board properly powered?
• Is the amplifier enabled (M114=1 on Turbo PMAC and indicator ON at the drive)?
• Is the amplifier in fault condition? If so, why?
If an ADC response is received only on one phase, the phase outputs and inputs may not be properly
matched. For example, the Phase B ADC may be reading current from the phase commanded by the
Phase C PWM output. Confirm this by trying other combinations of commands as shown in the six-step
test below, and checking which ADC responds to which phase command. If there is not a proper match,
change the wiring between Turbo PMAC and the drive. Changing the wiring between drive and motor
will not help here.
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
103
Turbo PMAC User Manual
Synchronous Motor Stepper Action
With a synchronous motor, this command should cause the motor to lock into a position, at least weakly,
like a stepper motor. This action may be received temporarily on an induction motor, due to temporary
eddy currents created in the rotor. However, an induction motor will not keep a holding torque
indefinitely at the new location.
Current Loop Polarity Check
Observe the signs of the ADC register values in M105 and M106. These two values should be of
approximately the same magnitude, and must be of the opposite sign from each other. (Again, remember
that these readings may appear noisy. Observe the base value underneath the noise.) If M105 is positive
and M106 is negative, the sign of the PWM commands matches the sign of the ADC feedback values. In
this case, the Turbo PMAC phase angle parameter I172 must be set to a value greater than 1024 (1365 for
a 3-phase motor).
If M105 is negative and M106 is positive, the sign of the PWM commands is opposite that of the ADC
feedback values. In this case, I172 must be set to a value less than 1024 (683 for a 3-phase motor).
CAUTION:
Make sure the I172 value is set properly attempting to close the digital current
loops on Turbo PMAC. Otherwise there will be positive feedback creating
unstable current loops, which could damage the amplifier and/or motor.
If M105 and M106 have the same sign, the polarities of the current sense circuitry for the two phases is not
properly matched. In this case, something has been mis-wired in the drive or between Turbo PMAC and
the drive to give the two phase-current readings opposite polarity. One of the phases will have to be fixed.
Do not attempt to close the digital current loops on Turbo PMAC until the polarities of the current sense
circuitry for the two phases have been matched properly. This will involve a hardware change in the
current sense wiring, the ADC circuitry, or the connection between them. As an extra protection against
error, make sure that Ixx57 and Ixx58 are set properly for I2T protection that will quickly shut down the
axis if there is saturation due to improper feedback polarity.
Troubleshooting
If not getting the current readings expected, probe the motor phase currents on the motor cables with a
snap-on hall-effect current sensor. If current is not seen when voltages are being commanded, check for
phase-to-phase continuity and proper resistance when the motor is disconnected.
Voltage Six-Step Test
For a complete test of the motor/drive connection, try all six sign combinations for a 3-phase motor, or
eight for a 4-phase motor. It is best to command all phase values on a single command line to get
simultaneous changes.
For a synchronous motor, this test will step the motor through one commutation cycle. It can be used to
confirm the size of the commutation cycle in counts, the pole count of the motor, and the commutation
direction sense.
What to Look For
In performing this test, check the following:
• Get a consistent relationship between M102 and M105 on Phase A, and between M104 and M106 on
Phase B.
• For synchronous motors (and probably for induction motors), the M101 position register should
change in approximately equal increments between each step.
• For synchronous motors (and probably for induction motors), the total change in M101 between the
initial Step 1 and the return to Step 1 should be approximately equal to Ixx71/Ixx70.
104
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
Turbo PMAC User Manual
•
For synchronous motors (and probably for induction motors), the physical change in rotor position
between the initial Step 1 and the return to Step 1 (mark the rotor if necessary) should be equal to 1
pole pair:
•
On a 2-pole motor, it should be one full mechanical revolution.
•
On a 4-pole motor, it should be one-half mechanical revolution.
•
On a 6-pole motor, it should be one-third mechanical revolution.
•
On an 8-pole motor, it should be one-fourth mechanical revolution.
•
On a 100-pole motor, it should be 7.2o mech.
Executing the Test
For Motor 1 using Servo IC 0, Channel 1 with a 3-phase motor, with 10% voltage commands being
acceptable, the commands could be:
M102=0 M104=I7000/10 M107=-I7000/10
M102=I7000/10 M104=0 M107=-I7000/10
M102=I7000/10 M104=-I7000/10 M107=0
M102=0 M104=-I7000/10 M107=I7000/10
M102=-I7000/10 M104=0 M107=I7000/10
M102=-I7000/10 M104=I7000/10 M107=0
M102=0 M104=I7000/10 M107=-I7000/10
;
;
;
;
;
;
;
Step
Step
Step
Step
Step
Step
Step
1:
2:
3:
4:
5:
6:
1:
A0,
A+,
A+,
A0,
A-,
A-,
A0,
B+,
B0,
B-,
B-,
B0,
B+,
B+,
C-:
C-:
C0:
C+:
C+:
C0:
C-:
0oelec.
60oelec.
120oelec.
+180oelec.
-120oelec.
-60oelec.
0oelec.
This test moves the motor through the commutation cycle in the positive direction if Ixx72 is less than
1024 (i.e. 683); it moves through the cycle in the negative direction if Ixx72 is greater than 1024 (i.e.
1365). On a synchronous motor, we can use the position reading to check the commutation polarity
match; this may be possible on an induction motor as well.
Action To Take
If the rotor position, as reflected by M101, changed in the wrong direction during this test, we could have
a commutation polarity mismatch. With such a mismatch, the motor would lock in (not run away) when
commanded. There are two possible fixes for this mismatch:
1. Reverse the feedback direction sense by changing I7mn0. However, this changes the direction sense
of the axis, which may not be tolerable.
2. Exchange two phase leads between amplifier and motor. Usually, this is done at the screw terminals
on the amplifier. Exchanging any two phases will change the polarity in the same way. However, the
relationship between the sensor zero position and Turbo PMAC’s commutation cycle zero position is
dependent on which two phases are exchanged.
For asynchronous induction motors, if the above test did not cause proper movement try each direction
polarity and see which works. This is described below.
Once the commutation polarity match has been established, the servo polarity match is established
automatically. Check this later by seeing that positive O-commands cause motor position to count in the
positive direction.
Remember that if I7mn0 is changed later to get the physical direction sense needed, exchange motor
phase leads to re-establish the commutation polarity match.
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
105
Turbo PMAC User Manual
Example
The table of results for a sample run of this test is:
Step
1
2
3
4
5
6
1
M102 (A)
+500
+500
0
-500
-500
0
+500
M104 (B)
0
-500
-500
0
+500
+500
0
M107 (C)
-500
0
+500
+500
0
-500
-500
Cycle Position
0oe
+60oe
+120oe
+180oe
-120oe
-60oe
0oe
Physical Position
4:00
3:00
2:00
1:00
12:00
11:00
10:00
M101 (counts)
8820
9184
9501
9845
10218
10532
10869
M105 (A)
<0
<0
≈0
>0
>0
M106 (B)
≈0
>0
>0
≈0
<0
≈0
<0
<0
≈0
From this test, we can conclude:
• PWM operation is fundamentally working (we got 6 approximately equal steps)
• We have a 4-pole motor because we moved 1/2 revolution
• Current ADC inputs are working: we got proportionate responses
• Sign of current is opposite to sign of voltage, so Ixx72 should be 683.
• We move 2049 counts in one cycle, so 4096 counts per revolution should be correct
• Encoder counter increased, so commutation polarity is correct
Cleaning Up
When done with this section of the testing, write zero values into the command registers and disable the
amplifier with the command:
M102=0 M104=0 M107=0 M114=0
With zero commands into all of the phases of the drive, with the drive either enabled or disabled, the
ADC registers should read nearly zero. Turbo PMAC can compensate for non-zero values with the offset
parameters Ixx79 (A-phase offset) and Ixx29 (B-phase offset). These offset parameters should hold
values of the opposite sign of the phases’ ADC values when there are zero PWM commands. Ixx29 and
Ixx79 magnitudes assume 16-bit values; since we are using 16-bit M-variables to look at the ADC
registers, regardless of the true resolution of the ADCs, we can just read the M-variable values at zero
command and use the opposite values for Ixx79 and Ixx29.
Debugging
With zero commands on the three output registers, you can observe any of the six PWM output signals
with an oscilloscope. All six should have 50% duty cycle (minus the deadtime set by I7m04) at the
frequency set by I7m00. All three top-PWM signals should be in phase with each other. All three
bottom-PWM signals should be in phase with each other, and one-half cycle out of phase with the top
signals. Observing a top and bottom pair, observe the deadtime between the top and bottom on times.
With a positive command into the A-phase, a negative command into the B-phase, and a zero command on
the C-phase, these waveforms should change. On the oscilloscope, observe PWMATOP1 on-time increase,
and PWMABOT1 on-time decrease, while maintaining the deadtime. Similarly, PWMBTOP1 on-time
should decrease, and PWMBBOT1 on-time should increase, maintaining the deadtime between them.
If the analog voltages representing the current measurements are available, these can be probed for
diagnostic purposes. Many direct PWM drives provide analog current measurement outputs and the A/D
conversion is done on an interface board. The probing is easy for these drives. If the A/D conversion is
done inside the drive, access to test points to probe these voltage levels. Consult the drive manual for
location and scaling of these signals.
106
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
Turbo PMAC User Manual
Establishing Basic Current Loop Operation
Once the proper operation of the Turbo PMAC PWM output circuits, the Turbo PMAC ADC input circuits,
and all of the drive and motor circuitry between them have been established, close the current loop.
The Turbo Setup program has both auto-tuning and interactive tuning procedures for the digital current
loop. Most people will use one or both of these procedures to tune the current loop. These procedures
can confirm the proper polarity of the current loop readings also. However, issue the commands
manually as explained in this section.
Purpose
The purpose of these next tests is to set the current loop gains for quick but stable current response. This
is done by giving the motor a current command and observing the current response.
The key to testing the current loop is the use of Turbo PMAC’s O commands, which close the current
loops while leaving open the position and velocity loops. The magnitude of the O-command value is that
of the torque-producing quadrature current command, expressed as a percentage of Ixx69; Ixx77 controls
the magnitude of the direct current command.
It is safest not to create any movement while testing the current loop; this can be done by commanding
only direct current, accomplished by setting Ixx77, then issuing an O0 command.
Digital Current Loop Gains
Ixx61, Ixx62, and Ixx76 are the gains of the PI (proportional-integral) current loop. They are used for
both phases of a multi-phase motor. Ixx61 is the integral gain term. There are two proportional gain
terms: Ixx62 is the forward-path proportional gain, and Ixx76 is the back-path proportional gain. Ixx62 is
multiplied by the current error (commanded minus actual) and the result is added into the output
command. Ixx76 is multiplied by the actual current value and subtracted from the output command. All
three gain terms have a range of 0.0 to 2.0, with 23-bit resolution for positive values.
PMAC2 Digital Current Loop
Current
Command
Forward Path
Proportional
Gain
+
(Id or Iq)
Prescaled to
range
-1.0 < I < 1.0
Kpf
PFM
Scale
Factor
U
+
Ix66
-1.0 < U < 1.0
+
-
Ki
1 - Z-1
Integral
Gain
Voltage
Command
(Vd or Vq)
+
Kpb
Back Path
Proportional
Gain
+
Kpb = Ix76 x 4 (0.0 < Ix76 < 1.0)
Kpf = Ix62 x 4 (0.0 < Ix62 < 1.0)
Ki = Ix61 x 4 (0.0 < Ix61 < 1.0)
Measured
Current
(Id or Iq)
Usually only one of Ixx62 or Ixx76 is used on a given motor; the other gain is set to 0.0. It is more
common to use Ixx61, the forward-path gain, because it provides greater responsiveness and bandwidth.
If Ixx76 is used instead of Ixx62, only the Ixx61 integral gain term directly connects the current command
to the output, and its integration effect filters the command, reducing responsiveness.
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
107
Turbo PMAC User Manual
However, if the command from the position/velocity servo loop is noisy, as can be the case with a lowresolution position sensor, this filtering effect can be desirable, and Ixx76 can provide better performance
than Ixx62.
Analytic Calculation of Current-Loop Gains
With some basic knowledge of motor and amplifier parameters, it is possible to calculate the current-loop
gains directly. It is strongly advised that these computed gains be checked against the values determined
through the auto-tuning or interactive tuning of the Turbo Setup program.
The motor parameters needed are:
• Rpn Motor phase-to-neutral resistance (Ohms)
= Rpp / √3 (Motor phase-to-phase resistance / √3)
• Lpn Motor phase-to-neutral inductance (Henries)
= Lpp / √3 (Motor phase-to-phase inductance / √3)
The amplifier parameters needed are:
• Isat Maximum (saturated) current reading from phase-current A/D converter (Amps).
This is a DC value, not an RMS AC value. This value can be derived from the current-sensor gain Kc
(volts/amp) and the maximum voltage in volts that the A/D-converter can read Vcmax: Isat= Vcmax/Kc.
• VDC DC bus voltage for the amplifier.
This can be derived from the AC RMS supply voltage VAC: VDC=VAC*√2.
Finally, the Turbo PMAC parameter needed is:
• TP Phase-update period (sec)
This can be derived from the phase update frequency fP in kHz: TP=1/(1000*fP)
Next, the following performance specifications for the current loop are required:
• ωn Desired natural frequency of the closed current loop in radians/sec
This can be derived from the desired natural frequency fn in Hz: ωn (rad/s) = 2πfn (Hz).
If the damping ratio (see below) is in the range 0.7 to 1.0, which it should be in most cases, the
desired bandwidth of the current loop basically is equal to the natural frequency. Usually values of
200 Hz to 400 Hz are used.
• ζ Desired damping ratio (dimensionless). A value of 0.7 here yields a step-response overshoot of
about 5%; a value of 1.0 here yields no overshoot.
Now we can compute the proportional current-loop gain Kcp and the integral current-loop gain Kci
according to the formulas:
K cp = I sat
(2ζω n L pn ) − R pn
V DC
K ci = I sat
TP ω n2 L pn
V DC
Finally, to compute the I-variables to represent these gains, we use the formulas:
Ixx62 + Ixx76 =
Ixx61 =
K cp ∗ PWM max cnt
4 ∗ Ixx66
K ci ∗ PWM max cnt
8 * Ixx66
Here, PWMmaxcnt is I7m00 for a channel directly driven by the Turbo PMAC; it is MI900, MI906, or
MI992 for a channel on a MACRO Station.
108
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
Turbo PMAC User Manual
The proportional gain term is expressed as the sum of two I-variables. Ixx62 is the “forward-path”
proportional gain term, directly responding to changes in the command values; Ixx76 is the “back-path”
proportional gain term, directly responding only to the actual current values. When high position
feedback resolution is used in the position/velocity loop, the quantization noise in the current command is
low, and it is better to use Ixx62. When low position-feedback resolution is used, it is better to use Ixx76.
Tradeoffs between responsiveness and smoothness can be obtained by varying the amount of the
proportional gain term allocated to each of these two variables.
Example:
The motor has a phase-to-phase resistance of 3.0 ohms, and a phase-to-phase-inductance of 39
millihenries. The amplifier phase-current sensors provide their maximum 5-volt output for 17.5 amps of
current, and the ADCs provide their full-range value for an input of 5V. The amplifier operates from an
AC supply voltage of 120Vrms. The Turbo PMAC is operating at the default phase update frequency of
9.03 kHz. A current-loop natural frequency of 200 Hz with a damping ratio of 0.7 is desired. The PWMmax-count variable is at the default value of 6528, and Ixx66 for the motor is at the recommended value
of 7181 (10% greater).
Lpn = Lpp / √3 = 0.039 / 1.732 = 0.0225 H
Rpn = Rpp / √3 = 3.0 / 1.732 = 1.732 ohms
Isat = Vcmax/Kc = 5.0 / (5.0/17.5) = 17.5 amps
VDC = VAC * √2 = 120 * 1.414 = 170 V
TP = 1 / (1000*fP) = 1 / (1000*9.03) = 0.000110 sec
ωn (rad/s) = 2 π ωn (Hz) = 2 * π * 200 = 1256 rad/sec
Kcp = Isat [(2 ζ ωn Lpn) – Rpn] / VDC = 17.5 [(2 * 0.7 * 1256 * 0.0225)- 1.732] / 170 = 3.89
Kci = Isat Tp ωn2 Lpn / VDC = 17.5 * 0.000110 * 12562 *0.0225 / 170 = 0.401
Ixx62 + Ixx76 = (Kcp * PWMmaxcnt) / (4 * Ixx66) = 3.89 / (4 * 1.1) = 0.884
Ixx61 = (Kci * PWMmaxcnt) / (8 * Ixx66) = 0.401 / (8 * 1.1) = 0.0456
Preparation for Experimental Tuning
Typically the experimental current-loop tuning is done through the Turbo Setup program or the PMAC
Executive program; both create the proper setup for the tests automatically. However, to do the test
manually, follow the instructions in this section.
To prepare Turbo PMAC for this test:
• Set Ixx00 for all other motors to 0 to de-activate them.
• Set Ixx01 for all other motors to 0 to turn off the commutation. This makes sure Turbo PMAC has
enough calculation time to gather data fast enough.
• Set I7m02 to 0 for divide-by-1 to make servo update rate equal to phase update rate. This permits
Turbo PMAC to gather data every phase update.)
To set up the motor under test:
• Set Ixx00 to 1 for the motor under test to activate it.
• Set Ixx01 to 1 for the motor under test to turn on phase/current calculations. These may have been
left at zero from the earlier tests.
• Give the motor a K command to turn off the outputs
•
•
•
•
Set Ixx71 to 1 to cripple the commutation algorithm and prevent movement during the test.
Make sure the other setup I-variables are set as instructed above. I7m02 and Ixx71 only should be
different from what they would be in the final application.
Start with Ixx62=0.1, Ixx61=0.0, and Ixx76=0.0 as current loop gains.
Set Ixx77 to 3000 to provide about 10% direct current command.
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
109
Turbo PMAC User Manual
In the Detailed Plot section of data gathering, specify data gathering at intervals of one servo cycle.
Select for gathering the commanded and actual direct current registers every servo cycle for the motor
under test. The addresses for these registers are found in the following table.
Commanded Direct Current Registers
Motor #
Address
Motor #
Address
Motor #
Address
Motor #
Address
1
2
3
4
5
6
7
8
Y:$00B8
Y:$0138
Y:$01B8
Y:$0238
Y:$02B8
Y:$0338
Y:$03B8
Y:$0438
9
10
11
12
13
14
15
16
Y:$04B8
Y:$0538
Y:$05B8
Y:$0638
Y:$06B8
Y:$0738
Y:$07B8
Y:$0838
17
18
19
20
21
22
23
24
Y:$08B8
Y:$0938
Y:$09B8
Y:$0A38
Y:$0AB8
Y:$0B38
Y:$0BB8
Y:$0C38
25
26
27
28
29
30
31
32
Y:$0CB8
Y:$0D38
Y:$0DB8
Y:$0E38
Y:$0EB8
Y:$0F38
Y:$0FB8
Y:$1038
Actual Direct Current Registers
Motor #
Address
Motor #
Address
Motor #
Address
Motor #
Address
1
2
3
4
5
6
7
8
Y:$00B9
Y:$0139
Y:$01B9
Y:$0239
Y:$02B9
Y:$0339
Y:$03B9
Y:$0439
9
10
11
12
13
14
15
16
Y:$04B9
Y:$0539
Y:$05B9
Y:$0639
Y:$06B9
Y:$0739
Y:$07B9
Y:$0839
17
18
19
20
21
22
23
24
Y:$08B9
Y:$0939
Y:$09B9
Y:$0A39
Y:$0AB9
Y:$0B39
Y:$0BB9
Y:$0C39
25
26
27
28
29
30
31
32
Y:$0CB9
Y:$0D39
Y:$0DB9
Y:$0E39
Y:$0EB9
Y:$0F39
Y:$0FB9
Y:$1039
Executing the Current-Loop Test
Once the data gathering has been set up, the following commands can be given:
DEFINE GATHER
#1
GAT O0
ENDG K
;
;
;
;
Reserve memory for data gathering buffer
Make sure proper motor is addressed
Start data gathering and give current command
Stop gathering and end current command
The data is then uploaded and plotted. The aim is to get as quick a response as possible to the
commanded value, without significant overshoot or any instability. Typical current loop proportional
gains are in the range 0.6 to 0.9. Typical current loop integral gains are around 0.01.
Clean-Up
When finished with this test, restore the following:
• Set I171 back to its proper value
• Set I7002 back to its proper value
There should now be properly operating current loops. Now, get the commutation algorithm working
properly. See the Finishing Setting up Turbo PMAC Commutation section.
Setting Up Turbo PMAC for Sine-Wave Output Control
This section explains how to set up the commutation scheme if Turbo PMAC is performing the
commutation for a motor, but not the digital current loop. In this mode, Turbo PMAC outputs two phase
current commands to the amplifier, usually as analog voltages through digital-to-analog converters
(DACs). In the steady state, these voltages are a sinusoidal function of time, so this mode is often called
sine-wave output.
Hardware Setup
For Turbo PMAC to operate a motor in the sine-wave commutation analog output mode, two analog
outputs are required for the motor. When used through a PMAC2-style Servo IC, this requires DACs on
both the A and B sub-channels of a given channel. These dual DACs can be found on Acc-8E or Acc-8A
(with Option 2) breakout boards, or on UMAC Acc-24E2A axis-interface/breakout boards.
110
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
Turbo PMAC User Manual
When used through a PMAC-style Servo IC, this requires the single DACs on two consecutive channels.
The higher (even) numbered DAC channel is the A DAC; the lower (odd) numbered DAC channel is the
B DAC.
DAC Output Signals
The A and B DAC outputs should be connected to the phase command inputs on the sine-wave input
amplifier. If the inputs on the amplifier are single-ended, use the DAC+ output only, and leave the
complementary DAC- outputs floating; do not ground them. If the inputs on the amplifier are
complementary, use both the DAC+ and DAC- outputs. In either case, tie the AGND reference voltage
on the output connector to the reference voltage for the amplifier input.
Amplifier-Enable and Fault Interface
On PMAC2-style interface and breakout boards, including the Acc-8A, 8E, and 24E2A, the amplifierenable outputs are dry-contact relays. Normally open, normally closed, sinking or sourcing
configurations from 12V to 24V can be chosen. Normally open contacts (closed when enabled) are
recommended for more fail-safe operation.
On PMAC-style boards, the amplifier-enable outputs are optically isolated solid-state drivers with 24V,
100mA capability. On most boards, sinking or sourcing drivers can be chosen by the selection of
socketed driver IC. The factory default is sinking drivers.
The amplifier fault inputs are optically isolated 12V to 24V inputs. On newer designs with surface-mount
ICs, the isolators are AC Optos so sinking or sourcing drivers can be used. On older designs with
through-hole ICs, sinking drivers are required.
Encoder Feedback
Usually, sine-wave control is done with either digital quadrature encoders connected directly into the
controller (the breakout board is just a pass-through for these signals) or analog sine-cosine encoders
processed through an interpolator accessory. When using an interpolated analog encoder for servo
feedback, typically the uninterpolated digital encoder counter is used for the commutation feedback,
which does not require the high resolution of the servo.
Hall-Effect Commutation Flags
PMAC2-style Servo ICs have supplemental flags for each channel labeled T, U, V, and W. The U, V,
and W flags are commonly used for hall-effect commutation signals (or their optical equivalent) that
provide power-up phase position information.
PMAC-style Servo ICs require the use of a second channel’s flags (usually the same channel as the
second DAC) as the supplemental flags for this purpose.
Turbo PMAC Parameter Setup
PMAC2-Style Servo IC Multi-Channel Setup
The PMAC2-style DSPGATE1 Servo ICs have a great deal of flexibility in supporting different hardware
interfaces. This means that there are certain registers that must be set up through I-variables to support
the desired mode of operation. This is not required if PMAC-style Servo ICs are used.
Hardware Clock Frequency Control: I7m03, MI903, MI907
I7m03 (MS{anynode},MI903 or MS{anynode},MI907 on a MACRO Station) determines the
frequency of four hardware clock signals used for the machine interface channels on Servo IC m. These
can probably be left at the default values. The four hardware clock signals are SCLK (encoder sample
clock), PFM_CLK (pulse frequency modulator clock, DAC_CLK (digital-to-analog converter clock), and
ADC_CLK (analog-to-digital converter clock).
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
111
Turbo PMAC User Manual
Only the DAC_CLK signal is directly used with the sine-wave output, to control the frequency of the
serial data stream to the DACs. The default DAC clock frequency of 4.9152 MHz is suitable for the
DACs on all Delta Tau hardware. Refer to the I7m03 description for detailed information on setting these
variables.
The encoder SCLK frequency should be at least 20% greater than the maximum count (edge) rate that is
possible for the encoder on any axis. Higher SCLK frequencies than this minimum may be used, but
these make the digital delay anti-noise filter less effective.
DAC Strobe Control: I7m05, MI905, MI909
Turbo PMAC generates a common DAC strobe word for each set of four machine interface channels. It
does this by shifting out a 24-bit word each phase cycle, one bit per DAC clock cycle, most significant bit
first. I7m05 (MS{anynode},MI905 or MS{anynode},MI909 on a MACRO Station) contains this
word for the channels on Servo IC m. The default value of $7FFFC0 is suitable for use with the 18-bit
DACs used with the PMAC2-style Servo ICs on Delta Tau interface and breakout boards. A value of
$7FFF00 is suitable for 16-bit DACs.
Individual-Channel Hardware Setup
For each machine interface channel n (n = 1 to 4) of Servo IC m used for sine wave analog outputs, a few
I-variables must be set up properly.
Encoder Decode Control: I7mn0, MI910
I7mn0 (MS{node},MI910 on a MACRO Station) must be set up to decode the commutation encoder
properly. Usually, a value of 3 or 7 is used to provide times-4 decode of a quadrature encoder (4 counts
per encoder line). The difference between 3 and 7 is the direction sense of the encoder; set this variable
so the motor counts up in the direction desired.
The polarity sense of the Ixx72 commutation phase angle parameter must match that of I7mn0 for a
particular wiring; if it is wrong, it will lock into a position rather than generate continuous torque. A test
for determining this polarity match is given below. Remember that if I7mn0 is changed on a working
motor, change Ixx72 as well.
Output Mode Control: I7mn%6, MI916
I7mn6 (MS{node},MI916 on a MACRO Station) must be set to 1 or 3 to specify that outputs A and B
for Channel n are in DAC mode, not PWM. A setting of 1 puts output C (not used for servo or
commutation tasks in this mode) in PWM mode; a setting of 3 puts output C in PFM mode.
Output Inversion Control: I7mn7, MI917
I7mn7 (MS{node},MI917 on a MACRO Station) controls whether the serial data streams to the DACs
on Channel n are inverted or not. The default value of 0 (non-inverted) is suitable for use with the
recommended Acc-8E analog interface board. Inverting the bits of the serial data stream has the effect of
negating the DAC voltage. In a commutation algorithm this is equivalent to a 180o phase shift, which
would produce runaway if the system were working properly before the inversion.
Parameters to Set Up Motor Operation
Several I-variables must be set up for each Motor xx to enable and configure the sine-wave output for that
motor. Of course, Ixx00 must be set to 1 for any active motor, regardless of the output mode for that
motor, and Ixx01 bit 0 must be set to 1 to enable commutation, as covered earlier.
Command Output Address: Ixx02
When using PMAC-style Servo ICs for sine-wave control, Ixx02 must specify the address of an evennumbered DAC register. In this mode, this will cause the Turbo PMAC to use both this DAC and the
next lower-numbered DAC (which is one address higher). Two channels must be used to commutate one
motor. For example, if the address of Servo IC 0 DAC 4 is specified, Turbo PMAC will use DACs 3 and
4. The following table shows the possible Ixx02 values for this mode:
112
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
Turbo PMAC User Manual
Sine-Wave Mode Command Output Addresses –
PMAC-style Servo ICs (Y-registers)
IC# - Chan#
0 – 1&2
0 – 3&4
1 – 1&2
1 – 3&4
Ixx02
$078002
$07800A
$078102
$07810A
IC# - Chan#
2 - 1&2
2 - 3&4
3 - 1&2
3 - 3&4
Ixx02
$078202
$07820A
$078302
$07830A
IC# - Chan#
4 - 1&2
4 - 3&4
5 - 1&2
5 - 3&4
Ixx02
$079202
$07920A
$079302
$07930A
IC# - Chan#
6 - 1&2
6 - 3&4
7 - 1&2
7 - 3&4
Ixx02
$07A202
$07A20A
$07A302
$07A30A
IC# - Chan#
8 - 1&2
8 - 3&4
9 - 1&2
9 - 3&4
Ixx02
$07B202
$07B20A
$07B302
$07B309
Servo ICs 0 and 1 are on the Turbo PMAC itself.
Servo ICs 2 – 9 are on Acc-24P/V or Acc-51P boards.
Channels 1 – 4 on odd-numbered Servo ICs are Channels 5 – 8 on the boards.
When using PMAC2-style Servo ICs for sine-wave control, Ixx02 must specify the address of the A DAC
register for a channel. In this mode, this will cause the Turbo PMAC to use both the A and the B DACs
for the channel. The following table shows the possible Ixx02 values for this mode:
Sine-Wave Mode Command Output Addresses – PMAC2-style Servo ICs (Y-registers)
IC# - Chan#
0-1
0-2
0-3
0-4
1-1
1-2
1-3
1-4
Ixx02
$078002
$07800A
$078012
$07801A
$078102
$07810A
$078112
$07811A
IC# - Chan#
2-1
2-2
2-3
2-4
3-1
3-2
3-3
3-4
Ixx02
$078202
$07820A
$078212
$07821A
$078302
$07830A
$078312
$07831A
IC# - Chan#
4-1
4-2
4-3
4-4
5-1
5-2
5-3
5-4
Ixx02
$079202
$07920A
$079212
$07921A
$079302
$07930A
$079312
$07931A
IC# - Chan#
6-1
6-2
6-3
6-4
7-1
7-2
7-3
7-4
Ixx02
$07A202
$07A20A
$07A212
$07A21A
$07A302
$07A30A
$07A312
$07A31A
IC# - Chan#
8-1
8-2
8-3
8-4
9-1
9-2
9-3
9-4
Ixx02
$07B202
$07B20A
$07B212
$07B21A
$07B302
$07B30A
$07B312
$07B31A
Servo ICs 0 and 1 are on the Turbo PMAC2 itself or on Acc-2E 3U-format stack boards.
Servo ICs 2 – 9 are on Acc-24x2 or Acc-51E boards.
Channels 1 – 4 on odd-numbered Servo ICs are Channels 5 – 8 on dual-Servo-IC boards.
When performing sine-wave output control over the MACRO ring, Ixx02 must specify the address of the
MACRO node register for the A DAC causing the A and B DACs to be used. For the MACRO Type 1
protocol used by the Delta Tau MACRO Station, this is the address of Register 0 of the node. The
following table shows the possible Ixx02 values for this mode:
Sine-Wave Mode Command Output Addresses – MACRO ICs (Y-registers)
IC# - Node#
0-0
0-1
0-4
0-5
0-8
0-9
0 - 12
0 - 13
Ixx02
$078420
$078424
$078428
$07842C
$078430
$078434
$078438
$07843C
IC# - Node#
1-0
1-1
1-4
1-5
1-8
1-9
1 - 12
1 - 13
Ixx02
$079420
$079424
$079428
$07942C
$079430
$079434
$079438
$07943C
IC# - Node#
2-0
2-1
2-4
2-5
2-8
2-9
2 - 12
2 - 13
Ixx02
$07A420
$07A424
$07A428
$07A42C
$07A430
$07A434
$07A438
$07A43C
IC# - Node#
3-0
3-1
3-4
3-5
3-8
3-9
3 - 12
3 - 13
Ixx02
$07B420
$07B424
$07B428
$07B42C
$07B430
$07B434
$07B438
$07B43C
If using the older Type 0 MACRO protocol, add 1 to the value shown in the above table (e.g. $078420
becomes $078421).
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
113
Turbo PMAC User Manual
Commutation Phase Angle: Ixx72
Ixx72 sets the angle from Phase A to Phase B as a fraction of the commutation cycle. Turbo PMAC splits
the commutation cycle (360oe) into 2048 parts. For a 3-phase motor, the angle from A to B is either 1/3
of a cycle (Ixx72=683) or 2/3 of a cycle (Ixx72=1365). For a 2-phase or 4-phase motor, the angle from A
to B is either 1/4 of a cycle (Ixx72=512) or 3/4 of a cycle (Ixx72=1536).
The proper choice of Ixx72 is dependent on the commutation feedback encoder’s direction sense as
determined by its wiring and the encoder decode variable I7mn0, and on the wiring of the phases of the
motor. This choice is generally determined experimentally through a test explained below. Changing
Ixx72 between, say, 683 and 1365 is equivalent to exchanging two phase wires of the motor.
Establishing Basic Output Operation
A quick test can establish basic operation of the commutation outputs, the drive and motor, and the
feedback. The test uses the output-offset variables Ixx29 and Ixx79 to force current directly into the
particular phases and drive the motor like a stepper motor. The test can be used to verify that –
• Output voltages are obtained from the Acc-8E board
• Currents flow in the phases of the motor
• The currents cause the motor to lock into a position (it only does this well for a synchronous motor),
and to set the proper polarity of the Ixx72 commutation phase angle parameter (it only does this well
for a synchronous motor; for an asynchronous induction motor, the polarity of Ixx72 may have to be
determined by trial and error).
Executing the Test
This test, which can be done easily from the terminal window of the Executive program by typing in a
few simple commands, is best illustrated by an example, which will use Motor 1. It should first be done
on a bare motor with no load for safety reasons:
M101->X:$078001,24,S
#1O0
I129=2000
M101
382
I179=2000
M101
215
; Encoder 1 phase position register
; Command zero output
; Positive offset of 2000 bits on 1st phase
; Request position (after motor settles)
; PMAC responds with position
; Positive offset of 2000 bits on 2nd phase
; Request position (after motor settles)
; PMAC responds with position
If the servo-loop feedback has been established already with Ixx03, the position query P command or the
position window in the Executive program can be used instead of the Mx01 encoder position register.
Verifying Basic Operation
Setting a non-zero value for Ixx29 should force a voltage on the A-phase DAC, which can be read with a
voltmeter or oscilloscope. It should also force current in the matching phase of the motor, which can be
measured with a current probe. Setting a non-zero value for Ixx79 should do the same for the B-phase DAC.
A synchronous motor should lock into a position and hold it when an Ixx29 offset is given. An induction
motor may lock in briefly for a brief period of time due to short-term eddy currents in the rotor.
When the Ixx79 offset is added, a synchronous motor should lock into a new position a fraction of a cycle
away from the earlier position. An induction motor may do this also, but probably not as strongly.
114
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
Turbo PMAC User Manual
Evaluating the Polarity Match
Determine the proper setting of Ixx72 by looking at the direction of motion between the two steps. If the
position changed in the negative direction, set Ixx72 less than 1024 – to 683 for a 3-phase motor, or 512
for a 2- or 4-phase motor. If the position changed in the positive direction, set Ixx72 greater than 1024 –
to 1365 for a 3-phase motor, or 1536 for a 2- or 4-phase motor.
For the motor in this example, we conclude that we want a value of 512 if it is a 4-phase motor or 683 if it
is a 3-phase motor. If the encoder direction is subsequently changed for system reasons, change I172 to
match.
Finishing Setting up Turbo PMAC Commutation (Direct PWM or Sine
Wave), Synchronous Motors
By this point, proper operation of the digital current loops should be established for direct PWM control,
or basic operation should be established for analog sine-wave control. The commutation I-variables
Ixx70 and Ixx71 (commutation cycle size), Ixx72 (commutation phase angle), and Ixx83 (commutation
feedback address) already should be set properly.
The next steps, explained in this section, are common for both types of control, with a shared
commutation algorithm.
Confirming Commutation Polarity Match
For the Turbo PMAC commutation algorithms to work properly, the polarity of the output phases must
match the feedback polarity. If there is a mismatch, the algorithm will lock up the motor at a point of
zero torque.
Testing Commutation Polarity Match
With a synchronous motor, we try applying both a direct current command and a quadrature current
command. Because we have not established a phase reference yet, we cannot be sure that a quadrature
current command really produces quadrature current. But if the commutation polarity is correct, at least
one of the commands should cause steady movement of the motor.
First, we apply a direct current command with:
Ixx77=3000 O0
; ~10% direct current command
If this does not produce steady movement, we apply a quadrature current command with:
Ixx77=0 O10
; 10% quadrature current command
To finish the test, we issue a K command and make sure Ixx77 has been returned to 0.
If one of these commands produces steady movement, the commutation polarity is correct, and we can
move on to the next stage of establishing a phase reference. However, if neither of these commands
produces steady motion, we probably have commutation polarity mismatch. To correct the mismatch, see
Correcting Polarity Mismatch, below.
Correcting Polarity Mismatch
To correct a commutation polarity mismatch, there are two possible options:
1. Reverse the feedback direction sense by changing I7mn0. However, this changes the direction sense
of the axis, which may not be tolerable.
2. Reverse the output direction sense. For analog sine-wave output, this can be done by changing Ixx72,
for example from 1365 to 683, or by exchanging two phase leads between amplifier and motor. For
direct PWM, this must be done by exchanging phase leads. Usually this is done at the screw
terminals on the amplifier. Exchanging any two phases will change the polarity in the same way.
However, the relationship between the sensor zero position and Turbo PMAC’s commutation cycle
zero position is dependent on which two phases are exchanged.
After changing the polarity match by one of the above methods, repeat the test to make sure that the
problem has been solved.
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
115
Turbo PMAC User Manual
Establishing a Phase Reference
Purpose
When commutating a synchronous multi-phase motor such as a permanent-magnet brushless motor, the
commutation algorithm must know the absolute position of the rotor. With an absolute sensor such as a
resolver, the phase referencing must be done just once, on assembly of the system. With an incremental
sensor such as an incremental optical encoder, the phase referencing must be done every time the system
is powered up. If incremental sensor power on signal is lost, even if controller power is retained, the
phase referencing must be done again before enabling the signal.
Hall-effect commutation sensors, or their equivalent on an optical encoder, are absolute, but of a very low
resolution (+30oe). In a high-performance application, they are suitable to create a rough, temporary
phase reference, permitting movement until a more accurate reference is established.
The index pulse on an incremental encoder is absolute with high accuracy, but in general, there must be
movement before this pulse is reached. This requires at least a rough phasing, either from a lowresolution absolute sensor such as hall effect, or from a power-on phasing search.
WARNING:
It is important for the safety of the machine that a reliable phase referencing
method be used, whether with an absolute or incremental sensor. If the phase
reference is incorrect by more than 1/4 of the phasing cycle, runaway will occur
when the servo loop is closed. Test the phase referencing carefully with a bare
motor before attaching a load, to make sure the method is reliable. Before
attaching a load, make sure that the Turbo PMAC fatal following error limit
parameter Ixx11 and the amplifier overcurrent fault are active and working
properly. Also make sure that required mechanical protections are in place.
Preparation
These tests require that both the commutation and current loop be working properly. Double-check that
the setup variables are correct for these actions, especially ones that may have been changed for earlier
tests. For motor 1, make sure:
• I100=1 to activate the motor
• I101=1 to enable commutation
• I170 and I171 are set to their proper value
For these tests, we will want access to the motor phase position register, where Turbo PMAC keeps track
of where it is in the phase cycle. The phase position register is 48 bits long, using both X and Y memory.
The Y-memory portion of this register has only fractional information, so we will use only the X-memory
portion. Its units are (counts*Ixx70). The registers are:
Phase Position Angle Registers
Motor #
Address
Motor #
Address
Motor #
Address
Motor #
Address
116
1
X:$00B4
9
X:$04B4
17
X:$08B4
25
X:$0CB4
2
X:$0134
10
X:$0534
18
X:$0934
26
X:$0D34
3
X:$01B4
11
X:$05B4
19
X:$09B4
27
X:$0DB4
4
X:$0234
12
X:$0634
20
X:$0A34
28
X:$0E34
5
X:$02B4
13
X:$06B4
21
X:$0AB4
29
X:$0EB4
6
X:$0334
14
X:$0734
22
X:$0B34
30
X:$0F34
7
X:$03B4
15
X:$07B4
23
X:$0BB4
31
X:$0FB4
8
X:$0434
16
X:$0834
24
X:$0C34
32
X:$1034
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
Turbo PMAC User Manual
This register normally varies from -Ixx71/2 to +Ixx71/2, although if monitoring it, sometimes it will jump
by Ixx71 units and be temporarily outside this range. This is normal behavior. Access to this register is
useful in many ways for establishing a phase reference. Define the suggested M-variable for the Motor 1
phase position register:
M171->X:$00B4,0,24,S
; Motor 1 phase position (counts*Ixx70)
Add this M-variable to the Watch window.
Current-Command Six-Step Test
The basic technique we will use here, either for a one-time phase reference with an absolute sensor or
power-up phase reference with an incremental sensor, is all or part of the current command six-step test.
This is similar to the voltage command six-step test described above, except the current loops are active.
We use the ADC input offset registers to bias the phase current feedback, and hence the phase command
outputs, to drive the motor as a stepper motor to a particular location in the commutation cycle, usually
the 0o position. Then we can write a 0 to the phase position register.
Ixx29 is the A-phase offset; Ixx79 is the B-phase offset. The third phase is not directly commanded;
Turbo PMAC will command it automatically as part of the digital current loop to balance the first two
phases. For motor 1, the following sequence of commands for the current six-step test, and the expected
results, could be:
#1O0
I179=3000 I129=0
I179=3000 I129=-3000
I179=0 I129=-3000
I179=-3000 I129=0
I179=-3000 I129=3000
I179=0 I129=3000
I179=3000 I129=0
;
;
;
;
;
;
;
;
Open
Step
Step
Step
Step
Step
Step
Step
loop command of zero magnitude
1: (A) 0oelec.; (B) 180oelec.
2: (A) -60oelec.; (B) 120oelec.
3: (A) -120oelec.; (B) 60oelec.
4: (A) 180oelec.; (B) 0oelec.
5: (A) 120oelec.; (B) -60oelec.
6: (A) 60oelec.; (B) -120oelec.
1: (A) 0oelec.; (B) 180oelec.
Case (A) is the proper result for all direct PWM setups (Ixx82>0), regardless of the setting of Ixx72. It is
the proper result for sine-wave output setups (Ixx82=0) with Ixx72<1024. Case (B) is the proper result
for sine-wave output setups (Ixx82=0) with Ixx72>1024.
These commands will force about 1/10 of maximum current into phases to drive the motor to known
positions in the phase cycle. Remember to clear the offsets when you are finished with this test:
I179=0 I129=0
Direction-Balance Fine-Phasing Test
The stepper motor phasing test will establish a phase reference typically to within 1 or 2 degrees. This is
adequate for many purposes, but for complete optimization of the motor phase reference, it is necessary to
perform another test, described below. This test finds the best phase reference by making sure that key
performance measures are the same in both directions. Usually the improvement seen in performance
from this fine phasing is better smoothness, not increased torque.
The use of current-loop integrator registers as explained below can be used only in direct PWM systems.
The tests can still be run on sine-wave output systems, but the measurement to be compared in both
directions is the motor velocity. This can simply be read in the position window of the PMAC Executive
Program. This measurement, which is also possible on direct PWM systems, is not quite as sensitive to
phase differences as the measurement explained below, but can still result in an improvement.
This test needs to be performed only once for a given motor. Its purpose is to establish a relationship
between the motor phase angle and an absolute sensor on the motor (e.g. resolver or incremental encoder
index pulse). Most motor manufacturers who mount feedback devices in the factory do not specify a
mounting repeatability tolerance (between motor phase angle and sensor angle) tighter than 1 or 2
degrees, so the results of this test do not necessarily carry from one motor to another of a given design.
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
117
Turbo PMAC User Manual
Note:
Generally, this test is not appropriate for linear motors, because of the relatively
uncontrolled movement it produces. It should only be done on unloaded rotary
motors. On linear motors, a fine phasing test can be done by adjusting the phase
position register so that no movement occurs when a large value of Ixx77 (e.g.
16,000) is given with an O0 command. The test should start with small values, and
movement quickly stopped with a K command.
Preparation
In the Detailed Plot menu of the data gathering section of the PMAC Executive program, set up to gather
the Direct Integrator Output and Quadrature Integrator Output registers. The gathering period should be
set to about 10 servo cycles. The addresses of the registers for each of the motors is:
Direct Integrator Output Registers
Motor #
Address
Motor #
Address
Motor #
Address
Motor #
Address
1
Y:$00BC
9
Y:$04BC
17
Y:$08BC
25
Y:$0CBC
2
Y:$013C
10
Y:$053C
18
Y:$093C
26
Y:$0D3C
3
Y:$01BC
11
Y:$05BC
19
Y:$09BC
27
Y:$0DBC
4
Y:$023C
12
Y:$063C
20
Y:$0A3C
28
Y:$0E3C
5
Y:$02BC
13
Y:$06BC
21
Y:$0ABC
29
Y:$0EBC
6
Y:$033C
14
Y:$073C
22
Y:$0B3C
30
Y:$0F3C
7
Y:$03BC
15
Y:$07BC
23
Y:$0BBC
31
Y:$0FBC
8
Y:$043C
16
Y:$083C
24
Y:$0C3C
32
Y:$103C
4
X:$023C
12
X:$063C
20
X:$0A3C
28
X:$0E3C
5
X:$02BC
13
X:$06BC
21
X:$0ABC
29
X:$0EBC
6
X:$033C
14
X:$073C
22
X:$0B3C
30
X:$0F3C
7
X:$03BC
15
X:$07BC
23
X:$0BBC
31
X:$0FBC
8
X:$043C
16
X:$083C
24
X:$0C3C
32
X:$103C
Quadrature Integrator Output Registers
Motor #
Address
Motor #
Address
Motor #
Address
Motor #
Address
1
X:$00BC
9
X:$04BC
17
X:$08BC
25
X:$0CBC
2
X:$013C
10
X:$053C
18
X:$093C
26
X:$0D3C
3
X:$01BC
11
X:$05BC
19
X:$09BC
27
X:$0DBC
Executing the Test
Before performing this test, first use the stepper-motor method of phasing to get close. Then, in the
terminal window of the Data Gathering section of the PMAC Executive, you can use the following set of
commands (wait a couple of seconds between commands):
DEFINE GATHER
GAT O10
O-10
ENDG K
;
;
;
;
Reserve memory for gathered data
Positive command
Negative command
Stop gathering and kill motor
Upload the gathered data and plot direct and quadrature voltage vs. time. The goal is to have the average
direct voltage reading be the same for the moves in both directions. The quadrature voltage will change
in sign at the move reversal. To adjust the system, wait until the motor is stopped after the test (M171 is
constant) and make a small adjustment to the M171 phase position register with a command like:
M171=M171+5
Then repeat the test as needed until the direct voltage readings are as close as possible in both directions.
This test is sensitive to a count of phasing error, so the last change should probably be +1 count.
Using the Test Results for Absolute Sensor
This test is useful only when matching the super-accurate phase position to an absolute position sensor or
the index pulse of an incremental sensor. With an absolute sensor, assign an M-variable to the sensor
register, and add this to the Watch window. For example:
M175->TWR:0,0
118
; Abs. pos. of 1st resolver on 1st Acc-8D Opt 7 R/D
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
Turbo PMAC User Manual
Make sure the motor is completely at rest. Now multiply the sensor position value read by I170, and
subtract this from the phase position read by M171. (If the motor is moved manually so that M171=0, the
product can be negated). Enter this value into I175 using a statement such as:
I175=M171-(M175*I170)
Finally, set up I181 to read the absolute sensor on subsequent Turbo PMAC resets and store these values
with the SAVE command. Perform another phase reference on this motor.
Using the Test Results for Incremental Index Pulse
For the incremental encoder index pulse, we will use the position capture feature to note where the index
is. Set variable I7mn2 (MS{node},MI912 for a MACRO Station) to 1 if there is a high-true index
pulse, or to 9 if there is a low-true index pulse. (To see which it is, define Mxx19 to its suggested
definition and put it in the Watch window. Generally, if it is 0 there is a high-true pulse.) With a
PMAC2-style Servo IC, to make sure the effective index pulse is only 1 count wide, set I7mn4
(MS{node},MI912 for a MACRO Station) to 1, and I7mn5 (MS{node},MI912 for a MACRO
Station) to the appropriate value for the encoder.
Now assign an M-variable to the encoder flag capture register:
M103->X:$078003,0,24,S
; Encoder 1 flag capture register
Add this to the Watch window. With the motor at rest, note the phase position value in M171 and the
encoder position register in M101. Write these values down. Now turn/push the motor manually in the
direction to home to the machine until M103 changes. The new value is the value of the encoder register
captured at the index pulse.
Subtract your starting M101 value from this new M103 value. Multiply the difference by I170 and add
this to the starting M171 value. The result is the value we will write to the phase position register when
we are settled at the index to refine our initial rough phasing. Mathematically speaking:
IndexPhasePos = I 170 * (IndexM 103 − StartM 101) + StartM 171
Alternately, in a technique that is easier mathematically but harder physically, put M119 in the Watch
window (or the index signal on an oscilloscope) and turn the motor shaft until it stops on the index pulse.
Read the M171 phase position register value. This is the value we will write to the phase position register
when we are settled at the index to refine our initial rough phasing.
Using Hall-Effect Sensors for Phase Reference
Hall-effect sensors or their optical equivalents on a commutation encoder, for a 3-phase motor can be
used for rough phasing on power-up without the need for a phasing search move. This initial phasing
provides reasonable torque, but it will need to be corrected for top operation. Usually the correction is
done when the index pulse is reached, in the same technique that is described above for the correction
after a power-on phasing search move.
Hall-effect sensors usually map out 6 zones of 60oelec. each. In terms of Turbo PMAC’s commutation
cycle, the boundaries should be at 180o, -120o, -60o, 0o, 60o, and 120o. Typically a motor manufacturer
will align the sensors to within a few degrees of this, because these are the proper boundary points if all
commutation is done from the commutation sensors. If mounting the hall-effect sensors, take care to
align the boundaries at these points. The simplest way is to force the motor to the zero degree point with
a current offset (as shown above) and adjust the sensor while watching its outputs to get a boundary as
close as possible to this point.
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
119
Turbo PMAC User Manual
Preparation
Define M-variables to the hall-effect or equivalent inputs. Suggested definitions for Channel 1 on a
PMAC2-style Servo IC are:
M124->X:$078000,20
M125->X:$078000,21
M126->X:$078000,22
M127->X:$078000,23
M128->X:$078000,20,4
;
;
;
;
;
Channel
Channel
Channel
Channel
Channel
1
1
1
1
1
W flag
V flag
U flag
T flag (not usually hall)
TUVW as a 4-bit value
Make these definitions and add these variables to the Watch window. (Delete other variables that are no
longer monitored.) With the motor killed, move the motor slowly by hand to verify that the inputs
expected to change do change.
Executing the Test
To map the hall-effect sensors, we will use the current-loop six-step test, or a variant of it, to force the
motor to known positions in the commutation cycle, and observe the states of the hall-effect signals. The
current-loop test shown above should force the motor right to the boundaries of the hall-effect zones. If
you use these commands, move the motor by hand a little bit at each point to observe the transition.
Force the motor to the expected mid-point of each hall-effect zone instead (or in addition). To do this, the
command sequence would be:
#1O0
I179=3000 I129=-1500
I179=1500 I129=-3000
I179=-1500 I129=-1500
I179=-3000 I129=1500
I179=-1500 I129=3000
I179=1500 I129=1500
I179=3000 I129=-1500
;
;
;
;
;
;
;
;
Open
Step
Step
Step
Step
Step
Step
Step
loop command of zero magnitude
1: (A)-30oelec.; (B)150oelec.
o
2: (A)-90 elec.; (B)90oelec.
o
3: (A)-150 elec.; (B)30oelec.
o
4: (A)150 elec.; (B)-30oelec.
o
5: (A)90 elec.; (B)-90oelec.
o
6: (A)30 elec.; (B)-150oelec.
o
1: (A)-30 elec.; (B)150oelec.
Case (A) is the proper result for all direct PWM setups, regardless of the setting of Ixx72. It is the proper
result for sine-wave output setups with Ixx72<1024. Case (B) is the proper result for sine-wave output
setups with Ixx72>1024.
Remember to clear the offsets when finished with this test:
I179=0 I129=0
It is advisable to create a table listing the values of M124 through M128 for each position. An example
table would be:
Step
M179
M129
1
2
3
4
5
6
1
+3000
+1500
-1500
-3000
-1500
+1500
+3000
-1500
-3000
-1500
+1500
+3000
+1500
-1500
Cycle
Pos.
-30o
-90o
-150o
+150o
+90o
+30o
-30o
Physical
Position
3:30
2:30
1:30
12:30
11:30
10:30
9:30
M101
(counts)
-9001
-9343
-9673
-10030
-10375
-10709
-11050
M126
(U)
0
1
1
1
0
0
0
M125
(V)
1
1
0
0
0
1
1
M124
(W)
0
0
0
1
1
1
0
M128
(TUVW)
2
6
4
5
1
3
2
Note:
If the T flag input is 1, the values of Mx28 will be 8 greater than what is shown in
the table.
120
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
Turbo PMAC User Manual
Using the Test Results
To execute a power-on phasing using the hall-effect sensors, you can use new modes of the Ixx81 poweron phase position parameter, or write a simple PLC program that executes once on power-up/reset.
Setting bit 23 of Ixx81 to 1 specifies a hall-effect power-on phase reference. In this case, the address
portion of Ixx81 specifies a Turbo PMAC X-address, usually that of the flag register used for the motor,
the same address as in Ixx25.
Turbo PMAC expects to find the hall-effect inputs at bits 20, 21, and 22 of the specified register. In a flag
register, these bits match the CHWn, CHVn, and CHUn inputs, respectively. Hall-effect inputs are
traditionally labeled U, V, and W.
Each hall-effect signal must have a duty cycle of 50% (180oe). PMAC can use hall-effect commutation
sensors separated by 120oe. There is no industry standard with hall-effect sensors as to direction sense or
zero reference, so this must be handled with software settings of Ixx81.
Bit 22 controls the direction sense of the hall-effect sensors as shown in the following diagrams, where a
value of 0 for bit 22 is standard and a value of 1 is reversed:
UVW Value: 1
3
2
6
4
5
30
-30
90
-90
150
-150
-150
150
-90
90
1
U
0
1
V
0
1
W
0
Standard: -30
Reversed: 30
This diagram shows the hall-effect waveforms with zero offset, defined such that the V-signal transition
when the U-signal is low (defined as the zero point in the hall-effect cycle) represents the zero point in
PMAC’s commutation cycle.
If the hall-effect sensors do not have this orientation, bits 16 to 21 of Ixx81 can be used to specify the
offset between PMAC’s zero point and the hall-effect zero point. These bits can take a value of 0 to 63
with units of 1/64 of a commutation cycle (5.625oe).
The offset can be computed using the mapping test shown above. In our example, the hall effect zero
(HEZ) point was found to be between 30oe and 90oe, so we will call 60oe. The offset value can be
computed as
Offset =
o
HEZ %360
* 64
360 o
The offset computed here should be rounded to the nearest integer.
In our example, this comes to:
Offset =
60 o % 360
360 o
o
∗ 64 =
60 o
∗ 64 = 10.667 ≈ 11 = 0 Β hex
360 o
The test showed that the hall-effect sensors were in the standard direction, not reversed, so bit 22 is left at
zero. With bit 23 (a value of 8 in the first hex digit) set to 1 to specify hall effect sensing, the first two
hex digits of Ixx81 become $B5. If Flag register 1 at address $C000 were used for the hall-effect inputs,
Ixx81 would be set to $B5C000.
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
121
Turbo PMAC User Manual
The description of Ixx81 in the Software Reference Manual shows the common values of offsets used, for
all the cases where the zero point in the hall-effect cycle is at a 0o, 60o, 120o, 180o, -120o, or -60o point –
where manufacturers generally align the sensors.
Note:
Ixx81 in Turbo is used for address only (i.e. same as Ixx25). Ixx91 in Turbo is
used for bits 16-21, 22 and 23.
Note that Ixx75 is not used for the phase position offset in this method. It can be used to store the final
correction based off fine phasing.
Overall Procedure Summary
The full phase reference then consists of the following steps:
1. Do a rough phase reference using the hall-effect sensors as specified by Ixx81, either automatically
on power-up/reset if Ixx80=1, or on the $ command if Ixx80=0.
2. Do a homing search move on the motor, using the index pulse as part of the home trigger.
3. Wait for the motor to settle “in-position” (following error less than Ixx27) at the home position using
the motor in-position status bit – suggested M-variable Mxx40 – [WHILE(M140=0)...]
4. Force the motor phase position register to the pre-determined value at this point with a command like
Mxx71=Ixx75.
PLC-Based Hall-Effect Reference
Alternately a power-on PLC program could be used to do the hall-effect phasing. This is useful if extra
error trapping is desired, or if sensors of a different format are used.
A program based on the results of our example table would be:
;****************** Set-up and Definitions ********************
CLOSE
; Make sure all buffers are closed
M148->X:$0000C0,8,1
; Motor 2 phasing error fault bit
M171->X:$0000B4,0,24,S
; Motor 2 phase position register
;****************** Program to do phasing search **************
OPEN PLC 1 CLEAR
M148=1
; Tentatively set phasing error bit
IF (M128&7=2)
; Hall Effect State 1 (0 to -60 deg)?
M171=I171/-12
; Set phase angle to -30 deg
P170=1
; Phasing OK flag
ENDIF
IF (M128&7=6)
; Hall Effect State 2 (-60 to -120 deg)?
M171=I171*-3/12
; Set phase angle to -90 deg
P170=1
; Phasing OK flag
ENDIF
IF (M128&7=4)
; Hall Effect State 3 (-120 to -180 deg)?
M171=I171*-5/12
; Set phase angle to -150 deg
P170=1
; Phasing OK flag
ENDIF
IF (M128&7=5)
; Hall Effect State 4 (180 to 120 deg)?
M171=I171*5/12
; Set phase angle to 150 deg
P170=1
; Phasing OK flag
ENDIF
IF (M128&7=1)
; Hall Effect State 5 (120 to 60 deg)?
M171=I171*3/12
; Set phase angle to 90 deg
P170=1
; Phasing OK flag
ENDIF
IF (M128&7=3)
; Hall Effect State 6 (60 to 0 deg)?
122
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
Turbo PMAC User Manual
M171=I171/12
P170=1
ENDIF
IF (M128&7=0 OR M128&7=7)
P170=0
ENDIF
IF (P170=1)
M148=0
CMD"#1J/"
ELSE
CMD"#1K"
ENDIF
DISABLE PLC 1
CLOSE
; Set phase angle to 30 deg
; Phasing OK flag
; Invalid states
; Phasing not OK
; Phasing OK?
; Clear phasing error bit
; Enable motor
; Not OK; disable motor
; So program will not repeat
Notes on this program:
• The reason for the &7 (bit-by-bit AND with 7 [0111]) operation is to remove the effect of the T-flag
input, which is the most significant bit of the 4-bit M-variable.
• This phasing estimate has a potential error of +30oe.
This tentative phasing will serve until we move the motor to the encoder index pulse, at which time we
will force M171 to the value we found with the fine phasing.
Power-On Phasing Search
If a non-absolute sensor is used for commutation, Turbo PMAC must perform a search move for the
proper phasing reference every time it powers up (with an absolute sensor, this needs to be done only
once in the development of the system). There are several ways to do this phasing search. Turbo PMAC
has two automatic methods executed by firmware; other methods or enhancements of these methods can
be executed with PLC programs.
A power-on phasing search permits commutation of permanent-magnet brushless motors without the need
for a more expensive and possibly less accurate absolute sensor. However, a phasing search may not be
dependable in some applications; in these cases an absolute sensor will be required.
The estimate from a power-on phasing search should be within +1-2oe of the true zero position, so many
people will just use the phasing established here throughout the application. It is possible also to adjust
the estimate when settled at the index pulse, using the results of the fine phasing test described above.
WARNING:
An unreliable phasing search method can lead to a runaway condition. Test the
phasing search method carefully to make sure it works properly under all
conceivable conditions. Make sure the Ixx11 fatal following error limit is active
and as tight as possible so the motor will be killed quickly in the event of a serious
phasing search error.
When properly phased, a positive O-command should cause movement in the positive direction; a
negative O-command should cause movement in the negative direction. If the opposite results are
received, there will be a dangerous runaway condition when the servo loop is closed.
Two-Guess Phasing Search
Turbo PMAC’s first automatic phasing search method is called the two-guess phasing search, because it
makes two arbitrary guesses as to the phase position, briefly applies a torque command using each guess,
and observes the response of the motor to each command. Based on the magnitude and direction of the
two responses, Turbo PMAC calculates the proper phasing reference point. It then starts the commutation
based on this reference, and closes the servo loop to hold position.
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
123
Turbo PMAC User Manual
The two-guess phasing search is quick and requires little movement. It works well provided that external
loads such as gravity and friction are low. However, if there are significant external loads, it may not
prove to be a reliable phasing-search method (and unreliable phasing search methods can be dangerous);
if this is the case, another method such as the stepper-motor method described below should be used.
The two-guess method is selected by setting Ixx80 to 0 or 1. With Ixx80 at 0, the phasing search is not
executed automatically during the power-on/ reset cycle; a $ command must be used to execute the
phasing search. With Ixx80 at 1, the phasing search will automatically be executed during the power-on
reset cycle; also, it can be subsequently executed with a $ command.
Two parameters must be specified to tell Turbo PMAC how to do this phasing search. Ixx73 specifies the
magnitude of the torque command during each guess, with units of 16-bit DAC bits. Typical values are
2000 to 6000; 4000 (about 1/8 of full range) is a usual starting point. Ixx74 sets the duration of each
torque command and the evaluation of its response, with units of servo cycles. Typical values are 3 to 10;
5 (about 2 msec at the default servo update) is a usual starting point.
Stepper-Motor Phasing Search
The other automatic method of phasing search for a synchronous motor is the stepper-motor method.
This method forces current through particular phases of the motor, as a stepper-motor controller would,
and waits for it to settle. With proper operation, this will be at a known position in the commutation
cycle. This method is equivalent to two steps of the current-loop six-step test described above.
The stepper-motor phasing search requires more movement and more time than the two-guess method,
but it is more reliable in finding the phase accurately in the presence of large external loads.
The stepper-motor method is selected by setting Ixx80 to 2 or 3. With Ixx80 at 2, the phasing search is
not executed automatically during the power-on/reset cycle; a $ command must be used to execute the
phasing search. With Ixx80 at 3, the phasing search will be executed automatically during the power-on
reset cycle (this is not recommended); also, it can be subsequently executed with a $ command.
In this method, Ixx73 controls the magnitude of the current through the phases, with 32,767 representing
full range. Typically a value near 3000, about 1/10 of full range, will be used, although the actual value
will depend on the loads.
Ixx74 controls the settling time for each of the two steps used in the search. In this mode, the units of
Ixx74 are servo cycles*256, about 1/10 sec with the default servo cycle time. Typically a settling time of
1-2 seconds is used.
In the stepper-motor phasing search, Turbo PMAC first forces current to put the motor at the +/-60o point
in the phasing cycle and then waits for the settling time. Then it forces current to put the motor at the 0o
point in the phasing cycle and again waits for the settling time. It checks to see that there has been at least
1/16 cycle (22.5o) movement between the two steps. If there has been, it forces the phase position register
to 0, clears the phasing-search-error motor status bit, and closes the servo loop. If it has detected less
movement than this, it sets the phasing-search-error bit, and disables (kills) the servo loop.
If the stepper motor phasing search is done outside of the power-on/reset cycle, the phasing search
algorithm will fail on detection of an amplifier fault or overtravel limit condition. Turbo PMAC will set
the phasing-search-error bit and disable the servo loop. If done inside the power-on/reset cycle, Turbo
PMAC cannot detect these errors automatically, but the search will likely fail due to lack of movement.
Custom Phasing Search Methods
It may be necessary or desirable to write a custom phasing-search algorithm. Usually these are executed
as Turbo PMAC PLC programs, but often they can be tried and debugged using on-line commands. The
on-line commands are particularly useful if the phasing search is done only in development to establish a
reference for an absolute sensor.
124
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
Turbo PMAC User Manual
Most custom algorithms are variations on the stepper-motor phasing search method. They use the phasecurrent offset values Ixx29 and Ixx79 with an O0 command to force current into particular phases so the
motor will lock at a certain physical position in its phasing cycle. The following table shows the positions
in the phasing cycle created by different combinations of Ixx29 and Ixx79 for 3-phase motors. Usually
the magnitudes of the non-zero values are 2000 to 3000:
Ixx29
Ixx79
(A). Phase Pos.
(B). Phase Pos.
=0
>0
0oe
+180oe
<0
>0
-60oe
120oe
<0
=0
-120oe
60oe
=0
<0
+180oe
0oe
>0
<0
120oe
-60oe
>0
=0
60oe
-120oe
Case (A) shows the resulting positions for all direct PWM systems and for sine-wave output systems with
Ixx72<1024.
Case (B) shows the resulting phase positions for sine-wave output systems with Ixx72>1024.
For example, the following set of on-line commands typed into the terminal window of the Executive
program could be used to force a motor to the zero position in its phasing cycle, set the phase position
register as zero, and enable the motor.
#1O0
I129=0
I179=3000
M171=0
I179=0
J/
;
;
;
;
;
;
Enable the motor with open-loop zero magnitude
No offset on Phase A
Positive offset on Phase B to force to 0 deg
Write zero into phase position register
No offset on Phase B
Close servo loop
The time between typing the commands would provide sufficient delay for settling into position.
The following PLC program is a good starting point for variants on the stepper-motor phasing search
method. Extensions to this program could be to phase two gantry motors simultaneously or to step out of
a position limit. This example uses Ixx73 and Ixx74 as they would be used in the automatic steppermotor phasing search method.
;****************** Set-up and Definitions ********************
CLOSE
; Make sure all buffers are closed
M248->X:$000140,8,1
; Motor 2 phasing error fault bit
M271->X:$000134,0,24,S
; Motor 2 phase position register
;****************** Program to do phasing search **************
OPEN PLC 1 CLEAR
CMD"#2O0"
; Force zero-magnitude open-loop
M248=1
; Tentatively set phasing error bit
P229=I229
; Save real Phase A bias
P279=I279
; Save real Phase B bias
I229=-I273
; Force negative bias into A
I279=I273
; Force positive bias into A
I5111=I274*256
; Starting value for countdown timer
WHILE (I5111>0)
; Wait for prescribed time
ENDWHILE
P271=M271
; Store phase position at this point
I229=P229
; Restore real bias to A for 0 deg
I5111=I274*256
; Starting value for countdown timer
WHILE (I5111>0)
; Wait for prescribed time
ENDWHILE
P271=P271-M271
; Get difference between two positions
IF (I282>0 OR I272<1024)
; Direct PWM or check analog phase
M271=0
; Set phase position to zero
ELSE
; Analog system with Ixx72>1024
M271=I271/2
; Set phase position to 180 deg
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
125
Turbo PMAC User Manual
I279=P279
IF (ABS(P271)>I271/12)
M248=0
CMD"#2J/"
ELSE
CMD"#2K"
SEND"PHASING FAILED"
DISABLE PLC 1
CLOSE
;
;
;
;
;
;
Restore real bias to B
Greater than 1/12 cycle?
Clear phasing error bit
Close servo loop
Not enough movement
Bad phasing, kill
; Keep from executing again
WARNING:
Make sure an algorithm of this type can be executed reliably. Do not attempt this
algorithm if the position sensor or drive is unpowered or faulted. Turbo PMAC
does not permit the open-loop enabled state required for this PLC if it is into
overtravel limits. The automatic overtravel limit functions may have to be
disabled with Ixx13, Ixx14, and Ixx24. Special logic may be required to step out
of a limit before the full phasing can be done. Remember that improper phase
referencing can lead to runaway conditions. Make sure that both Turbo PMAC
fatal following error limit Ixx11 and amplifier overcurrent fault protections are
active and working.
Final Phase Correction with Index Pulse
To make the final phase correction, move the Turbo PMAC motor to the index pulse and force the value
obtained during the fine phasing test into the phase position register. Before doing this, the
position/velocity servo loop must be reasonably tuned; the tuning for this loop is the same as for any other
PMAC motor.
Usually the move to the index pulse will be the homing search move, where the trigger for the home
position includes the leading edge of the index pulse; typically the first index pulse inside the home flag
pulse. I-variables I7mn2 and I7mn3 control the home trigger. Typically the preliminary phasing will
permit a reasonable move all the way to the home position. It is also possible to perform a preliminary
homing search to the first index pulse, correct the phase, and then do the real homing search.
Once arriving at a known position in the commutation cycle, use the SETPHASE command to force a
predetermined value into the phase position register. When the SETPHASE command is executed, the
value in motor parameter Ixx75 is copied into the phase position register immediately. The SETPHASE
command can be used as an on-line command, in a motion program, or in a PLC program.
A sample motion program segment that performs the homing search and phase correction is:
HOME1
WHILE (M140=0) WAIT
SETPHASE
126
; Command homing search move
; Loop until in position
; Force value into phase position register
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
Turbo PMAC User Manual
Finishing Setting up Turbo PMAC Commutation (Direct PWM or Sine
Wave), Asynchronous (Induction) Motors
Turbo PMAC commutation of an AC induction motor requires the setup of two I-variables that can be left
at 0 for permanent-magnet brushless motors. One variable is the Ixx77 magnetization-current parameter
(which is usually left at 0 for permanent-magnet motors, but can be changed for them). The other
variable is the Ixx78 slip-gain parameter (which must be left at 0 for permanent-magnet motors).
Typically, the Turbo Setup program can be used to set Ixx77 and Ixx78 automatically. The program
stimulates the induction motor to infer its parameters, and sets these terms appropriately for the results it
gets. This section explains analytical and experimental methods for setting these parameters. These settings
are independent of any mechanical load, so any tests can and should be done with an unloaded motor.
Calculating Ixx78 Slip Constant
Calculating from Name Plate Data
The slip constant parameter Ixx78 for an induction motor can be calculated simply from basic parameters
for the motor and for the Turbo PMAC. The following is needed:
• The rated speed for the motor, usually given in revolutions per minute (rpm).
• The electrical line frequency given for this rated speed, usually given in Hertz (Hz), or cycles/sec.
• The number of poles for the motor.
• The Turbo PMAC’s phase update period, usually given in microseconds (µs).
The rated speed can be subtracted from the line frequency (after conversion to consistent units) to get the
slip frequency. This can be multiplied by the phase update period (again after conversion to consistent
units) to get the Ixx78 slip constant. The formula is:
Ixx78 = (ω e − ω m )* T p *
I mag _ std
32 ,768
where:
ωe is the electrical frequency given, in radians/sec. To calculate from frequency in Hertz, multiply by 2π
(6.283).
ωm is the rated mechanical pole frequency, in radians/sec. To calculate from motor rated speed in rpm
and the number of poles, divide the speed in rpm by 60, multiply by 2π (6.283), then multiply by the
number of poles and divide by 2.
Tp is the Turbo PMAC’s phase update time in seconds. To convert from microseconds, divide by one
million. For a Turbo PMAC2, the phase update time can be calculated as:
Tp =
[2 * I (I 19 − 7 ) + 3]* [I (I 19 − 6 ) + 1]
11796480
where I19 is the Turbo PMAC2 parameter containing the number of the clock direction I-variable for the
Servo IC or MACRO IC that is the source of the phase and servo clocks for the system. Usually I19 is set
to 6807 for a Turbo PMAC2 Ultralite to specify MACRO IC 0 (so I6800 and I6801 set the phase update);
usually it is set to 7007 for a board-level Turbo PMAC2 to specify Servo IC 0 (so I7000 and I7001 set the
phase update); usually it is set to 7207 for a UMAC Turbo to specify Servo IC 2 (so I7200 and I7201 set
the phase update).
Imag_std is the value of the magnetization current parameter Ixx77 that would produce the same rated
speed/torque point as the direct operation off the AC lines. For a first calculation, use a value of 3500
here. Usually, it is close enough. If the value of Ixx77 is set as explained in the next section for this type
of operation, come back and adjust this calculation.
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
127
Turbo PMAC User Manual
Example:
A 4-pole induction motor has a rated speed of 1740 rpm at a 60 Hz electrical frequency. It is being
controlled from a UMAC Turbo with default clock source and frequency. The electrical frequency is:
 rad 
 cyc 
 rad 
 = 377.0
 * 2π 

 sec 
 sec 
 cyc 
ω e = 60
The mechanical pole frequency is:
rad
 rad   poles  1  cyc 
 rad 
 rev  1  min 
 * 4
 = 364.4
 = 1740
* 
 * 2π 
* 
sec
 sec 
 min  60  sec 
 cyc   rev  2  pole 
ωm 
In a UMAC Turbo, the default clock source is Servo IC 2 (I19=7207). The default value for I7200 is
6527, and the default value for I7201 is 0. The phase update time can be calculated as:
Tp =
(2 * 6527 + 3)* (0 + 1) = 0.000111 sec
117964800
Ixx78 can now be calculated as:
Ixx78 = (377.0 − 364.4 )* 0.000111 *
3500
32768
= 0.000149
Calculating from Rotor Time Constant
Occasionally, the L/R electrical time constant of the induction motor’s squirrel-cage rotor can be obtained
from the manufacturer (this is distinct from, and much larger than, the L/R electrical time constant of the
stator windings). The Ixx78 slip constant can be calculated easily from this value by the equation:
Ixx78 =
Tp
Tr
where Tp is Turbo PMAC’s phase update time, and Tr is the rotor’s electrical time constant. Remember to
use the same units for both times.
Example
If running with a phase update frequency of 8 kHz and there is a rotor time constant of 0.75 seconds,
calculate:
Ixx78 =
Tp
Tr
=
0.000125
0.75
= 0.000167
Experimentally Optimizing Slip Constant
For a given magnetization current, the optimum slip constant will maximize the acceleration capabilities
of the motor. Changes from the optimum value of Ixx78 in either direction will degrade performance.
Simple tests employing data gathering while using a low-valued O-command (e.g. O10) to accelerate the
motor, permit easy optimization or verification of optimization of the Ixx78 value. If the best value of
Ixx77 magnetization current has not been selected, use a value of 3000 for these tests.
Setting Ixx77 Magnetization Current
Once there is a good value for the Ixx78 slip constant, find the best value of the Ixx77 magnetizationcurrent parameter. Ixx77 sets the commanded value for the direct current component in commutation, the
component in phase with the rotor’s measured/estimated magnetic field orientation. Ixx77 determines the
rotor’s magnetic field strength and so the torque constant Kt and back-EMF constant Ke for the motor. If
Ixx77 is not so high that it magnetically saturates the rotor, torque and back-EMF constants will be
proportional to Ixx77.
128
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
Turbo PMAC User Manual
The higher the value of Ixx77 (before saturation), the more torque is produced per unit of quadrature current
commanded from the servo loop, but the higher the back-EMF generator voltage produced per unit of motor
velocity, so the lower the maximum velocity can be achieved from a given supply voltage. The lower the
value of Ixx77, the less torque is produced per unit of commanded quadrature current, but the lower the
back-EMF voltage produced per unit of velocity, so the higher the velocity that can be achieved.
In most applications, a single value of Ixx77 will be set and left constant for the application. However, it
is possible to change Ixx77 dynamically as a function of speed, lowering it at high speeds so as to keep
the back-EMF under the supply voltage, extending the motor’s speed range. Generally, this technique is
known as field weakening” and can be implemented in a PLC program.
If the Turbo Setup program is not used to set the value of the Ixx77 magnetization-current parameter, it is
best to do so experimentally. With a good value of Ixx78 set, simply issue a low-valued O-command
(e.g. O10) at each of several settings of Ixx77 and observe the end velocity the unloaded motor achieves.
This can be done by watching the real-time velocity read-out in the Executive program’s position
window. If using the data gathering feature, also note the rate of acceleration to that speed.
This velocity is known as the base speed for the motor for that setting. Typically, a value of 3200 to 3500
for Ixx77 will achieve the approximately a base speed equivalent to the rated speed of the motor when run
directly from a 50 Hz or 60 Hz line.
If the test values of Ixx77 are low enough that none of them magnetically saturate the rotor, the base
speeds in the test will be approximately inversely proportional to the value of Ixx77 (and the accelerations
to that speed will be approximately proportional to Ixx77). If you start increasing Ixx77 into the range
that causes magnetic saturation of the rotor, increases in Ixx77 will not cause further lowering of base
speed and further increase in rate of acceleration to that speed.
Many users will want a value of Ixx77 as high as possible without causing rotor saturation. These users
will want to find values of Ixx77 that do cause saturation, then reduce Ixx77 just enough to bring it out of
saturation. The Turbo Setup program finds this setting automatically.
Direct Microstepping of Motors (Open-Loop Commutation)
Turbo PMAC has the ability to do open-loop microstepping (direct microstepping) of standard stepper
motors, working off internally generated pseudo-feedback for both commutation and servo algorithms.
This technique is different from using Turbo PMAC with a pulse-and-direction output to command an
external microstepping drive; that technique does not utilize Turbo PMAC’s commutation algorithms at all.
When microstepping, Turbo PMAC provides two analog outputs that are used as current commands for
phases of the motor. Typically for a microstepping motor, the two phases are electrically independent and
90o out of phase with each other. In this case, the two outputs are simply bi-directional current
commands for the H-bridge amplifiers driving each phase. These amplifiers can be simple torque-mode
(current-mode) DC brush motor amplifiers.
Turbo PMAC’s microstepping algorithm provides 2048 microsteps per electrical cycle, which is 512
microsteps/step. On a typical 200-step/revolution motor, this amounts to 102,400 microsteps per
revolution. With the default phase update frequency of 9 kHz, Turbo PMAC can slew at over 4,600,000
microsteps/second (9000 full steps per second). With a small number of motors and/or fast versions of
the Turbo PMAC, higher phase-update frequencies can be used.
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
129
Turbo PMAC User Manual
Setting the I-Variables
Setting up a motor for microstepping is simply a matter of setting motor I-variables according to the
following list. Since there is no feedback, there is no tuning necessary.
Commutation Enable: Ixx01
Set Ixx01 to 3 to enable Turbo PMAC commutation, reading a Y-register for commutation position
feedback. (See the Ixx83 section below.)
Command Output Address: Ixx02
Set Ixx02 to the lower address of the pair of output DACs being used (e.g. $078002 for DAC1 and DAC2
on a Turbo PMAC or DAC1A and B on a Turbo PMAC2, $078420 for MACRO IC 0 Node 0 Registers 0
and 1 for DACs on a remote MACRO Station. This setting is the same as for using these registers with
normal, closed-loop commutation. The following tables show the settings for DAC pairs in PMAC-style
Servo ICs, in PMAC2-style Servo ICs, and in MACRO ICs.
Sine-Wave Mode Command Output Addresses –
PMAC-style Servo ICs (Y-registers)
IC# - Chan#
0 – 1&2
0 – 3&4
1 – 1&2
1 – 3&4
Ixx02
$078002
$07800A
$078102
$07810A
IC# - Chan#
2 - 1&2
2 - 3&4
3 - 1&2
3 - 3&4
Ixx02
$078202
$07820A
$078302
$07830A
IC# - Chan#
4 - 1&2
4 - 3&4
5 - 1&2
5 - 3&4
Ixx02
$079202
$07920A
$079302
$07930A
IC# - Chan#
6 - 1&2
6 - 3&4
7 - 1&2
7 - 3&4
Ixx02
$07A202
$07A20A
$07A302
$07A30A
IC# - Chan#
8 - 1&2
8 - 3&4
9 - 1&2
9 - 3&42
Ixx02
$07B202
$07B20A
$07B302
$07B309
Servo ICs 0 and 1 are on the Turbo PMAC itself.
Servo ICs 2 – 9 are on Acc-24P/V or Acc-51P boards.
Channels 1 – 4 on odd-numbered Servo ICs are Channels 5 – 8 on the boards.
Sine-Wave Mode Command Output Addresses – PMAC2-style Servo ICs (Y-registers)
IC# - Chan#
0-1
0-2
0-3
0-4
1-1
1-2
1-3
1-4
Ixx02
$078002
$07800A
$078012
$07801A
$078102
$07810A
$078112
$07811A
IC# - Chan#
2-1
2-2
2-3
2-4
3-1
3-2
3-3
3-4
Ixx02
$078202
$07820A
$078212
$07821A
$078302
$07830A
$078312
$07831A
IC# - Chan#
4-1
4-2
4-3
4-4
5-1
5-2
5-3
5-4
Ixx02
$079202
$07920A
$079212
$07921A
$079302
$07930A
$079312
$07931A
IC# - Chan#
6-1
6-2
6-3
6-4
7-1
7-2
7-3
7-4
Ixx02
$07A202
$07A20A
$07A212
$07A21A
$07A302
$07A30A
$07A312
$07A31A
IC# - Chan#
8-1
8-2
8-3
8-4
9-1
9-2
9-3
9-4
Ixx02
$07B202
$07B20A
$07B212
$07B21A
$07B302
$07B30A
$07B312
$07B31A
Servo ICs 0 and 1 are on the Turbo PMAC2 itself or on Acc-2E 3U-format stack boards.
Servo ICs 2 – 9 are on Acc-24x2 or Acc-51E boards.
Channels 1 – 4 on odd-numbered Servo ICs are Channels 5 – 8 on dual-Servo-IC boards.
130
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
Turbo PMAC User Manual
Sine-Wave Mode Command Output Addresses – MACRO ICs (Y-registers), Type 1 Protocol
IC# - Node#
0-0
0-1
0-4
0-5
0-8
0-9
0 - 12
0 - 13
Ixx02
$078420
$078424
$078428
$07842C
$078430
$078434
$078438
$07843C
IC# - Node#
1-0
1-1
1-4
1-5
1-8
1-9
1 - 12
1 - 13
Ixx02
$079420
$079424
$079428
$07942C
$079430
$079434
$079438
$07943C
IC# - Node#
2-0
2-1
2-4
2-5
2-8
2-9
2 - 12
2 - 13
Ixx02
$07A420
$07A424
$07A428
$07A42C
$07A430
$07A434
$07A438
$07A43C
IC# - Node#
3-0
3-1
3-4
3-5
3-8
3-9
3 - 12
3 - 13
Ixx02
$07B420
$07B424
$07B428
$07B42C
$07B430
$07B434
$07B438
$07B43C
Direct Microstepping Enable: Ixx96
Set the command-output mode variable Ixx96 to 1 to tell Turbo PMAC it is using direct microstepping for
this motor.
Encoder Conversion Table Entries: I8xxx
Set up a two-line entry in the encoder conversion table (ECT) for each direct microstepping motor to
process data from the phase position register for the motor. The first setup line (I-variable) has a “6” as
the first hex digit, telling the table to read a 48-bit Y/X register, followed by the phase-position register
address in the last five hex digits. The following table shows the required first setup line for each motor:
Motor
First Setup
Line
Motor
First Setup
Line
Motor
First Setup
Line
Motor
First Setup
Line
1
2
3
4
5
6
7
8
$6000B4
$600134
$6001B4
$600234
$6002B4
$600334
$6003B4
$600434
9
10
11
12
13
14
15
16
$6004B4
$600534
$6005B4
$600634
$6006B4
$600734
$6007B4
$600834
17
18
19
20
21
22
23
24
$6008B4
$600934
$6009B4
$600A34
$600AB4
$600B34
$600BB4
$600C34
25
26
27
28
29
30
31
32
$600CB4
$600D34
$600DB4
$600E34
$600EB4
$600F34
$600FB4
$601034
The second setup line of each entry should be set to $00B018. The first three hex digits specify that 11
bits of the 48-bit register are to be used. The last three hex digits specify that these bits start with Bit 24
of the 48-bit Y/X register, which is Bit 0 of the X register. In other words, the low 11 bits of the X
register are used.
For example, to use the first two lines of the ECT to process Motor 1’s phase position register, I8000
would be set to $6000B4, and I8001 would be set to $00B018.
Position Feedback Addresses: Ixx03 and Ixx04
Set both Ixx03 and Ixx04 to the address of a register in the encoder conversion table that has processed
data from the phase position register. Remember to set them to the address of the second line of the entry.
For example, for Motor 1 to use the result of the two-line first entry in the ECT in the above section’s
example I103 and I104 would be set to $3502, the address of the second line. Starting in firmware
revision 1.937, this can be done by assigning the value to the address of the conversion table I-variable
(e.g. [email protected], [email protected]). Because the phase position register, with 2048 microsteps per
commutation cycle, is used for position feedback, one count is therefore defined as a microstep – 1/2048
of a cycle, or 1/512 of a full step for a 2-phase motor.
Position Scale Factors: Ixx08 & Ixx09
Set both the Ixx08 and Ixx09 scaling factors to 32.
Servo Loop Gains: Ixx30 – Ixx35
•
•
Set the Ixx30 proportional gain term to 8192.
Set the Ixx31 derivative gain term to 0.
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
131
Turbo PMAC User Manual
•
•
•
•
Set the Ixx32 velocity feedforward term to 2048.
Set the Ixx33 integral gain term to 0.
Set the Ixx35 acceleration feedforward term to 2048.
Set this command output limit to 32,767.
Commutation Cycle Size: Ixx70 & Ixx71
Set Ixx70 to 1 and Ixx71 to 2048 to provide 2048 counts (microsteps) per electrical cycle (512
microsteps/step).
Commutation Phase Angle: Ixx72
Set the Ixx72 commutation phase-angle parameter to 512 or 1536 for the usual 2-phase microstepping
motor. Changing between these two values changes the direction sense of positive rotation. If you want
to try microstepping a 3-phase motor, use 683 or 1365. Changing between the two possible settings for a
given number of phases reverses the direction that is “counting up” for the motor.
Current Magnitude: Ixx77
Set the Ixx77 “magnetization current” parameter to control the amount of current used in the phases. This
holds the maximum number of 16-bit DAC bits (or equivalent that will be used to command a DAC
output (current command to the amplifier). For instance, a value of 16,384 provides a +/-5V sinusoidal
output on each phase. This value can be changed at any time to change the amount of current used.
Slip Gain: Ixx78
Set the Ixx78 “slip gain” parameter equal to 2.0/N, where N is the number of phasing cycles per servo
cycle, as set by E3-E6 on a Turbo PMAC, I7002 on a Turbo PMAC2, I7202 on a typical UMAC Turbo,
or I6802 on a Turbo PMAC2 Ultralite. The default setting of these jumpers or variables provides an N of
4, so Ixx78 would be set to 0.5 with the default setting.
Commutation Position Address:Ixx83
Set the Ixx83 commutation position address parameter to the address of the previous phase position
register for the motor. The addresses for each motor are shown in the following table.
Previous Phase Position Registers (Y-addresses)
Motor
Ixx83
Motor
Ixx83
Motor
Ixx83
Motor
Ixx83
1
2
3
4
5
6
7
8
$0000B2
$000132
$0001B2
$000232
$0002B2
$000332
$0003B2
$000432
9
10
11
12
13
14
15
16
$0004B2
$000532
$0005B2
$000632
$0006B2
$000732
$0007B2
$000832
17
18
19
20
21
22
23
24
$0008B2
$000932
$0009B2
$000A32
$000AB2
$000B32
$000BB2
$000C32
25
26
27
28
29
30
31
32
$000CB2
$000D32
$000DB2
$000E32
$000EB2
$000F32
$000FB2
$001032
What to Do Next
Once the appropriate steps in this section have been taken, the motor’s commutation and current loop
should be operating correctly. The motor should be turned in both directions with O-commands; positive
O-commands should cause the motor position to count in the positive direction, and negative Ocommands should cause the motor position to count in the negative direction.
Once this is done, the next step is to set up and tune the position/velocity loop servo, either the standard
PID loop, the Extended Servo Algorithm, or a user-written servo algorithm. This is done in the same
method as for Turbo PMAC motors without digital current loop and/or Turbo PMAC commutation. For
purposes of tuning, a system with PMAC commutation and/or current loop looks like a “torque mode”
drive to the position/velocity loop.
Remember to store the I-variable values that have been set here to the non-volatile flash memory with the
SAVE command.
132
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
Turbo PMAC User Manual
User-Written Phase Algorithms
Turbo PMAC supports the installation and automatic execution of user-written phase algorithms. These
can be used if the standard commutation/current-loop algorithms are not suitable to get the required
performance; alternately, they can be used for non-servo purposes, with the algorithm guaranteed to
execute at the phase update rate. This can be very valuable for fast updates of I/O.
User-written phase algorithms must be written in assembly language for the Motorola DSP56300 family
and assembled using a cross-assembler from Motorola, available at no cost from their website.
Only a single user-written phase algorithm may be installed in a Turbo PMAC. This algorithm can be
executed by any motor on the Turbo PMAC. If different motors should execute different algorithms, this
must be accomplished by branching within a single user-written phase algorithm.
Highly efficient user-written servo algorithms may be written in the assembly language for the DSP56300
family of processors used in the Turbo PMAC. This requires the use of a cross-assembler from Motorola,
obtainable at no cost from their website. It also requires a linking program from Delta Tau, called
CODET.EXE and running under Microsoft Windows operating systems, available at no cost from the
Delta Tau website.
Writing the Algorithm
The algorithm is written in Motorola DSP56300 assembly language using any standard text editor. The
code written is subject to the following restrictions.
Program Memory Space
The program must start at memory location P:$040800, so the first line of code must be:
ORG P:$40800
For a DSP56303 processor (80 MHz CPU Option 5Cx), the resulting code must end by memory location
P:$040BFF, providing a 1-Kword buffer for the program. For a DSP56309 processor (100 MHz CPU
Option 5Dx) or a DSP56311 processor (160 MHz CPU Option 5Ex), the resulting code must end by
memory location P:$044BFF, providing a 17-Kword buffer for the program. In all cases, if a user-written
phase program is used, the user-written servo program is limited to a 2-Kword buffer (P:$040000 –
P:$0407FF).
Conditions on Entry
On entry into the user-written phase, the program can expect the following data for the executing motor in
internal DSP registers:
• The R0 register contains the address of the first status word for the executing motor (e.g. $0000B0 for
Motor 1).
• The N4 register contains the block length of the motor servo registers ($80 presently), which may be
useful in incrementing from motor to motor. This must not be changed.
• The R4 register contains the base address of the first status word for the next higher-numbered motor.
This must not be changed. If you subtract the contents of the N4 register from this value, you will get
the base address of the executing motor’s phasing data.
• The N0 register contains a value of 2.
The torque (quadrature current) command from the motor’s servo algorithm may be found in the motor’s
quadrature command register. This is the register at an address 15 ($F) greater than the address in the R0
register.
Conditions on Exit
On exit from the user-written phase, the algorithm must already have written its outputs into the registers
that perform the action – the function that setup variable Ixx02 performs in the factory provided phase
algorithms. Turbo PMAC firmware will not perform this function after exiting from a user-written phase
routine. A user-written phase algorithm may use the Ixx02 register to tell it in which addresses to place
its output values.
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
133
Turbo PMAC User Manual
The last line of the user-written phase must be RTS (ReTurn from Subroutine).
Available Registers
The following data registers may be used by the user-written phase:
• Internal DSP registers R0, N0, R1, N1, R5, and N5 may be used, and do not need to be restored when
done.
• Internal DSP registers M0, M1, M4, M5, R4, and N4 may be used, but must be restored to previous
values when done.
• Motor commutation registers X:$000xB2/32 through X:$000xBF/3F that are not I-variable registers
may be used to hold values from cycle to cycle. They are not used by any Turbo PMAC firmware as
long as the user-written phase is activated.
• Global registers X/Y:$0010F0 – $0010FF may be used. They are not used by any Turbo PMAC
firmware tasks, other than being set to 0 on power-up/reset.
• Registers in the user buffer established by the DEFINE UBUF command may be used. They are not
used by any Turbo PMAC firmware tasks.
• Other registers may be used as well, but it is possible for certain tasks of Turbo PMAC firmware to
overwrite these. For example, it is possible to use the registers for some P or Q-variables for the userwritten servo, but assigning a value to one of these variables will overwrite the register. It is also
possible to use the I-variables for Turbo PMAC’s standard servo algorithms as gains for the userwritten servo.
• The 2048-entry sine and cosine tables used by Turbo PMAC’s built-in phase routine are located at
addresses $003800 – $003FFF. The sine table is in Y-memory; the cosine table is in X-memory.
Turbo PMAC firmware automatically sets up these tables at power-up/reset; it does not write to these
registers afterwards.
Programming Restrictions
Any levels of the DSP’s stack may not be used, so no DO or JSR instructions are permitted.
Internal DSP address registers R2, R3, R6, and R7 may not be used; modifier registers M2, M3, M6, and
M7; offset registers N2, N3, N6 and N7.
Assembling the Algorithm
The assembly language algorithm must be assembled into DSP56300 machine code using Motorola’s
cross assembler for the computing platform. Follow the instructions from Motorola to do this.
Linking the Algorithm
Use the Delta Tau applet “CODET.EXE”, available on the Delta Tau website to convert the file that
results from the Motorola assembler into a format that can be directly downloaded to the Turbo PMAC.
This file should be archived on your computer or network.
Downloading the Algorithm
Use any version of the PMAC Executive program to download this resulting file into Turbo PMAC’s
program memory. Remember that it is downloaded into volatile RAM memory. To have the Turbo
PMAC retain this algorithm, issue a SAVE command before the controller is reset or power is removed
from it.
Executing the Algorithm
Set bit 1 of Motor xx variable Ixx59 to 1 (Ixx59 = 2 or 3) to select the user-written phase algorithm for
this motor. If bit 0 of Ixx01 is also set to 1, the user-written phase algorithm will execute for this motor
every phase cycle (every [I7+1] phase interrupts, every phase interrupt with the default I7 value of 0),
regardless of whether the motor is open-loop, closed-loop, enabled or disabled. Ixx00 for the motor does
not even have to be set to 1.
134
Setting Up Turbo PMAC-Based Commutation and/or Current Loop
Turbo PMAC User Manual
SETTING UP THE ENCODER CONVERSION TABLE
Turbo PMAC uses a two-step process to work with its feedback and master position information for the
servo algorithm, to provide maximum power and flexibility. (Note that the commutation algorithms
generally use raw feedback registers unprocessed by the conversion table.) For most Turbo PMAC users
with quadrature encoder feedback, this process can be virtually transparent, with no need to worry about
the details. However, some users will need to understand this conversion process in some detail to make
the changes necessary to use other types of feedback, to optimize their system, or to perform special
functions. This section is for those users.
The first stage in the position processing uses the hardware registers such as encoder counters with
associated timers, A/D registers, or accessory cards for parallel input. These work continually without
direct software intervention (although they can be configured through software) with data typically
latched on the servo interrupt. This stage is covered in the Hardware Reference manual for the particular
device and the Setting up Feedback section of this manual. Beyond this point, the process is softwarecontrolled.
Most controllers use the raw data in the hardware registers directly for feedback or master position data.
However, Turbo PMAC has an intermediate step using a software structure called the Encoder
Conversion Table to pre-process the information in the latched registers. This table tells PMAC what
registers to process, and how to process them; it also holds the intermediate processed data.
PMAC Encoder Conversion Table Principle
Servo Address
I-Variables
To Servo
Algorithms
Data
Ix03: Position Loop Feedback Address
Ix04: Velocity Loop Feedback Address
Ix05: Master Position Address
Ix93: Time Base Source Address
Address
Encoder Conversion
Table
PMAC Hardware
Registers
Conversion Instructions: Address
Process & Address
Processed
Feedback Data
RAM
Data
Raw
Feedback
Data
Feedback
Data Signals
Encoder Counters
& Timers, Latches, ADC's
This pre-processing step is necessary, or at least advantageous for several reasons:
• There are several pieces of raw data that need to be combined into one resulting position number, as
with the counter and timers for “1/T” encoder interpolation.
• There is some checking for special conditions (such as rollover or sudden changes indicative of noise,
which needs to be done to ensure robust data.
• Mathematical processing of the raw data, such as addition, subtraction, differentiation, or integration,
needs to be done to get the resulting value.
• The raw data just needs to be transferred and shifted for the servo algorithm to use.
While the primary purpose of the table is to produce processed (converted) values for feedback and
master position data for the servo algorithms to use, other uses are possible. The table’s guaranteed
execution at the beginning of each servo cycle, and its many tools for processing data, permit other uses
as well.
Motor Compensation Tables and Constants
135
Turbo PMAC User Manual
Conversion Table Execution
The conversion table executes automatically at the beginning of each servo cycle, immediately after the
servo interrupt. The entire active part of the table executes before any servo loops execute that cycle.
Each entry in the table is executed every servo interrupt, even if the result is used less often (as when a
motor’s own servo cycle is extended with Ixx60) or not at all.
The table is executed in order from top to bottom each cycle. If the result of one entry is used as the
source for another entry, usually it is desirable to have the second entry further down in the table, so a
servo cycle’s delay is not introduced.
Conversion Table Structure
The Encoder Conversion Table has two columns, one in the X memory space of the processor, and one in
the Y memory space. The X-column holds the converted data, while the Y-column holds the addresses of
the source registers, and the conversion methods used on the data in each of those source registers.
Basically, you set up the table by writing to the Y-column, and PMAC uses the Y-column data to fill up
the X-column each servo cycle.
The table is a series of entries, each of which produces one final result value. Each entry can be 1, 2, or 3
lines (rows) long, with both setup and result data in each line. The final result of the entry is in the last
data line. The entire table is 192 lines long, although typically not all lines are used in a given
application.
Turbo PMAC Encoder Conversion Table
X-Memory
(Results)
Y-Memory
(Set-up)
1. Single Line Entry
Bits
5-23
Result: Integer
Bits
0-4
Fraction
Bits
20-23, 19
Method, mode bit
Bits
0-18
Source Address
2. Multi-Line Entry
Bits
Bits
5-23
0-4
(Intermediate Result)
•••
Result: Integer
Fraction
Bits
Bits
20-23, 19
0-18
Method, mode bit
Source Address
(Conversion Factors)
•••
Setup Structure
Each setup line is represented by an I-variable. The first setup line of the table is represented by I8000,
the second by I8001, and so on; the last setup line of the table is represented by I8191. Because
individual entries are 1, 2, or 3 lines long, each entry uses 1, 2, or 3 setup I-variables, consecutively
numbered.
These I-variables are 24-bit unsigned variables, usually represented as 6-digit hexadecimal values; they
are reported as hexadecimal values if I9 is set to the default of 2, or to 3. Detailed descriptions of the
values these I-variables can take are given in the Software Reference Manual description of I-variables
I8000 – I8191.
136
Motor Compensation Tables and Constants
Turbo PMAC User Manual
Executive Program Setup Menu
Generally, the setup of the encoder conversion table I-variables can be done interactively through the
configuration menu in the PMAC Executive Program. The following section explains how the setup can
be done directly and “manually.”
Entry First Setup I-Variable
The first setup I-variable for an entry has three parts: the conversion method, a mode-switch bit, and the
address of the source data. The conversion method, which determines what processing is to be done, is
contained in the first hex digit (bits 20 – 23). The mode-switch bit, which specifies variants to the
processing, is contained in bit 19. The source address, which specifies which register or set of registers
contains the raw data, is contained in bits 0 – 18.
The following table shows how the first I-variable for an entry is structured:
Entry First Setup Line (I-Variable) Structure
Bit
Digit
Contents
23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7
1
2
3
4
Method
*
Source Address
6
5
5
4
3
2
1
6
0
* Mode-switch bit
For example, the parallel extension (method digit $8) of an incremental encoder from a PMAC-style
(mode-switch bit 0) Servo IC 0 Channel 1 (source address $078000) in the first setup line of the
conversion table would be:
I8000=$878000
The same parallel extension (method digit $8, but from a PMAC2-style (mode-switch bit 1) Servo IC 0
Channel 1 (source address $078000) in the first setup line of the conversion table would be:
I8000=$8F8000
The structures of any subsequent lines for an entry are dependent on the method; each type is discussed
below.
The following table summarizes the operation specified by each method digit.
Method
Digit
# of
lines
Process Defined
$0
1
$1
1
$2
2
$3
3
$4
2
$5
2
$6
2
$7
3
$8
1
$9
2
1/T Extension of
Incremental Encoder
Acc-28 style A/D
converter (high 16 bits,
no rollover)
Parallel Y-word data, no
filtering
Parallel Y-word data,
with filtering
Time Base scaled digital
differentiation
Integrated Acc-28 style
A/D converter
Parallel Y/X-word data,
no filtering
Parallel Y/X-word data,
with filtering
Parallel Extension of
Incremental Encoder
Triggered Time Base,
frozen
Motor Compensation Tables and Constants
1st Additional Line
2nd Additional Line
None
-
-
0 = signed data
1 = unsigned data
-
-
Width/Offset Word
-
Width/Offset Word
Max Change per Cycle
Time Base Scale Factor
-
Input Bias
-
Width/Offset Word
-
Width/Offset Word
Max Change per Cycle
-
-
Time Base Scale Factor
-
Mode Switch
0 = normal shift
1 = unshifted
0 = normal shift
1 = unshifted
None
0 = signed data
1 = unsigned data
0 = normal shift
1 = unshifted
0 = normal shift
1 = unshifted
0 = PMAC IC
1 = PMAC2 IC
0 = PMAC IC
1 = PMAC2 IC
137
Turbo PMAC User Manual
$A
2
$B
2
$C
1
$D
3
$E
1
$F
-
$F/$0
3
$F/$2
2
$F/$3
3
Triggered Time Base,
running
Triggered Time Base,
armed
Incremental Encoder, no
or HW 1/T extension
Exponential filter of
parallel data
Sum or difference of
entries
(Extended entry – type
determined by 1st digit of
2nd line)
High-Resolution
Interpolator
0 = PMAC IC
1 = PMAC2 IC
0 = PMAC IC
1 = PMAC2 IC
0 = No extension
1 = Hardware 1/T
None
Byte-wide parallel Yword data, no filtering
Byte-wide parallel Yword data, with filtering
0 = normal shift
1 = unshifted
0 = normal shift
1 = unshifted
0 = no integration
1 = integrate result
0 = PMAC IC
1 = PMAC2 IC
Time Base Scale Factor
-
Time Base Scale Factor
-
-
-
Max Change per Cycle
-
Filter Gain (Inverse
Time Constant)
-
-
-
$0 Method digit &
Address of 1st A/D
converter
$2 and Width/Offset
Word
$3 and Width/Offset
Word
A/D Bias Term
Max Change per Cycle
Because most source addresses have a 7 in the second hex digit, the effect of setting the mode-switch bit
to 1 is to change the second hex digit from 7 to F. For example, an Acc-28 ADC entry that is $178005
with the mode-switch bit at 0 is changed to $1F8005 with the mode-switch bit at 1.
In the Software Reference Manual under the descriptions for I8000 – I8191, there is a detailed description
for the setup of each type of entry, with extensive tables of the common source addresses for each. The
PMAC Executive program’s setup menu for the conversion table has good pick lists of source addresses
for each method.
Entry Subsequent Setup I-Variables
If the method defined in the first line specifies an entry of more than one line, the next one or two setup
lines (I-variables) in the table are further setup values for this entry. The meaning of these subsequent
lines is dependent on the method; the above table gives a brief description of what these are, and the
Software Reference Manual describes these in detail.
Next Entry
After the last line of any entry, whether the entry is one, two or three lines long, the next line in the table
is the first line of the next entry. For example, if the first line in the table is a three-line entry, I8000,
I8001, and I8002 will be used to set up that entry, and I8003 will be the first setup line of the next entry.
End of Table
If a setup line (I-variable) that could be the first line of an entry is set to 0, this marks the end of the active
table. Even if there are valid entries below this point in the table, they will not be executed.
Result Structure
The column of the encoder conversion table that is in X-memory contains the processed results of the
table. The 24-bit data in this column for the last line of the entry (the only line for a single-line entry) is
the final processed result. This value will be used for subsequent operations in the Turbo PMAC. The
data in this column for previous lines (if any) of the entry contain intermediate results; these generally are
for internal use only.
138
Motor Compensation Tables and Constants
Turbo PMAC User Manual
Final Result Format
In general, the 24-bit final result is structured as 19 bits of integer (bits 5 – 23) and five bits of fraction
(bits 0 – 4). The integer value represents whole number of least significant bits (LSBs) of a counter
(counts), a latched register or an analog-to-digital converter. In other words, the LSB of the main source
register is placed in bit 5 of the result register, a 5-bit shift. The fractional component represents a
measured or estimated value of finer resolution (which could always be zero in some methods); with 5
bits of fractional data, the result will have units of 1/32 of a count. Note that several methods yield a
measurement resolution that is finer than what the software considers a count.
However, this is not always so. It is possible in several conversion methods to specify an unshifted
conversion. When this is done, the LSB of the source data is placed in bit 0 of the result word, not the
usual bit 5. In the high-resolution analog encoder conversion, 10 bits of fraction are produced for each
count of the hardware encoder counter, so the result has units of 1/1024 of a quadrature count, or 1/4096
of an encoder line.
When the final result value of the table entry is used for feedback or master position data by a Turbo
PMAC servo loop, the servo algorithm software will always consider this value to have units of 1/32 of a
count – to have the bit representing one count or LSB located in bit 5 – even if the conversion did not
place it there. This can lead to a difference in scaling between a software count and a hardware count. If
an unshifted conversion is used, the software will consider the LSB of the source register to be 1/32 of a
count. If the high-resolution analog encoder conversion is used, the software will consider one hardware
count of the quadrature counter to be 32 software counts, yielding 128 counts per line of the analog
encoder.
Intermediate Results
If the entry in the conversion table is more than one line in length, the lines in the result column before the
final line hold intermediate results. Typically, these are for internal use only; if there is any utility for the
user, this is documented in the Software Reference Manual for the conversion method.
Using the Results
Most often, the processed data in the conversion table is used for feedback or master data for the servo
algorithms through one of the four following I-variables:
• Ixx03
Motor xx Position-Loop Feedback Address
• Ixx04
Motor xx Velocity-Loop Feedback Address
• Ixx05
Motor xx Master Position Address
• Isx93
Coordinate System x Time Base Source Address
The value of each of these variables is the address of the register that will be read each servo cycle to get
the appropriate information for the servo algorithms. These variables always use the X-register at the
specified address (which is the result column of the table).
There are two ways to specify the value of these variables, and therefore the address that they use. The
first is to specify the address directly, and usually as a hexadecimal number (with $ prefix). The second
way is to specify the address indirectly, by reference to the I-variable at that address, using the ‘@’ symbol
in front of that I-variable name. This permits you to use the result without having to look up the address.
The following table shows the I-variables and the addresses for the first 16 lines of the table, the most
commonly used. A table showing these for all 192 lines of the table is given in the formal description of
these variables in the Software Reference manual.
Motor Compensation Tables and Constants
139
Turbo PMAC User Manual
Line
I-variable
Address
Line
I-variable
Address
1
2
3
4
5
6
7
8
I8000
I8001
I8002
I8003
I8004
I8005
I8006
I8007
$3501
$3502
$3503
$3504
$3505
$3506
$3507
$3508
9
10
11
12
13
14
15
16
I8008
I8009
I8010
I8011
I8012
I8013
I8014
I8015
$3509
$350A
$350B
$350C
$350D
$350E
$350F
$3510
For example, to use the result in the eighth line of the table for position-loop feedback for Motor 4, either
of the following commands could be used:
I403=$3508
[email protected]
Even if the address is specified indirectly, it will still be reported back directly. So in the above example,
regardless of the way the value was specified, if Turbo PMAC is queried for the value of the I-variable, it
will return the address –
I403
$3508
Conversion Methods
Each conversion method is covered in detail in the description of the table’s I-variables in the Software
Reference manual. This section gives a brief overview.
Digital Quadrature Encoders
The most common conversion is for digital quadrature encoders (or other pulse trains, such as pulse and
direction that feed these counters). These are single-line entries. The source addresses are always the
base addresses of encoder channels in the Servo ICs. A full list of these addresses is given in the
Software Reference manual; a quick list of common addresses is given here:
On-board PMAC-style channels 1 – 4: $78000, $78004, $78008, $7800C
On-board PMAC2-style channels 1 – 4: $78000, $78008, $78010, $78018
First four accessory PMAC2-style channels: $78200, $78208, $78210, $78218
There are three conversion methods for this signal format:
• Software 1/T Extension
• Hardware 1/T Extension
• No Extension
Software 1/T Extension (Method digit $0)
This is the most common method, and the one that is used in the table by default. This method uses the
quadrature counter for the integer count value, and computes the ratio of two timers to obtain the five bits
of fractional count value.
140
Motor Compensation Tables and Constants
Turbo PMAC User Manual
PMAC 1/T Extension
Servo
Interrupts
A
B
T
1
T
T
2
1
T
2
Velocity Estimation : V = K
n
T1
Position Estimation : Pn = Counter +/-
T2
T1
Note:
For PMAC2-style Servo ICs of Revision D or newer (introduced 2002), ICchannel variable I7mn9 must be set to the default value of 0 so the two timers can
be read. (Backward compatibility is maintained.)
Examples:
I8000=$078000
I8001=$078004
; Software 1/T extension of Servo IC 0 Channel 1
; Software 1/T extension of Servo IC 0 Channel 2 (PMAC1)
Hardware 1/T Extension (Method digit $C, mode-switch bit 1)
This method lets the Servo IC compute the fractional count value in its hardware. It requires PMAC2style Servo ICs of Revision “D” or newer (introduced 2002), with IC-channel variable I7mn9 set to 1 to
enable the hardware calculations.
This method permits new hardware position-capture and position-compare circuits with sub-count
resolution to be used with digital quadrature encoders. It also saves a small amount of computational
time.
Examples:
I8000=$CF8000
I8001=$CF8008
; Hardware 1/T extension of Servo IC 0 Channel 1
; Hardware 1/T extension of Servo IC 0 Channel 2 (PMAC2)
No Extension (Method digit $C, mode-switch bit 0)
This method computes no fractional count value, leaving the 5 fractional bits at 0, and just using the
hardware counter for the whole count values. This is primarily used for the simulated feedback from the
pulse-and-direction output for open-loop stepper control, in which the fractional count estimation can
cause unwelcome dithering.
Motor Compensation Tables and Constants
141
Turbo PMAC User Manual
Examples:
I8008=$C78200
I8009=$C78208
; No extension of Servo IC 2 Channel 1
; No extension of Servo IC 2 Channel 2 (PMAC2 IC)
Analog Sine/Cosine Encoders
Turbo PMAC supports two methods of conversion for interpolating analog sine/cosine encoders: one lowresolution (128 or 256 states per line) and one high-resolution (4096 states per line). The source
addresses are always the base addresses of encoder channels in the Servo ICs.
Low-Resolution Interpolation (Method digit $8)
This method reads five bits of externally calculated fractional count data presented in parallel to the Turbo
PMAC, and uses the quadrature counter for whole-count data. It is intended for use with the interpolator
circuitry on the Acc-8D Option 8 and the Acc-8A through normal PMAC servo channels. The modeswitch bit is set to 0 if the data is read through a PMAC-style Servo IC; it is set to 1 if the data is read
through a PMAC2-style Servo IC. This is a one-line entry.
A full list of source addresses is given in the Software Reference Manual. Common source addresses are:
PMAC on-board channels 1 – 4: $78000, $78004, $78008, $7800C
PMAC2 on-board channels 1 – 4: $78000, $78008, $78010, $78018
Note that the encoder counter used must be set up for times 4 decode by the decode variable I7mn0 for
the encoder channel, and that only one of the two possible direction senses may be used. With Delta Tau
accessories and standard wiring, the I7mn0 variable must be set to 7. If this is wrong, the direction senses
of the fractional and whole-count data will not match.
Examples:
I8004=$878100
I8005=$8F8018
; Low-res interpolation of Servo IC 1 Channel 1 (PMAC1 IC)
; Low-res interpolation of Servo IC 0 Channel 4 (PMAC2 IC)
High-resolution Interpolation (Method digit $F, extended method digit on second line $0)
This method calculates 10 bits of fractional count data (1024 states) per quadrature count – 4096 states
per line – by reading A/D converters on an Acc-51 board for the sine and cosine signals and computing
the arctangent. The whole count data comes from the digital quadrature counter. The mode-switch bit is
set to 0 if the data is read through a PMAC-style Servo IC (Acc-51P); it is set to 1 if the data is read
through a PMAC2-style Servo IC (Acc-51E, 51C).
A full list of source addresses is given in the Software Reference Manual. Common source addresses are:
Acc-51P as Servo IC 2, channels 1 – 4: $78200, $78204, $78208, $7820C
Acc-51E as Servo IC 4, channels 1 – 4: $79200, $79208, $79210, $79218
The second line contains the address of the first of the two ADCs read, with a $0 in the first hex digit.
For the Acc-51P, the ADC address is two higher than the channel base address used in the first line; for
the Acc-51E and 51C, it is five higher.
Common source addresses for the ADCs are:
Acc-51P as Servo IC 2, channels 1 – 4: $78202, $78206, $7820A, $7820E
Acc-51E as Servo IC 4, channels 1 – 4: $79205, $7920D, $79215, $7921D
The third line is a 24-bit bias term that is subtracted from both 24-bit words containing the ADC values
(even though the ADCs only occupy the top 14 bits) before the arctangent calculations are done.
The encoder counter used must be set up for times 4 decode by setting variable I7mn0 for the encoder
channel to 3 or 7. If this variable is changed, for example to change the direction sense of the counter, the
new value must be saved and the Turbo PMAC reset in order for this conversion to work properly (for the
direction sense of the fractional data to match the directional sense of the whole-count data).
142
Motor Compensation Tables and Constants
Turbo PMAC User Manual
Examples:
I8000=$F78200
I8001=$078202
I8002=$004000
; Hi-res interpolation of Servo IC 2 Channel 1 (PMAC1 IC)
; Read ADCs from Servo IC 2 Channel 1
; Bias term of 4 LSBs of 12-bit ADCs
I8000=$FF8300
I8001=$078305
I8002=$000000
; Hi-res interpolation of Servo IC 3 Channel 1 (PMAC2 IC)
; Read ADCs from Servo IC 3 Channel 1
; Zero bias term
Acc-28 Analog-to-Digital Converters
The conversion table can process the data from the 16-bit A/D converters of Acc-28 boards or their
equivalent. The source address is that of the A/D converter register in the IC that interfaces to the accessory.
Standard A/D conversion (Method digit $1)
This method takes the value in the A/D converter (without integrating it) and shifts it to leave the LSB in
bit 5 of the result – what Turbo PMAC software considers a count. If the A/D converter provides a
signed value (Acc-28A), the mode-switch bit should be set to 0; if the A/D converter provides an
unsigned value (Acc-28B, 28E), the mode-switch bit should be set to 1. This is a single-line entry.
Common source addresses are:
Acc-28A/B into PMAC channel 1 – 4: $78006, $78007, $7800E, $7800F
Acc-28B into PMAC2 channel 1A, 1B, 2A, 2B: $78005, $78006, $7800D, $7800E
Acc-28E as first I/O board, channel 1 – 4: $78C00, $78C01, $78C02, $78C03
Examples:
I8004=$178006
I8005=$1F8105
I8006=$1F8C00
; Conversion of Servo IC 0 ADC 1 (PMAC1 IC), signed
; Conversion of Servo IC 1 ADC 1A (PMAC2 IC), unsigned
; Conversion of first Acc-28E Channel 1, unsigned
Integrated A/D conversion (Method digit $5)
This method takes the value in the A/D converter, shifts it so the LSB is in bit 5 and adds it to the
previous cycle’s result to yield a numeric integration. If the A/D converter provides a signed value (Acc28A), the mode-switch bit should be set to 0; if the A/D converter provides an unsigned value (Acc-28B,
28E), the mode-switch bit should be set to 1. This is a double-line entry. The second line holds a bias
term, expressed as a 24-bit number, that is subtracted from the source A/D data (whose LSB is in bit 8)
before the numerical integration.
Examples:
I8004=$578006
I8005=$000300
; Integration of Servo IC 0 ADC 1 (PMAC1 IC), signed
; Bias term (3 LSBs of 16-bit ADC)
Parallel Feedback
The conversion table can process parallel feedback data read from a register or set of registers. While this
is fundamentally a copying operation, it serves several important functions:
• It copies the data into an X-register, where the servo-loop I-variables can access it (most I/O
registers are in Y-memory).
• It can shift the data so the LSB of the source register ends in bit 5 of the result, which PMAC
firmware treats as a count.
• It can mask out unwanted bits from the source word, and properly handle rollover and software
extension.
• It can perform a filtering function, mitigating the effect of noise in significant data bits.
Parallel feedback data can be read from a continuous range of bits up to 24-bits wide from within a 48-bit
range, or from 2 or 3 byte-wide pieces in consecutive registers, as from many of the UMAC I/O boards.
Five method digits for the first setup line of the entry in the conversion table handle this case:
Motor Compensation Tables and Constants
143
Turbo PMAC User Manual
•
$2: The 48 possible bits are from the Y-register at the specified address and the Y-register at the next
higher address. The conversion table performs no filtering.
• $3: The 48 possible bits are from the Y-register at the specified address and the Y-register at the next
higher address. The conversion table can perform filtering of the source data.
• $6: The 48 possible bits are from the Y-register at the specified address and the X-register at the
same address. The conversion table performs no filtering.
• $7: The 48 possible bits are from the Y-register at the specified address and the X-register at the
same address. The conversion table can perform filtering of the source data.
• $F: The data is in byte-wide pieces at the Y-register specified address and adjacent higher
address(es). In this case, the first hex digit of the entry’s second setup line must be either $2 (no
filtering) or $3 (with filtering).
In all of these cases, the bit-19 mode-switch bit in the first setup line of the entry is set to 0 to place the
LSB of the source data in bit 5 of the result, where Turbo PMAC firmware treats it as a count. If the
mode-switch bit is set to 1, the source LSB is placed in bit 0 of the result, where Turbo PMAC firmware
treats it as 1/32 of a count. This alternate setting can be used for very high-resolution feedback, where
numerical velocity saturation could be a problem.
Bits 0 – 18 of the first setup line of the entry specify the address of the source data. In the case of
methods $2 and $3, or $F/$2 and $F/$3, some of the source data could also come from the next higher
addresses.
The second setup line of the entry specifies which part of the source address data is used. This line is split
into four parts: the first hex digit, the second and third hex digits, the fourth hex digit, and finally the fifth
and sixth hex digits.
Second Setup Line for Parallel Data Entries
Hex Digits
Contents
1
Aux Meth
2&3
Bit Width
4
Byte
5&6
LSB Location
In the case of methods $2, $3, $6, and $7, the first digit is always 0. In the case of method $F, the first
digit is $2 if no filtering is to be done; it is $3 if filtering is to be done.
The second and third digits specify the bit width – how many bits are to be used.
In the case of methods $2, $3, $6, and $7, the fourth digit is always 0. In the case of method $F, the
fourth digit specifies which byte the data comes from: 0 for the low byte, 1 for the middle byte, and 2 for
the high byte.
The last two digits specify which bit of the source byte or word is to be used as the LSB. For byte-wide
data this is always in the range $00 to $07. For data from a 48-bit field, this value could be up to value of
48 minus the bit width.
If the source data is to be filtered, there is a third setup line in the entry, which specifies the maximum
change in the source data (in units of LSBs of the used source data) in a single servo cycle that is to be
regarded as real. If the source changes by more than this (maybe due to noise in a high bit), the result will
change by this amount only. This is an important protection against noise and other possible anomalies.
144
Motor Compensation Tables and Constants
Turbo PMAC User Manual
Examples:
These first three 3-line entries process 3 16-bit inputs on a 48-bit Acc-14D/V card
I8000=$378A00
I8001=$010000
I8002=256
I8003=$378A00
I8004=$010010
I8005=256
I8006=$378A01
I8007=$010008
I8008=256
; Filtered parallel Y-data from first Acc-14D/V
; Use 16 bits starting at bit 0
; Max change of 256 LSBs per servo cycle
; Filtered parallel Y-data from first Acc-14D/V
; LSB from bit 16, use 16 bits (high 8 bits from next address)
; Max change of 256 LSBs per servo cycle
; Filtered parallel Y-data from first Acc-14D/V
; LSB from bit 8, use 16 bits
; Max change of 256 LSBs per servo cycle
These next several examples read MACRO registers unshifted
I8000=$2F8420
I8001=$018000
I8002=$2F8424
I8003=$018000
; Unshifted unfiltered read of IC 0 Node 0 Reg 0 (address $078420)
; Use 24 bits starting at bit 0
; Unshifted unfiltered read of IC 0 Node 1 Reg 0 (address $078424)
; Use 24 bits starting at bit 0
This next example reads data in byte-wide pieces from an Acc-14E
I8000=$F78C00
I8001=$314000
I8002=150
; Read from $78C00
; Filtered parallel Y-data 20 bits, low byte, start at bit 0
; Max change of 150 LSBs per servo cycle
This example reads a timer register as parallel data for MLDT feedback
I8003=$378000
I8004=$013000
I8005=75
; Filtered parallel read from Servo IC 0 Channel 1
; Read 19 bits starting at bit 0
; Max valid change of 75 LSBs per servo cycle
Exponential Filter
The conversion table can perform filtering functions on incoming data by executing the exponential-filter
entry. This entry implements a low-pass (single-pole IIR) filter on the source data. It is mainly useful for
master position data; its use for feedback data can be problematic because of the delays it introduces and
the resulting stability problems in the feedback loop.
This is a three-line entry. The first line uses a $D method digit, followed by the address. The source data
must be in the X-register of this address, and usually is the result of a previous entry in the conversion
table (addresses $3501 - $35C0).
The second line contains the maximum change in the source data that will be permitted in a single servo
cycle.
The third line contains the filter gain. Essentially, this is an inverse time constant; the gain term is set to
223/(Tf+1), where Tf is the filter time constant expressed in servo cycles.
Example:
I8004=$07820C
I8005=$D03504
I8006=4000
I8007=524288
[email protected]
; 1/T interpolation of Servo IC 2 Encoder 4
; Exponential filter of I8004 result
; Max permitted change (LSBs/servo cycle)
; Filter gain for 15-servo-cycle time constant [223/(15+1)]
; Use filtered value as master position for Motor 12
Sum and Difference
A sum and difference entry in the conversion table can be used to do many useful tasks:
• Adding or subtracting two results in the table
• Changing the sign of a result in the table
• Integrating the value of a result in the table
• Integrating the sum or difference of two results in the table
Motor Compensation Tables and Constants
145
Turbo PMAC User Manual
This can be used, for example, to average the readings of two sensors on opposite sides of a rotary table to
take out the eccentricity, or to calculate the skew on a gantry by taking the difference of sensors on the
two sides.
This entry can use as source data only results in the conversion table. It is strongly encouraged to use
results from earlier in the table; using results from later in the table means results from the previous servo
cycle are used.
This is a one-line entry, split into four parts.
The first part is the first hex digit, which is always “E” for this type of entry.
• The second part is the second hex digit (bits 16 – 19), which contains four mode-switch bits:
•
Bit 19, if set to 1, integrates the final result, adding the sum or difference calculated each cycle to
the final result of the previous cycle.
•
Bit 18, if set to 1, instructs Turbo PMAC to ignore the second source, permitting this entry to
simply negate and/or integrate a single value.
•
Bit 17, if set to 1, negates the second entry before adding it into the result, creating an effective
subtraction.
•
Bit 16, if set to 1, negates the first entry before adding it into the result.
• The third part uses the third and fourth hex digits as a number representing the offset from the
beginning of the table of the second source used. This can be calculated by subtracting 8000 from the
I-variable number that corresponds to this result.
• The fourth part uses the fifth and sixth hex digits as a number representing the offset from the
beginning of the table of the first source used. This can be calculated by subtracting 8000 from the Ivariable number that corresponds to this result.
Examples:
Using the following entries in the table as source data:
•
I8000=$078200
I8001=$078208
I8002=$078210
I8003=$078218
I8004=$F78C00
I8005=$314000
I8006=256
; 1/T interpolation of Servo IC 2 Encoder 1
; 1/T interpolation of Servo IC 2 Encoder 2
; 1/T interpolation of Servo IC 2 Encoder 3
; 1/T interpolation of Servo IC 2 Encoder 4
; Byte-wide parallel read of Acc-14E data
; Filtered, 20 bits, low byte, start at bit 0
; Max change of 256 LSBs per servo cycle
The sum and difference entries can be used to do several processes:
I8012=$E00001
I8013=$E10203
I8014=$E40006
I8015=$EC000D
; Sum of results from I8000 and I8001
; Subtract result of I8003 from result of I8002
; Negate result of I8006
; Integrate result of I8013
Time Base (Differentiation)
Time-base entries in the conversion table provide scaled digital differentiation of source data. Usually, the
source data is encoder count value, so the differentiation calculates a frequency value. This frequency value
can then be used as an external time base for a coordinate system in order to slave the motion of axes in the
coordinate system to the master encoder, creating an electronic cam. Refer to the Synchronizing Turbo
PMAC to External Events section for details in the use of this data for time base control.
146
Motor Compensation Tables and Constants
Turbo PMAC User Manual
Untriggered vs. Triggered Time Base
A time-base entry can be untriggered or triggered. Untriggered time base entries are simpler, but do not
have a means for starting the time-base tracking of the master encoder precisely at a specific position of
the master. The triggered time base entries use the hardware-capture feature of the Turbo PMAC Servo
ICs to precisely latch the master position that is the starting point for the programmed motion sequence
that is slaved to the master.
Time-base entries are two-line entries. The first line contains the method digit in bits 20 – 23 ($4 for
untriggered, $9 for triggered frozen state, $B for triggered armed state, and $A for triggered running
state), the bit-19 mode-switch bit for triggered entries (=0 for PMAC-style Servo ICs, =1 for PMAC2style Servo ICs), and the source address of the data to be differentiated in bits 0 – 18.) The second line
contains the time-base scale factor which multiplies the differentiated source data.
Source Register – Untriggered
In untriggered time base, usually the source data is the result of a previous conversion in the table –
source address $3501 to $35C0 – typically the result of a 1/T interpolation of an incremental encoder.
The 1/T interpolation can be important to reduce the quantization noise generated by the digital
differentiation of the encoder data; the result has five bits of fraction, estimating the encoder position to
1/32 of a count.
The untriggered time base takes this source data, subtracts the previous cycle’s source data from it,
multiplies the difference by two times the time-base scale factor, and stores this value as its result. If the
result is to be used as the time-base master for a coordinate system, variable Isx93 for the coordinate
system should contain the address of this result. In this case, the time-base scale factor should be set to a
value of 131,072 divided by the real-time input frequency for the master (expressed in counts/msec), and
the coordinate system motion programmed assuming that the master is always putting out this frequency.
Source Register – Triggered
In triggered time base, the source data is directly from the encoder channel in the Servo IC – the source
address is the base address of that channel in the IC (e.g. $78000 for Servo IC 0 Channel 1, $78200 for
Servo IC 2 Channel 1). In this method, the time-base conversion calculates the 1/T interpolation itself,
computing 8 bits of fractional data, and an intermediate result in units of 1/256 of a count. This
intermediate result is stored in the first line of the entry. The extra fractional bits can reduce quantization
errors further if the input signal is of very high quality.
In the running state (method digit $A), the triggered time base takes this intermediate data, subtracts the
previous cycle’s intermediate data, multiplies the difference by 2 times the time-base scale factor, and
stores this value as its result. In the frozen state (method digit $9), the result value is always 0. In the
armed state (method digit $B), if the trigger for the channel has not occurred, the result value is 0;
however, if the trigger for the channel has occurred, the trigger-captured position is used as the previous
cycle’s value to compute the result, and the state is automatically changed to “running.”
Result Data
If the result is to be used as the time-base master for a coordinate system, variable Isx93 for the
coordinate system should contain the address of this result. In this case, the time-base scale factor should
be set to a value of 16,384 divided by the real-time input frequency for the master (expressed in
counts/msec), and the coordinate system motion programmed assuming that the master is always putting
out this frequency.
Typically, a motion program will set the method digit to $9 immediately before computing the trajectory
of the first move to be executed after the trigger. The motion program automatically suspends further
execution after computing this trajectory because this move cannot start actual execution due to the
“frozen time base” (but note that there will be no calculation delays in starting the move on the trigger.
Motor Compensation Tables and Constants
147
Turbo PMAC User Manual
Next, a PLC program will change the method digit to $B to arm the time base. Finally, the conversion
table will change the method digit to $A itself when it sees the capture trigger, as explained above.
Examples:
Untriggered time base
I8003=$079218
I8004=$403504
I8005=640
[email protected]
; 1/T conversion of Servo IC 2 Channel 4 encoder
; Untriggered time base from result of I8003
; TBSF=131072/204.8 (RTIF=204.8 cts/msec)
; Use I8005’s result as time base for C.S. 1
Triggered time base
I8008=$AF9218
I8009=256
I5293=$350A
148
; Triggered time base from PMAC2 Servo IC 2 Channel 4, running
; TBSF=16384/64 (RTIF=64 cts/msec)
; Use I8009’s result address as time base for C.S. 2
Motor Compensation Tables and Constants
Turbo PMAC User Manual
SETTING UP THE SERVO LOOP
Turbo PMAC can close a digital servo loop automatically for each activated motor. The purpose of the
servo loop is to command an output in such a way so as to try to make the actual position for the motor
match the commanded position. How well it does this depends on the tuning of the servo loop filter – the
setting of its parameters – and the dynamics of the physical system under control.
Servo Update Rate
The servo loop is closed (updated) at a frequency determined by jumpers on a Turbo PMAC, or Ivariables on a Turbo PMAC2. On a Turbo PMAC, the servo interrupt frequency is set by jumper E98,
jumpers E29-E33 (which divide down the master clock to generate the phase clock), and jumpers E3-E6
(which divide down the phase clock to generate the servo clock). On a Turbo PMAC2, the servo interrupt
frequency is set by I-variables I7m00, I7m01, and I7m02 for the Servo IC m that is the clock source for
the system (or I6800, I6801, and I6802 for a Turbo PMAC2 Ultralite). Refer to the Turbo PMAC System
Configuration and Auto-Configuration section for details on the source of these clock signals and how to
set their frequencies.
Parameter Ixx60 permits you to lower the servo-loop closure rate for an individual Motor xx by
specifying the number of servo interrupts to be skipped between closures. The default value of 0 causes
closure every interrupt. Ixx60 is useful to slow down the servo update rate for a particular motor, while
leaving the faster rate for other motors; it is also useful to test quickly whether you can get the required
performance on all motors with a slower servo update; in addition, it can be used slow the update rate
below 1 kHz. However, generally it is more computationally efficient to slow down the update rate for all
motors using the jumpers.
Reasons to Change Servo Update Rate
How fast should the servo loops be updated in your system? For most applications, the default setting of
a 2.26 kHz (442 µsec) update can be retained. There are two basic reasons to change this time:
1. Reason to Increase Rate: If not getting the dynamic performance required, speed up the servo
update rate (decrease the update time). In most systems, a faster update rate means that a stiffer and
more responsive loop can be closed, resulting in smaller errors and lags.
2. Reasons to Decrease Rate: If the routines of lower priority than the servo loop are not executing
fast enough, consider slowing down the servo update rate (increasing the update time). It may well be
updating faster than is required for the dynamic performance needed. If so, processor time is being
wasted on needless extra updates. For example, doubling the servo update time from 442 µsec to 885
µsec (halving the update rate from 2.26 kHz to 1.13 kHz), virtually doubles the time available for
motion and PLC program execution, allowing much faster motion block rates and PLC scan rates.
There are some systems that get better performance with a slower servo update rate. Generally these
are systems with relatively low encoder resolution, usually an encoder only on the load, where the
derivative gain can not be raised enough to give adequate damping without causing an unstable buzz
due to amplified quantization errors. In this case, slowing down the update rate (increasing the
update time) can help to give adequate damping without excessive quantization noise.
Ramifications of Changing the Servo Update Rate
If the servo update time is changed, many of the existing servo gains Ixx30 to Ixx39 will behave
differently. To retain equivalent servo performance, change these values. Refer to the detailed
description of each gain Ixx30-Ixx35 in the I-variable descriptions of the Software Reference Manual to
see how these change. Refer to the Notch Filter section below to see how to re-compute the notch filter
parameters Ixx36-Ixx39.
Motor Compensation Tables and Constants
149
Turbo PMAC User Manual
Changing the servo update rate changes the percentage of processor time devoted to the servo tasks,
which can have important implications for lower-priority tasks, such as motion-program and PLCprogram calculations. Refer to the Computational Features section for details on how to evaluate these
changes.
If the servo update time is changed with the jumpers, change global parameter I10 to match the change so
that commanded trajectories are executed at the right speed. I10 does not have to be changed to match
changes in Ixx60 for individual motors.
Types of Amplifiers
Turbo PMAC can interface to a variety of amplifier types. The type of amplifier used for a particular
motor or hydraulic valve has important ramifications for the tuning of the servo loop. Each of the
common types is explained below.
Amplifiers for Which Servo Produces Velocity Command
Several types of amplifiers expect a velocity command out of the Turbo PMAC servo loop. The main
types of amplifiers in this class are:
• Analog-input velocity-mode servo amplifiers
• Pulse-and-direction-input amplifiers
• Hydraulic-valve amplifiers
If the command value from the Turbo PMAC servo loop, regardless of signal type, is a velocity
command, no velocity loop needs to be closed in the Turbo PMAC. With the standard PID loop, this
means that the derivative (D) term Ixx31 can be set to 0.
Analog-Input Velocity-Mode Amplifiers
Analog-input velocity-mode servo amplifiers close a velocity loop in the amplifier using the signal from
the Turbo PMAC as the commanded velocity and sensor feedback for the actual velocity. It is vital that
the amplifier’s velocity loop be tuned properly before attempting to tune the Turbo PMAC’s servo loop
around it.
The velocity loop of a velocity-mode drive must be well tuned with the load that it will drive before the
Turbo PMAC’s position loop is tuned. Because the velocity-loop tuning is load dependent, the amplifier
manufacturer cannot do the final tuning; the machine builder must tune the loop. The velocity step
response must not have any significant overshoot or ringing; if it does, it will not be possible to close a
good position loop around it with Turbo PMAC. The Turbo PMAC Executive Program’s tuning section
has a function called Open-Loop Tuning that can be used to give velocity command steps to the amplifier
and to observe the response plotted on the screen. This makes it easy to tune the amplifier, or to confirm
that it has been well tuned.
Pulse-and-Direction-Input Amplifiers
Pulse-and-direction-input amplifiers interpret each pulse as a commanded position increment. To
generate pulse-and-direction commands, the Turbo PMAC servo loop computes a pulse frequency value
that is sent to pulse-frequency modulation circuitry. This frequency value is effectively a velocity
command.
Amplifiers with this style of interface are of two types – Stepper Drive and Stepper Replacement
Amplifier.
Stepper Drive: There is no position feedback to this drive. Usually, there is no encoder at all for these
motors, so the Turbo PMAC must use the output pulse train as simulated feedback. This requires use of
an encoder channel on Turbo PMAC, even though no encoder is physically connected. If there is an
encoder on the stepper motor, it can be used in either of two ways:
150
Motor Compensation Tables and Constants
Turbo PMAC User Manual
1. It can be used as regular feedback to the Turbo PMAC, just as on a servo motor. In this method, the
key issue is the resolution and phasing of the encoder edges relative to the steps or microsteps
produced by the drive – some deadband may have to be created with Ixx64 and Ixx65 to prevent
hunting at rest.
2. The encoder can just be used for position confirmation at the end of moves. However, this technique
requires the use of two encoder channels on the Turbo PMAC: one for the simulated feedback of the
pulse train, and one for the confirmation encoder.
Stepper-Replacement Servo Amplifiers
These take position feedback from the servo motor and close all the loops inside the drive. Do not use the
encoder signal from the drive for feedback into Turbo PMAC’s servo loop, because the position loops in
the drive and the controller will conflict with each other. With these drives, use the commanded pulse
train from the Turbo PMAC as simulated feedback.
When using simulated feedback, it is possible to set up the Turbo PMAC servo gains solely with analytic
methods. See the Setting up Turbo PMAC2 for Pulse-and-Direction Control section for details. When
using real encoder feedback, tune the servo loop just as for an analog velocity-mode drive.
Hydraulic-Valve Amplifiers
Hydraulic-valve amplifiers, whether for servo valves or proportional valves, control a fluid-volume flow
proportional to their command input. Since fluid flow into or out of a hydraulic cylinder is proportional
to the velocity of the moving member of the cylinder, the command into the valve’s amplifier is
effectively a velocity command.
Amplifiers for Which Servo Produces Torque/Force Command
Several types of amplifiers require the Turbo PMAC servo loop to close the velocity loop as well, making
the output of this servo loop a torque or force command. If Turbo PMAC is not doing commutation for
this motor, the torque/force command is output to the amplifier; if Turbo PMAC is doing the
commutation, this command is an input to the commutation algorithm. The main types of amplifiers that
require the controller to close the velocity loop are:
• Analog-input torque-mode amplifiers
• Sinusoidal-input amplifiers
• Direct-PWM power-block amplifiers
If the command value from the Turbo PMAC servo loop, regardless of signal type, is a torque or force
command, the Turbo PMAC servo must close the velocity loop for the motor. With the standard PID loop,
this means that the derivative (D) term Ixx31 must be set to a non-zero value. This derivative action is
required to get the damping action needed for stability. Because motors produce a torque or force
proportional to motor current, the torque/force command out of the servo can also be considered a current
command.
There is no need to tune anything in the amplifier with the load attached to the motor, because no
velocity-loop closure is done in these types of amplifiers. Any tuning that may be required is dependent
only on motor properties, so potentially this can even be done by the amplifier manufacturer.
Analog-Input Torque-Mode Amplifiers
Analog-input “torque-mode” amplifiers accept an analog voltage that represents a torque/force, and hence
current, command. These amplifiers close a current loop inside, and if for brushless motors, perform the
motor phase commutation as well. Another name occasionally used for these types of amplifiers is the
transconductance amplifier, signifying that a voltage input results in a proportional current output.
Sinusoidal-Input Amplifiers
A sinusoidal-input amplifier accepts two phase-current commands that are sinusoidal functions of time in
the steady state. This type of amplifier expects the controller to calculate the commutation, using the
torque/force command from the position/velocity-loop servo as the current-magnitude command into the
commutation. The amplifier performs the current-loop closure in this style.
Motor Compensation Tables and Constants
151
Turbo PMAC User Manual
Direct-PWM Power-Block Amplifiers
A direct-PWM power-block amplifier accepts phase voltage commands encoded as the actual pulsewidth-modulated on-off commands for the power transistors. This type of amplifier expects the controller
to calculate the commutation and current loop, using the torque/force command from the
position/velocity-loop servo as the current-magnitude command into the commutation and current loop.
The amplifier performs no control functions in this style.
PID/Feedforward/Notch Servo Filter
Turbo PMAC’s PID/feedforward/notch servo filter (PID filter for short) is by far the most commonly
used servo filter. It is easy to understand and tune, yet powerful enough to provide excellent control of
the vast majority of systems.
The PID filter is selected by default on the Turbo PMAC. To use the PID filter on Motor xx, make sure
that bit 0 of Ixx59 is set to 0, disabling any user-written servo algorithm, and Iyy00/50 (I3300 for Motor
1, I3350 for Motor 2, etc.) is set to 0, disabling the Extended Servo Algorithm.
PMAC PID + NOTCH Servo Filter
K (1-z -1 )
vff
K
Reference
Position
+
-
-1
aff
-2
(1-2z +z )
Big Step/
Deadband
Filter
Notch Filter
+
+
K
-
+
-1
-2
-1
-2
1+n1 z +n 2z
p
1+d z +d z
1
2
Notch Coefficients
n1 : Ix36
IM
Ki
1-z -1
n 2 : Ix37
K
d
d 1 : Ix38
d 2 : Ix39
K : Proportional Gain (Ix30)
p
K : Derivative Gain (Ix31)
d
K : Velocity Feedforward Gain (Ix32)
vff
K : Integral Gain (Ix33)
i
IM : Integration Mode (Ix34)
K : Acceleration Feedforward Gain (Ix35)
aff
1-z-1
Secondary
Position
("Velocity")
Feedback
(usually
the same)
Primary
Position
Feedback
PID Feedback Filter
The PID feedback filter consists of proportional (P), integral (I), and derivative (D) terms, each with its
own contribution to the control effort.
Ixx30 Proportional Gain Term
The proportional gain term set by Ixx30 provides the basic corrective action for position errors, providing
a control effort proportional to the size of the error to try to reduce the error. Proportional gain alone acts
like a spring, and the magnitude of the proportional gain term is the spring constant; the higher this gain
term, the stiffer the spring action.
In Turbo PMAC, Ixx30 is an overall loop gain term, post-multiplying the other gain terms, and not just a
proportional gain term. This makes it possible to change Ixx30 alone if an external gain term (e.g.
encoder resolution or amplifier gain) is changed; it also normalizes the derivative and integral gain terms
into a time constant and inverse time constant, respectively.
152
Motor Compensation Tables and Constants
Turbo PMAC User Manual
Ixx31 Derivative Gain Term
The derivative gain term set by Ixx31 provides a damping effect by providing a contribution to the control
effort proportional to the actual velocity acting against that velocity. In this respect it acts much like a
dashpot or the shock absorber of a vehicle’s suspension. The higher the derivative gain term, the heavier
the damping action.
Some form of derivative action – effectively a velocity loop – is required for a stable position loop. If a
well-tuned velocity loop is closed in the amplifier, the Ixx31 derivative gain term in Turbo PMAC can be
set to 0. However, if there is no velocity loop closed externally, a positive value of Ixx31 will be required
for stable operation.
Note that in the Turbo PMAC, this gain acts on the derivative of the actual position, not on the derivative
of the position error, as in some other controllers. This permits the simple use of dual motor-and-load
feedback with a separate sensor on the motor for derivative action (specified by Ixx04) from the sensor on
the load for proportional and integral action (specified by Ixx03).
Ixx33 Integral Gain Term
The integral gain term set by Ixx33 provides for correction against steady-state errors caused by such
effects as friction, gravitational loads, cutting loads, and analog offsets. The integral gain term controls
how fast the position error integrator term charges up and discharges; the higher the gain, the faster the
action,
Ixx34 is a single-bit variable that controls the time in which the integral gain term is active. At the default
value of 1, the integral gain is active only when the commanded velocity is zero (at move end). When
Ixx34 is 0, the integral gain term is always active. When tuning the other terms, generally it is best to set
Ixx34 to 1 to reduce the errors as much as possible without the integrator; if the remaining errors are
small, it is then usually good practice to set Ixx34 to 0 to let the integrator dynamically compensate for
the remaining errors.
Feedforward Filter
A feedback filter is error driven, so an error must exist between the commanded and actual positions
before it takes any action. The actions of feedforward, on the other hand, are dependent only on the
commanded trajectory, and therefore do not require errors to cause action. The basic idea of feedforward
is to directly apply your best estimate of the control effort needed to execute the commanded trajectory,
without waiting for position errors to build up. The feedback terms then need to respond only to the
errors in this estimate, which typically are quite small.
In a well-tuned system, over 95 percent of the control effort can come from the feedforward terms, with
the feedback terms just providing small corrections for disturbances and imperfections in the estimate.
Turbo PMAC’s PID filter has velocity and acceleration feedforward terms, covered here, and a non-linear
friction feedforward term, covered below under the heading Servo Loop Modifiers.
Ixx32 Velocity Feedforward Term
The velocity feedforward term Ixx32 adds an amount to the control effort that is directly proportional to
the commanded velocity, to overcome potential position errors that would be proportional to velocity.
These errors can come from several sources. The first source, and the dominant one, is from the velocity
feedback term that provides the required damping for stability, whether done in the Turbo PMAC (the
Ixx31 term) or externally. Other minor sources of velocity related errors include magnetic losses in the
motor and actual viscous damping losses.
Properly set velocity feedforward will eliminate following error components that are proportional to
velocity. If the Turbo PMAC is closing the velocity loop for the motor, typically the optimal Ixx32 will
be equal to, or slightly greater than Ixx31.
Motor Compensation Tables and Constants
153
Turbo PMAC User Manual
Ixx35 Acceleration Feedforward Term
The acceleration feedforward term Ixx35 adds an amount to the control effort that is directly proportional
to the commanded acceleration, to overcome potential position errors that would be proportional to
acceleration. These errors come from the fundamental tendency of inertia to resist acceleration. Without
acceleration feedforward, there would be a component of the following error proportional to acceleration.
Properly set acceleration feedforward will eliminate following error components that are proportional to
acceleration. The Ixx35 acceleration feedforward term is an estimate of the inertia of the system, directly
providing a force or torque proportional to it and the commanded acceleration.
Actual PID/Feedforward Algorithm
The actual equation used in the PID/feedforward algorithm to compute the commanded output for Motor
xx is as follows:
CMDout(n) = 2-19 * Ixx30*[{Ixx08 * [FE(n) + (Ixx32*CV(n) +Ixx35*CA(n))/128 +
Ixx33*IE(n)/223]} -Ixx31*Ixx09*AV(n)/128]
where:
• CMDout(n) is the 16-bit output command (-32768 to +32767) in servo cycle n. It is converted to a 10V to +10V output. DACout(n) is limited by Ixx69.
• Ixx08 is an internal position scaling term for Motor xx (usually set to 96)
• Ixx09 is an internal scaling term for the velocity loop for Motor xx (usually set to 96)
• FE(n) is the following error in counts in servo cycle n, which is the difference between the
commanded position and the actual position for the cycle [CP(n) - AP(n)]
• AV(n) is the actual velocity in servo cycle n, which is the difference between the last two actual
positions [AP(n) - AP(n-1)] in counts per servo cycle
• CV(n) is the commanded velocity in servo cycle n: the difference between the last two commanded
positions [CP(n) - CP(n-1)] in counts per servo cycle
• CA(n) is the commanded acceleration in servo cycle n, which is the difference between the last two
commanded velocities [CV(n) - CV(n-1)] in counts per servo cycle
• IE(n) is the integrated following error in servo cycle n, which is:
n−1
∑ [ FE( 1 )
j=0
(for all servo cycles for which the integration is active. Ixx34=1 turns off the input to, but not the
output from the integrator when CV does not equal zero.)
Notch Filter
Turbo PMAC’s standard servo loop includes a notch filter. This is a second-order bi-quad filter acting on
the output of the PID section of the servo loop, one of whose main purposes is to create a notch
(frequency of low response) in the servo reaction for the purposes of fighting a resonance.
This filter has several possible uses:
• Anti-resonance (notch) filter
• Low-pass filter
• Velocity-loop integrator
• Lead-lag filter
Each use will be treated in its own section below.
154
Motor Compensation Tables and Constants
Turbo PMAC User Manual
Filter Structure
For those familiar with control theory (not necessary to use the filter!), the form of Turbo PMAC’s notch
filter system is:
N ( z ) 1 + N 1 z −1 + N 2 z −2 1 + Ixx36 z −1 + Ixx37 z −2
=
=
D( z ) 1 + D1 z −1 + D2 z −2 1 + Ixx38 z −1 + Ixx39 z −2
The I-variables Ixx36, Ixx37, Ixx38, and Ixx39 each have a range of -2.0 to +2.0; they are 24-bit values,
with one sign bit, one integer bit, and 22 fractional bits.
Use to Create a Notch
In feedback controls, a notch filter is an anti-resonance filter used to counteract a physical resonance.
While there are many different philosophies as to how to set up a notch filter, we recommend setting up a
lightly damped band-reject filter at about 90 percent of the resonant frequency, and a heavily damped
band-pass filter at a frequency somewhat greater than the resonant frequency (to reduce the highfrequency gain of the filter itself). The band-reject filter is implemented in the numerator of the filter
[N(z)], creating zeros in control terms; the band-pass filter is implemented in the denominator of the filter
[D(z)], creating “poles” in control terms.
Automatic Notch Specification
The Turbo PMAC Executive program allows you to set up a notch filter very simply, without the need to
understand how a notch filter works. The easiest way is to enter the frequency of the mechanical
resonance that you wish to control. The Executive program will compute the desired characteristics of
the band-reject and band-pass filters, compute the coefficients, and download them to Turbo PMAC.
Alternatively, you can individually specify the desired characteristics of the band-reject and band-pass
filters. The two characteristics for each part of the filter are the natural frequency ωn and the damping
ratio ζ. The Executive program will compute the coefficients to achieve those characteristics, and
download them to Turbo PMAC.
Manual Notch Specification
To calculate the notch filter coefficients manually, consider the continuous transfer function for a notch
filter:
2
s 2 + 2ζ z ω nz + ω nz
G (s ) =
2
s 2 + 2ζ p ω np + ω np
Start with five parameters for the filter:
• ωnz: the natural frequency of the zeroes in radians/second (not in Hertz)
• ζz: the damping ratio of the zeroes
• ωnp: the natural frequency of the poles in radians/second (not in Hertz)
• ζp: the damping ratio of the poles
• Ts: the servo-loop sampling period in seconds (not in msec)
To compute radians/second from Hertz, multiply by 2π (6.283). To compute the sampling period in
seconds from the sampling rate in kHz, first multiply the rate by 1000 to get Hz, and then take the
reciprocal. Remember that the sampling period is equal to (Ixx60+1) times the servo-interrupt period.
First, compute the following intermediate values:
2 T2
α z = 1 + 2ζ z ω nz Ts + ω nz
s
2 T2
α p = 1 + 2ζ p ω np Ts + ω np
s
Motor Compensation Tables and Constants
155
Turbo PMAC User Manual
Then compute the filter coefficients:
Ixx36 =
− (2ζ z ω nz Ts + 2 )
αz
Ixx37 =
Ixx38 =
(
1
αz
− 2ζ p ω np Ts + 2
)
αp
Ixx39 =
1
αp
Finally, modify the proportional-gain term to compensate for the DC-gain change that the filter creates:
2
ω np
αz
Ixx30 new = Ixx30old
2 α
ω nz
p
For example, suppose we have identified a 55 Hz resonance in our mechanical coupling. To compensate
for this, we decide to put a lightly damped band-reject filter (damping ratio 0.2) at 50 Hz natural
frequency, and a heavily damped band-pass filter (damping ratio 0.8) at 80 Hz natural frequency to limit
the high-frequency gain of the filter. The servo update time is the default of 442 microseconds.
Ts = 442 µ sec* 10 −6
sec
µ sec
= 0.000442 sec
ω nz = 2π * 50 Hz = 314.2 rad sec
ω np = 2π * 80 Hz = 502.7 rad sec
2 T2
α z = 1 + 2ζ z ω nz Ts + ω nz
s
= 1 + 2 * 0.2 * 314.2 * 0.000442 + 314.2 2 * 0.000442 2
= 1.0748
2 T2
α p = 1 + 2ζ p ω np Ts + ω np
s
= 1 + 2 * 0.8 * 502.7 * 0.000442 + 502.7 2 * 0.000442 2
= 1.4049
Next we compute the filter coefficients:
− (2ζ z ω nz Ts + 2 )
Ixx36 =
αz
=
Ixx37 =
Ixx38 =
(
− 2ξ p ω np Ts + 2
αp
)
=
Ixx39 =
− (2 * 0.2 * 314.2 * 0.000442 + 2 )
1.0748
1
αz
=
1
1.0748
= 0.930
− (2 * 0.8 * 502.7 * 0.000442 + 2 )
1.4049
1
αp
=
1
1.4049
= −1.912
= −1.677
= 0.712
Finally, we compute the DC gain adjustment, assuming for the example that our existing proportional
gain term Ixx30 had been 500,000:
156
Motor Compensation Tables and Constants
Turbo PMAC User Manual
Ixx30 new = Ixx30old
2
ω np
αz
502.7 2 1.0748
= 500 ,000
= 979 ,169
2
314.2 2 1.4049
ω nz α p
Use to Create a Low-Pass Filter
It is also possible to use this filter component as a low-pass filter if reducing roughness of operation is
more important than high system bandwidth. Typically, the low-pass filter is used if a low-resolution
position sensor is used.
Automatic Specification
The Turbo PMAC Executive program allows the set up of a low-pass filter easily, without the need to
understand how it works. Enter the cutoff frequency of the filter (the frequency above which should not
pass much signal strength) and choose whether to create a first-order or second-order filter. The
Executive program will compute the desired characteristics of the band-reject and band-pass filters,
compute the coefficients, and download them to Turbo PMAC.
Manual Specification
To calculate a low-pass filter manually, specify the cutoff frequency in radians per second, and the servoupdate frequency in seconds.
First-Order Filter: To calculate a first-order low-pass filter, consider the continuous transfer function
for the filter:
ω
1
F (s ) =
s/ω +1
=
s +ω
where ω is the cutoff frequency in radians per second (equal to 2πf, where f is the cutoff frequency in
Hertz). This value ω is equal to 1/τ, where τ is the time constant of the filter.
Next, convert this to a discrete-time transfer function using the approximation s=(1-z-1)/Ts, where Ts is the
servo update time in seconds.
F (z ) =
ω
1 − z −1
Ts
=
+ω
ωTs
1 − z −1 + ωTs
=
ωTs
1 + ω Ts
1
*
1−
1
1 + ωTs
z −1
In Turbo PMAC terms, the gain term is multiplied into the existing gain term Ixx30:
ωTs
Ixx30 new = Ixx30old *
1 + ωTs
The pole term uses the first-order notch filter pole parameter Ixx38. The other filter parameters Ixx36,
Ixx37 and Ixx39 are set to zero if the filter is used only as a low-pass filter.
Ixx38 = −
1
1 + ωTs
For example, to implement a first-order low-pass filter with a cutoff frequency of 50 Hz on a Turbo
PMAC with a servo update time of 442 µsec, we compute:
ωTs = 2* π * 50* 0.000442 = 0.139
Ixx30 new = Ixx30old *
0.139
= 0.122 * Ixx30old
1 + 0.139
1
Ixx38 = −
= −0.877
1 + 0.139
Ixx36 , Ixx 37 , Ixx39 = 0
Motor Compensation Tables and Constants
157
Turbo PMAC User Manual
Second-Order Filter: To calculate a second-order low-pass filter, we consider the continuous transfer
function for a generalized second-order filter:
ω n2
F (s ) =
s 2 + 2ζω n + ω n2
where ωn is the cutoff frequency of the filter in radians per second, and ς is the damping ratio – a value of
0.707 produces a Butterworth filter here.
First, compute the following intermediate value:
α = 1 + 2ζω n Ts + ω n2 Ts2
Then compute the filter coefficients:
Ixx38 =
− (2ζω n Ts + 2 )
α
Ixx39 =
1
α
Ixx36 and Ixx37 should be set to 0 if no other use is made of this filter.
Finally, modify your proportional-gain term to compensate for the DC-gain change that the filter creates:
Ixx30 new = Ixx30old
ω n2 Ts2
α
For example, to implement a second-order low-pass filter with a cutoff frequency of 60 Hz and a damping
ratio of 0.707 on a Turbo PMAC with a servo update time of 250 µsec, we compute the following:
ω n Ts = 2 * π * 60 * 0.000250 = 0.0942
a = 1 + 2 * 0.707 * 0.0942 + 0.0942 2 = 1.146
− (2 * 0.707 * 0.0942 + 2 )
Ixx38 =
= −1.861
1.146
Ixx39 =
Ixx30 new = Ixx30old
1
= 0.873
1.146
0.0942 2
= 0.00774 * Ixx 30old
1.146
Ixx36 , Ixx 37 = 0
Use to Create a Velocity-Loop Integrator
This filter can also be used to create an integrator inside the velocity loop, independent of the integral
gain term in the position loop. This additional integrator can provide additional stiffness and disturbance
rejection. However, it may hinder quick response to acceleration commands.
Manual Specification
Consider a PI filter in the velocity loop with transfer function:
V (z ) = K pv +
K iv
1 − z −1
where Kpv is the velocity-loop proportional gain, and Kiv is the velocity-loop integral gain. This can be
manipulated to produce:
V (z ) =
K pv (1 −
z −1
1−
) + K iv
z −1
=
K pv + K iv − K pv
1 − z −1
z −1
= (K pv


K pv
1 −
z −1 


K pv + K iv

+ K iv ) 
−
1
1− z
In Turbo PMAC terms, the gain term (Kpv+Kiv) is multiplied into the existing gain term Ixx30:
158
Motor Compensation Tables and Constants
Turbo PMAC User Manual
Ixx30new = Ixx30old * (K pv + K iv )
The zero and pole terms use the first-order notch filter parameters Ixx36 and Ixx38, respectively. The
second-order parameters Ixx37 and Ixx39 are set to zero if the filter is used only as an integrator.
Ixx36 = −
K pv
K pv + K iv
Ixx38 = −1
Use to Create a Lead-Lag Filter
This filter can be used simply as a lead-lag filter if the roots are real rather than imaginary. A lead-lag
filter is very similar in performance to a PID filter. It is useful when filter settings are determined
analytically rather than experimentally. When a basic lead-lag servo filter is desired, all servo gains
Ixx31 to Ixx35 should be set to zero; Ixx30 is still used as the generalized gain term.
The PMAC Executive program presently does not have any screens to assist in the automatic
specification of a lead-lag filter.
Manual Specification
The generalized analytical form of a digital lead-lag filter is:
L( z ) = K
(z + a ) (z + c )
(z + b ) (z + d )
where the (z+a)/(z+b) term is the lead filter, with a < b, the (z+c)/(z+d) term is the lag filter, with c > d, and
K is the DC gain term. In Turbo PMAC’s real-time implementation, the transfer function of the filter is:
L( z ) = K
1 + acz −1 + c 2 z −2
1 + bdz −1 + d 2 z −2
Turbo PMAC term Ixx30 is set to K; Ixx36 is set to ac; Ixx37 is set to c2; Ixx38 is set to bd; and Ixx39 is
set to d2.
Servo-Loop Modifiers
The PID filter has several modifying terms – non-linearities in control terminology – that can be
important to optimize the filter for performance and safety. Each is covered briefly below.
Motor Compensation Tables and Constants
159
Turbo PMAC User Manual
Ixx63: Integration Limit
Ixx63 is a saturation limit for the integrator, which limits the magnitude of the integrator output. If set to
a negative value, it trips the servo loop with a fatal following error if the integrator saturates. Setting
Ixx63 to 0 clears the integrator and thus its output. (Setting integral gain term Ixx33 to 0 only stops
further input to the integrator.)
Ixx64, Ixx65: Deadband Compensation
Ixx64 and Ixx65 permit both the creation of a deadband and compensation for physical deadband (as in
proportional hydraulic valves). Ixx65 specifies the magnitude of the deadband compensation zone in
terms of magnitude of the servo following error. Ixx64 specifies the gain inside that zone relative to the
overall proportional gain Ixx30. Positive values of Ixx64 increase the gain inside the zone, compensating
for physical deadband; negative values of Ixx64 decrease the gain inside the zone, with Ixx64=-16
producing true deadband.
n=-16
23
160
Motor Compensation Tables and Constants
Turbo PMAC User Manual
Ixx67: Following Error Limit
Ixx67 is a saturation limit on the magnitude of the following error input to the P and I terms of the filter.
It does not limit the true following error, or the error value compared to the Ixx11 and Ixx12 following
error limit parameters. Setting Ixx67 to 0 disables the PI control terms, while leaving the D derivative
term (i.e. the velocity loop) and the feedforward terms active. This setting is useful if only velocity
control is truly desired. If Ixx67 is set to 0, Ixx11 should also be set to 0 to make sure the motor does not
trip on a fatal following error fault.
Ixx68: Friction Feedforward
Ixx68 is a non-linear friction feedforward term. The magnitude of Ixx68, multiplied by the sign of the
instantaneous commanded velocity, is added directly to the output of the servo filter. It is meant to
compensate for dry (Coulomb) friction.
Ixx69: Output Limit
Ixx69 is a saturation limit on the output of the servo filter. If the output value is limited by Ixx69, the
input to the position-loop integrator is turned off automatically, and the output of the integrator remains
constant in this condition.
Ixx29, Ixx79: Offset Terms
If Turbo PMAC is not performing commutation for Motor xx, Ixx29 is a fixed offset term on the output
of the servo filter, in units of a 16-bit DAC (even if some other device is used). If Turbo PMAC is
performing commutation for the motor, Ixx29 and Ixx79 serve as phase offsets in the commutation
algorithm.
Extended Servo Algorithm
For systems with more difficult dynamics, such as multiple resonances and low-frequency resonances, the
Extended Servo Algorithm (ESA) may be used instead of the PID filter. The choice of servo filter may be
made on a motor-by-motor basis.
The ESA, while more powerful and flexible than the PID, is not as easy to understand or tune
interactively. The PEWIN32PRO Executive program has auto-tuning software for the ESA. However,
tuning the ESA manually requires significant control-theory knowledge and experience. This discussion
assumes such knowledge.
The ESA has 30 terms in seven blocks of polynomial coefficients. As with the PID, it does support single
and dual feedback using the Ixx03 and Ixx04 feedback address variables. However, the ESA is more
flexible with regard to what is done with dual feedback; it is not limited to separate position and velocity
loops. The feedback selected with Ixx03 and the commanded trajectory values are pre-multiplied by the
Ixx08 scale factor; the feedback selected with Ixx04 is pre-multiplied by the Ixx09 scale factor.
As with the PID, or a user-written servo, the ESA can be used with or without Turbo-PMAC based
commutation. If the motor is not commutated by Turbo PMAC, the computed control output is written to
the register specified by Ixx02; if it is commutated by Turbo PMAC, the control output is the torque
(quadrature) command into the commutation algorithm.
Motor Compensation Tables and Constants
161
Turbo PMAC User Manual
PMAC: Extended Servo Algorithm Block Diagram
1 - z -1
Ix68
KS (k 0 + ⋅ ⋅ ⋅ + k 3z -3 )
h 0 + h1 (1 - z -1)
1 + 4 (L1z −1 + ⋅ ⋅ ⋅ + L3z -3 )
r(z)
+
32 ⋅ Ix08
s0
TS (t 0 + ⋅ ⋅ ⋅ + t 3z -4 )
+
-
-
+
1
+
1 + 8 (r1z −1 + ⋅ ⋅ ⋅ + r4 z - 4 )
+
+
1 + d1z −1d 2 z - 2
Ix69
+
+
u(z)
15
2
DAC
-1
s1 (1 - z )
32 ⋅ Ix08
θ1(z)
Encoder #1
Encoder #2
Feedback
Loop #1
+
f0
GS (g 0 + g1 (1 - z -1 ))
+
-
-
f1 (1 - z -1 )
32 ⋅ Ix08
θ2(z)
Encoder #1
Encoder #2
Feedback
Loop #2
The ESA for Motor xx is selected by setting the first supplemental motor I-variable Iyy00/50 (I3300 for
Motor 1, I3350 for Motor 2, etc. – see full table in the Software Reference Manual) to 1, and by setting bit
0 of Ixx59 to the default of 0 to disable the “user-written servo.” With this setting, the servo loop terms
are supplemental I-variables Iyy10/60 through Iyy39/89 (I3310 – I3339 for Motor 1, I3360 – I3389 for
Motor 2, etc.). The following table shows the variables used for each gain term:
I-Var. for
OddNumbered
Motors
Iyy10
Iyy11
Iyy12
Iyy13
Iyy14
Iyy15
Iyy16
Iyy17
Iyy18
Iyy19
Iyy20
Iyy21
Iyy22
Iyy23
Iyy24
I-Var. for
EvenNumbered
Motors
Iyy60
Iyy61
Iyy62
Iyy63
Iyy64
Iyy65
Iyy66
Iyy67
Iyy68
Iyy69
Iyy70
Iyy71
Iyy72
Iyy73
Iyy74
Gain
Name
Range
s0
s1
f0
f1
h0
h1
r1
r2
r3
r4
t0
t1
t2
t3
t4
-1.0≤Var<+1.0
-1.0≤Var<+1.0
-1.0≤Var<+1.0
-1.0≤Var<+1.0
-1.0≤Var<+1.0
-1.0≤Var<+1.0
-1.0≤Var<+1.0
-1.0≤Var<+1.0
-1.0≤Var<+1.0
-1.0≤Var<+1.0
-1.0≤Var<+1.0
-1.0≤Var<+1.0
-1.0≤Var<+1.0
-1.0≤Var<+1.0
-1.0≤Var<+1.0
I-Var. for
OddNumbered
Motors
Iyy25
Iyy26
Iyy27
Iyy28
Iyy29
Iyy30
Iyy31
Iyy32
Iyy33
Iyy34
Iyy35
Iyy36
Iyy37
Iyy38
Iyy39
I-Var. for
EvenNumbered
Motors
Iyy75
Iyy76
Iyy77
Iyy78
Iyy79
Iyy80
Iyy81
Iyy82
Iyy83
Iyy84
Iyy85
Iyy86
Iyy87
Iyy88
Iyy89
Gain
Name
Range
TS
L1
L2
L3
k0
k1
k2
k3
KS
d1
d2
g0
g1
g2
GS
-223≤Var<223
-1.0≤Var<+1.0
-1.0≤Var<+1.0
-1.0≤Var<+1.0
-1.0≤Var<+1.0
-1.0≤Var<+1.0
-1.0≤Var<+1.0
-1.0≤Var<+1.0
-223≤Var<223
-1.0≤Var<+1.0
-1.0≤Var<+1.0
-1.0≤Var<+1.0
-1.0≤Var<+1.0
-1.0≤Var<+1.0
-223≤Var<223
The ESA consists of a series of blocks, most with multiple terms, each taking an input value, which could
be the output of another block, and computing an output value, which could be the input to another block.
Many of the blocks have polynomial transfer functions; an nth order polynomial implies the storage of n
cycles of history for the block.
Terms whose names consist of a letter and a number multiply a single control value that is i cycles old,
where i is the number in the name (e.g. t2 multiplies a value two cycles old). If the term is in the
numerator of the block, it multiplies an input value to that block; if it is in the denominator of the block, it
multiplies an output value from that block. These terms have a range of +/-1.0, with 24-bit resolution.
162
Motor Compensation Tables and Constants
Turbo PMAC User Manual
Terms whose names consist of two letters, with the second letter an S, multiply the results of an entire
block. These terms are treated as integers with a range of +/-8,388,608.
The PID terms Ixx30 – Ixx39, Ixx63 – Ixx65, and Ixx67 are not used. Ixx68 is used as the “friction
feedforward” term for the ESA, just as it is for the PID. Ixx69 is used for the ESA, but in a slightly
different manner from the PID. In the PID, Ixx69 is a truncation limit on the control effort output that
does not affect smaller command values; in the ESA it is an output scale factor that affects all output
command values.
Cascading Servo Loops
The open structure of Turbo PMAC’s servo loops and the ability to specify which registers are used for
its inputs and outputs provide the user with powerful capabilities such as the ability to “cascade” servo
loops. In this technique, the output of one servo loop (one Turbo PMAC “motor”) is used as an input to
another servo loop, bringing the capabilities of both loops to bear on a single actuator. The outer loop
does not drive an actuator directly; instead, it dynamically modifies the set point of the inner loop in an
effort to drive its own error to zero.
This technique has many possible uses. The most common is to be able to close an auxiliary loop around
a standard position loop. The auxiliary loop controls some quantity affected by the position loop’s
motion, such as torque or force applied, or distance from a surface. The coupling of the loops can be
turned on and off, permitting easy switching between control modes.
Common uses of this technique include:
• Web tensioning
• Torque-limited screwdriving
• Metal bending
• Controlled-force part insertion
• Height control over uneven surface (e.g. for auto-focus)
The inner loop in these applications is typically a standard position loop driving a real actuator with a
standard position feedback device such as an encoder or resolver. The first step in setting up such an
application is to get this loop working in standard positioning mode (running at continuous velocity if
appropriate).
The outer loop in these applications uses a feedback sensor measuring whatever quantity the outer loop is
to be controlled. Often these force or torque transducers such as strain gages or tensioning dancer arms,
or distance (gap) transducers employing capacitive or ultrasonic mechanisms.
By engaging and disengaging the outer loop, the user can switch between standard position control using
just the inner loop, as when not meeting the resistance of a surface, and control of the auxiliary function,
as when pushing with controlled force against a surface. The transition is simple to perform, and smooth
in operation.
A second use of this technique is to build a more complex filter than you can with the standard filter for a
single motor (e.g. incorporating a double notch filter). By using the output of the first filter as the input to
the second, you can chain them together and get the action of both filters between the command and the
output. While the general principle is the same, the details of the setup and the process for getting this
going will differ. The sections immediately following cover the process for setting up hybrid control. A
special section further down describes the differences in setting up using two loops to control a single
quantity.
Motor Compensation Tables and Constants
163
Turbo PMAC User Manual
Selecting Turbo PMAC Motors to Use
Any two Turbo PMAC motors can be used for the inner and outer loops. If no integration is required in
passing the information from outer loop to inner loop (see below), using a lower-numbered motor for the
outer loop will avoid adding a servo-cycle delay. This has the possibility of delivering higher
performance in closing the outer loop.
However, in many applications, the performance of the outer loop will be high enough even with the
added delay, and some users will find it easier to add the additional outer-loop motor as a highernumbered motor than the real motors. For example, Motors 1, 2, and 3 are the real X, Y, and Z axes,
respectively; Motor 4 is added as the W axis to close a loop around motor 3 as the Z axis.
Inner Loop General Setup
In hybrid control applications, first set up the inner loop as a standard positioning motor, get it well tuned
and operating as needed when controlling this actuator in position/velocity mode. For the hybrid control,
we will simply add a command from the outer loop through the master position input; otherwise,
operation of this inner loop remains the same.
Outer Loop General Setup
Set up the outer loop to use the alternate sensor as its feedback with its Ixx03 and Ixx04 feedback address
parameters (through the conversion table). Often, the control quantity of this loop will not be a position
value, but since all Turbo PMAC terminology is in terms of position, you should beware of possible
confusion. One least-significant bit (LSB) of the feedback sensor is considered a count by Turbo PMAC.
As with positioning motors, the scale factor of the axis-definition statement can permit you to program
this motor in engineering units (e.g. Newtons or pounds of force).
You will not be able to tune the outer loop until you have linked it with the inner loop. The next section
describes the steps in linking the loops.
Cascaded Loops for Hybrid Control
Virtual PMAC Motor
Commanded
Auxiliary
Trajectory
Ixx02
+
or
Ixx02
Servo
Filter
PMAC Motor for Physical Device
Ixx05
Σ∆t
or
Ixx05
Open
Register
-
Ixx06
+
Commanded
Position
Trajectory
Auxiliary
Loop
164
Ixx02
Output
Device
Amp
Motor
Position Loop
Actual
Position
Ixx03,
Ixx04
Servo
Filter
+
Actual
Auxiliary
Value
ECT
Entry
Ixx03,
Ixx04
ECT
Entry
ECT
Source
Address
E
ECT
Source
Address
S
M
Position
Sensor
Auxiliary
Sensor
Motor Compensation Tables and Constants
Turbo PMAC User Manual
Joining the Loops
After the inner loop is working properly, and have done the basic setup of the outer loop, join the loops
together.
To Integrate Outer Loop Command or Not
Sometimes the output from the outer loop will be numerically integrated before being used as a position
input to the inner loop. If it is integrated, the outer loop’s command itself will effectively be a velocity
value; if it is not integrated, the value will be a position value. In general, if the steady-state condition
with the outer loop engaged has a non-zero velocity, as in web-tensioning applications, this value should
be integrated. However, if the steady-state condition is at zero velocity, as in part-insertion applications,
integration should not be used.
Outer Loop Output Address: Ixx02
The key to this technique is to link the output of the outer loop to the master position of the inner loop.
The Ixx02 variable for the outer loop’s motor that specifies the address of the servo output can specify
any open register. This is just a temporary holding register for the value, so it does not need to be the
register of an output device. Commonly the open RAM registers in the address range $0010F0 - $0010FF
are used to hold this value.
If Ixx01 for this motor is set to 2, the register specified by Ixx02 will be an X register. (Note that the
outer-loop motor will never be commutated, so it will never be necessary to set bit 0 of Ixx01 to 1.) If
you do not need to integrate this value before using it for the inner loop (the decision is discussed below),
writing this value to an X register will permit the inner loop to read this value directly, without any
processing by the encoder conversion table. However, if the value must be integrated, Ixx01 for the
outer-loop motor should be set to 0 so the value is written to a Y register.
Integrating in the Conversion Table
If the command output of the outer loop is integrated before using it as an input into the inner loop, treat
the holding register the same as the register from an Acc-28 A/D converter. A conversion-table entry that
integrates a register of Acc-28 style has a leading hex method digit of $5. The register is to be treated as a
signed quantity, so the bit-19 mode switch bit is set to 0. For example, if the holding register were at
address Y:$0010F0, the entry’s first line would be $5010F0.
The second line of the entry is an offset value that is subtracted from the reading before integration. Since
this is using a computed and not measured value, this should be set to 0.
Inner Loop Master Address: Ixx05
The Ixx05 master position address variable for the inner loop’s motor contains the address of the result in
the encoder conversion table. This must be an X register. If the outer loop wrote its output directly to an X
register, the inner-loop motor’s Ixx05 can contain the address of this same register. For example, if Motor 4
were to pick up a value left in open register X:$10F1, this could be specified with I405=$10F1.
If the outer-loop’s command were processed through the conversion table, this Ixx05 will contain the
address of the result in the conversion table. Remember that you can set this address by reference to the
conversion-table I-variable number. If the conversion-table entry were set up with I8004 and I8005, you
could specify the use of the resulting value for Motor 5 with [email protected]
Inner-Loop Following Enable and Mode: Ixx06
The Ixx06 variable for the inner loop’s motor controls whether or not the outer loop is engaged. When bit
0 of Ixx06 is set to 0, the outer loop is not engaged, and the inner loop will function independently. When
bit 0 of Ixx06 is set to 1, the outer loop is engaged, and its output will command a modification to the
total commanded position of the inner loop.
Motor Compensation Tables and Constants
165
Turbo PMAC User Manual
Ixx06 for the inner loop’s motor also controls how the outer loop’s corrections interact with commanded
positions for the inner loop. When Ixx06 bit 1 (the following mode control bit) is set to 0, the inner
loop’s commanded positions are relative to a fixed origin, and these commanded moves effectively cancel
out whatever corrections have come in through the master position port. When Ixx06 bit 1 is set to 1
(offset mode), the corrections that come in through the master position port effectively offset the origin
for programmed commanded moves, permitting commanded moves and master corrections to be
superimposed. This distinction in mode is true even if following is disabled.
When the outer loop is engaged (Ixx06 bit 0 = 1), usually the following must be in offset mode (Ixx06 bit
1 =1), making the required value of Ixx06 be 3 for this operation. Even if there are no explicit commands
in the motion program for the axis assigned to the inner loop’s motor at this time, any motion command
for the coordinate system containing this motor implicitly commands that motor to its previous
commanded position. If the following is not in offset mode, this will take out the corrections that have
come in since the last programmed move or move segment.
When the following is disabled (Ixx06 bit 0 = 0), to command the inner loop’s motor to a definite
physical position, put the following in normal mode (Ixx06 bit 1 = 0), making the required value of Ixx06
be 0 for this operation.
Inner-Loop Master Scale Factor: Ixx07
The Ixx07 variable for the inner loop’s motor, the master scale factor, is a gain term for the outer loop in
this use. Set to 1 to keep the net outer-loop gain (inner-loop Ixx07 times outer-loop Ixx30) as low as
possible. It can be set to a negative value if necessary to invert the sense of the coupling between the two
loops.
Tuning the Outer Loop
In the cases of hybrid control, typically you will need only proportional gain (Ixx30) in the outer loop or
possibly integral gain as well (Ixx33). Most applications will require no derivative gain (Ixx31), and
because in most applications the outer loop is just trying to maintain a constant command value, usually
feedforward terms (Ixx32 and Ixx35) are not important.
If you are integrating the outer loop’s command value before using it in the inner loop, your Ixx30
proportional gain term probably will be extremely low (often around 10).
It is possible to use the Executive’s tuning tools to tune the outer loop gains as you would a standard loop.
Programming the Outer Loop Motor
With the outer loop engaged, commanding the position of the outer-loop motor will cause the outer loop’s
feedback loop to calculate offsets into the inner loop command in an attempt to drive the outer-loop’s
feedback device to the commanded value. This outer-loop command can be a motor jog command, or it
can be a programmed axis command. If a programmed axis command, the axis to which the outer-loop
motor is assigned can be in the same coordinate system as the inner-loop motor, or in a different
coordinate system.
Most commonly, the outer-loop motor will be assigned to an axis in the same coordinate system as the
inner-loop motor, and commanded in the same coordinate system. Axis-naming conventions and
standards (e.g. RS/EIA-267) consider these as secondary axes and suggest the name of U when matched
with an X axis, V when matched with Y, and W when matched with Z.
166
Motor Compensation Tables and Constants
Turbo PMAC User Manual
Setup Example
In this example, Motors 1, 2, and 3 are the X, Y, and Z-axes, respectively, in Coordinate System 1 of a
Cartesian stage. Each uses quadrature feedback with 0.1-micron resolution, and is programmed in
millimeters. Motor 4 is used to control the gap height of the vertical tool over the surface. It uses a capacitive
gap sensor through an Acc-28 16-bit A/D converter, with the LSB of the ADC measuring 0.25 microns. It is
assigned to the W-axis in the same coordinate system, also programmed in millimeters (of gap).
I8003=$1F8E00
; ECT 4th entry: unsigned A/D from Y:$78E00
Gap virtual motor basic setup
I400=1
I401=2
I402=$10F0
[email protected]
[email protected]
;
;
;
;
;
Activate Motor 4
Output to an X register
Output to open register X:$10F0
Use result from ECT 4th entry for position
Use result from ECT 4th entry for velocity
Vertical positioning motor link to virtual motor
I305=$10F0
I307=1
; Use #4 output as master input
; Set lowest possible master scale factor
Coordinate system setup
&1
#1->10000X
#2->10000Y
#3->10000Z
#4->4000W
;
;
;
;
X-axis positioning in millimeters
Y-axis positioning in millimeters
Z-axis positioning in millimeters
Vertical gap in millimeters
Changing the Mode of Control
Whenever the following mode control bit is changed, re-align the relationship between the corresponding
motor and axis positions for the inner loop. While this is done automatically any time a program is
started with an R or S command, if this bit is changed in the middle of a motion program, command this
re-alignment explicitly by issuing a PMATCH command. This is an on-line command; from a motion
program, it must be issued with the CMD"PMATCH" syntax.
This command generally must be bracketed before and after with DWELL commands; the first to stop any
lookahead and blending (this can be a DWELL 0), and the second to give the on-line command time to
execute in background from the command queue before the next programmed move is calculated.
Mode Changing Example
The following code segment shows how the transition to engaging the outer loop can be accomplished.
This example continues the one above, in which the Z-axis has been assigned to the inner loop’s motor, so
it is a position axis, and the W-axis has been assigned to the outer loop’s virtual motor.
Z10
DWELL0
I306=3
CMD"&1PMATCH"
DWELL10
W5
;
;
;
;
;
;
Pure position move on inner loop
Stop lookahead
Engage following, put in offset mode
Re-align motor and axis position
Give PMATCH command time to execute
Outer loop command
The following code segment shows how the transition to disengaging the outer loop can be accomplished.
W5
DWELL0
I306=0
CMD"&1PMATCH"
DWELL10
Z0
Motor Compensation Tables and Constants
;
;
;
;
;
;
Outer loop command
Stop lookahead
Disengage following, put in normal mode
Re-align motor and axis position
Give PMATCH command time to execute
Pure position move on inner loop
167
Turbo PMAC User Manual
Special Instructions for Extended Single-Loop Setup
When using cascaded servo loops for extended filtering of a single control quantity, several details of the
setup will be different. The outer loop will be set up first, using the real feedback device, and writing
directly to the registers that command the amplifier, commutating if necessary. This servo loop should be
tuned as well as possible by itself.
When ready to engage the inner loop as well, do the following:
1. Set Ixx01 for the outer-loop motor to 2 to point the output to an X register. This also disables
commutation for the motor if you had been using it before.
2. Set Ixx02 for the outer-loop motor to the address of an open register, usually $0010Fx.
3. Set Ixx00 for the inner-loop motor to 1 to activate it.
4. Set Ixx01 and Ixx02 for the inner-loop motor to the same values you had for the outer-loop motor
when running it alone. If commutating with Turbo PMAC, also set Ixx70 – Ixx84 for the inner-loop
motor to the same values you had for the outer-loop motor.
5. Set Ixx24 and Ixx25 flag parameters for the inner-loop motor to the same values you had for the
outer-loop motor when running it alone.
6. Set Ixx03 and Ixx04 feedback address parameters for the inner-loop motor to the address of a register
that will always be set to 0. Register $35C0 at the end of the conversion table is suggested.
7. Set Ixx05 for the inner-loop motor to the same address as the outer-loop motor’s Ixx02 to pick up the
intermediate value.
8. Set Ixx06 for the inner-loop motor to 1 to enable the use of the Ixx05 register.
9. Set the Ixx07 master position scale factor for the inner loop motor to 1.
10. Set gain term Ixx30 for the inner-loop motor to 65,536, and Ixx31 – Ixx39 to 0. This should make the
inner loop a pass-through and your performance should be the same as with the outer loop alone.
11. Now adjust the gains of the inner-loop motor to get the additional filtering you desire.
Cascaded Loops for Single Quantity Control
First PMAC Motor
Second PMAC Motor
Commanded
Trajectory
+
Servo
Filter
Ixx05
Ixx02
Ixx06
-
+
Actual
Position
Servo
Filter
+
(Always Zero)
Ixx02
Output
Device
Amp
M
Ixx03,
Ixx04
(Always Zero)
Ixx03,
Ixx04
168
ECT
Entry
ECT
Source
Address
E
Motor Compensation Tables and Constants
Turbo PMAC User Manual
User-Written Servo Algorithms
Turbo PMAC supports the installation and automatic execution of user-written servo algorithms. These
can be used if the standard PID and ESA filters are not suitable to get the required servo performance.
Alternately, they can be used for non-servo purposes, with the algorithm guaranteed to execute at the
servo update rate. This can be valuable for fast updates of I/O.
There are two methods for creating these user-written servo algorithms. The first is a compiled method
called Open Servo, using Turbo PMAC’s high-level language, compiled by the PEWIN32 PRO Executive
program. The second is an assembled method using a Motorola cross-assembler for the DSP56300
family. The compiled method is easier to use and more flexible, permitting the use of floating-point
mathematics. The assembled method generates code that is far more efficient and more compact. Both
methods are discussed below.
Only a single user-written servo algorithm may be installed in a Turbo PMAC. This algorithm can be
executed by any motor on the Turbo PMAC. If you desire that different motors execute different userwritten servo algorithms, this must be accomplished by branching within a single algorithm.
Open Servo Compiled Algorithms
Turbo PMAC’s Open Servo feature permits writing a custom algorithm in a high-level language that will
execute on Turbo PMAC’s high-priority servo interrupt. This algorithm can be used either for actual
servo control functions, or for other tasks that must execute at a very high priority, such as very highfrequency I/O, special pre-processing of feedback data, or special post-processing of servo commands.
Open Servo algorithms are compiled into DSP machine code in the host computer before being
downloaded into Turbo PMAC’s active memory. They may be retained in Turbo PMAC’s non-volatile
flash memory using the SAVE command. When executed, they replace only the standard servo-loop
algorithm for the motor. All other tasks, including trajectory generation, motion and PLC program
execution, and safety checking, are still executed by the Turbo PMAC’s built-in firmware.
The Open Servo feature is a second method for creating user-written servo algorithms in Turbo PMAC.
Previously, this could be done only by writing the algorithm in assembly language for the DSP56300
family using Motorola’s cross-assembler, and downloading the assembled code to the Turbo PMAC. The
Open Servo feature permits these algorithms to be written without the need to understand and use
assembly language.
Turbo PMAC can hold and execute only a single Open Servo algorithm. This algorithm can be run by
multiple motors. The sections below explain tools for this single algorithm to access motor-specific
registers; if different procedures are desired for different motors, this must be handled by explicit logic in
the algorithm.
The compiled Open Servo program is similar to the compiled PLC programs, but there are two key
differences:
• Open Servo algorithms run on the servo interrupt, with guaranteed execution every cycle (or the
Turbo PMAC will watchdog); compiled PLC programs either run on the real-time interrupt (PLCC 0)
with possible pre-emption by motion program calculations, or in background (PLCC 1 – 31) with no
deterministic execution rate.
• Open Servo algorithms have specific access mechanisms to special registers used for servo functions.
Requirements
The Open Servo requires a Turbo PMAC controller (Turbo PMAC, Turbo PMAC2, UMAC, or QMAC)
with version 1.938 or newer firmware to execute the algorithm. It requires a PC running PEWIN32PRO
version 3.2 or newer PMAC Executive program.
Motor Compensation Tables and Constants
169
Turbo PMAC User Manual
Computational Features
The Open Servo provides powerful computational features to permit easy writing of sophisticated and
flexible algorithms.
Access to Turbo PMAC Variables
Open Servo algorithms can utilize all of Turbo PMAC’s I, P, Q, and M-variables, reading and writing to
them as appropriate. As in other user programs, it uses floating-point arithmetic to process these variable
values, even those that are stored as fixed-point values (see Floating-Point vs. Fixed-Point Mathematics,
below). Q-variables are accessed from Open Servo algorithms according to the Coordinate System 1
addressing scheme, no matter which coordinate system the motor executing the Open Servo algorithm is
assigned to.
Compiler-Assigned Pointer Variables
For direct and efficient access to Turbo PMAC registers, Open Servo algorithms support two types of
pointer variables for which the register assignment is made at compilation time, not at program execution
time.
L-variables are pointers to short (24-bit) registers, treated as integer (fixed-point) values. These work in
the same way as L-variables do in compiled PLC programs. They can access either X or Y short
registers, either as entire 24-bit registers (treated as signed integers only), or as portions of the registers 1,
4, 8, 12, 16, or 20 bits wide (treated as signed or unsigned integers, except for 1-bit variables, which are
unsigned only).
F-variables are pointers to long (48-bit) registers. If the F-variable definition is an L format (e.g. F1>L:$10F0), the register is accessed as a 48-bit floating-point register. If the F-variable definition is a D
format variable (e.g. F2->D:$88), the register is accessed as a 48-bit signed integer, but conversion to
or from Turbo PMAC’s 48-bit floating-point format is performed automatically, so it can be used in
floating-point mathematics.
Note:
Do not confuse L-variables, which are short-word compiler pointers, with L-format
F-variables and M-variables, which are long-word variables.
Turbo PMAC itself cannot recognize L-variables or F-variables; these variables have meaning only to the
compiler on the host computer.
By contrast, when using Turbo PMAC’s M-variable pointers, the register assignment is made when the
line is executed, each time it is executed. This assignment requires about 600 nanoseconds additional
computation time (on a 100 MHz CPU) each time the variable is accessed. However, this does permit the
M-variable definition to be changed during execution, enabling techniques such as indirect addressing.
It is possible to use L-variables for fast integer arithmetic while retaining the run-time flexibility of Mvariable definitions, but this adds the run-time definition-access computational penalty described above.
Instead of directly defining L-variables to registers for the compiler, you can reference a range of Lvariables to Turbo PMAC M-variable definitions with the LMOVERLAY {start},{end} compiler
directive. This directive must precede the actual Open Servo program. For example, LMOVERLAY
10,20 instructs the compiler that the definitions of L10 through L20 are to be assigned at run time using
the definitions of M10 through M20 respectively at the time each statement is executed, not at
compilation time.
Using the M-variable definition and accessing this definition at run time permits indirect addressing
techniques through real-time modification of this M-variable definition using another pointer variable.
170
Motor Compensation Tables and Constants
Turbo PMAC User Manual
Floating-Point vs. Fixed-Point Mathematics
Each statement in the Open Servo can be executed using either floating-point or integer (fixed-point)
mathematics. In a floating-point statement, all variables used are processed through an intermediate
working format that is 48-bit floating-point, regardless of the storage format of the variable. Floatingpoint statements can utilize any of Turbo PMAC’s I, P, Q, or M-variables, and the compiler’s long Fvariable pointers. They cannot use the compiler’s short fixed-point L-variable pointers. All constants
used in these statements are stored as 48-bit floating-point values.
In an integer statement, all variables used are processed through an intermediate working format that is
24-bit signed integer, regardless of the storage format of the variable. Integer statements can only utilize
the compiler’s short fixed-point L-variable pointers. They cannot use the compiler’s long F-variable
pointers, or Turbo PMAC’s I, P, Q, or M-variables. All constants used in these statements are stored as
24-bit signed integers. All constants used and intermediate values computed must fit in the range of these
integers: -8,388,608 to +8,388,607. No mathematical functions (e.g. SIN, COS) may be used in an
integer statement.
If a constant appears on a line before any variable (e.g. IF (0=P1) …), the line is assumed to be
floating-point. This means that the statement IF (0=L1) is illegal, because it mixes floating-point and
fixed-point data types.
Short-word integer operations will execute more than 10 times faster than the same operations done as
long floating-point operations, but usually will not have the dynamic range to handle the actual control
calculations. They can be useful for efficient calculation of associated logic, however.
Note:
PMAC’s built-in servo algorithms (PID and Extended Servo Algorithm) use long
[48-bit/56-bit] fixed-point mathematics. This format is not supported in the Open
Servo.
If conversion between integer and floating-point data types is required, Open Servo provides the ITOF
(integer-to-float) and FTOI (float-to-integer) functions. The ITOF function (with an integer expression
as its argument) can be used in a floating-point statement, such as:
P20=ITOF(L10+L11)*3.14159/P628
The FTOI function (with a floating-point expression as its argument) can be used in a fixed-point
statement, such as:
L10=L9+FTOI(P5*100)-5
The FTOI function will round the value of the floating-point expression to the nearest integer.
It is not permissible to nest FTOI and ITOF functions within an expression.
Arrays. Open Servo algorithms support two types of arrays: variable arrays and register arrays. Both
provide useful capabilities.
Variable Arrays: Variable arrays work with the Turbo PMAC’s standard PMAC I, P, Q, and Mvariables. The number of the array index is placed inside parentheses, and specifies the variable number
for the specified type of variable. The expression that determines this number is a floating-point
expression, so it can use Turbo PMAC I, P, Q, or M-variables, constants (which will be treated as
floating-point values) and the compiler’s F-variables, but it cannot use the compiler’s L-variables (unless
the value has been converted to floating-point with the ITOF function). The resulting value of this
floating-point expression is rounded to the nearest integer automatically, to select the variable number to
be used. Some examples of statements using these variable arrays are:
P(P1)=P10*32
P30=I(ITOF(L10)*100+30)*P29
Motor Compensation Tables and Constants
171
Turbo PMAC User Manual
Register Arrays: Register arrays work with the compiler’s short L-variables and long F-variables.
These arrays must be declared to the compiler before the start of the actual Open Servo algorithm. In use,
the number of the array index is placed inside square brackets, and specifies the address offset from the
declared beginning of the array. The expression that determines this number is a fixed-point expression,
so it can only use the compiler’s L-variables and constants that fit within the range of a 24-bit signed
integer.
L and F-variable register arrays must be declared to the compiler before the start of the actual Open Servo
algorithm. Examples of these definitions are:
L100->X:$010000[32]
F200->D:$030040[64]
F300->L:$030080[128]
The declared array size must be a power of 2 in the range 2 to 8192. L-variable register arrays always use
full 24-bit X or Y registers, treating the values as signed integers.
Operators
As with any Turbo PMAC user program, Open Servo can utilize the following mathematical and logical
operators:
• + (addition)
• - (subtraction
• x (multiplication)
• / (division)
• % (modulo, remainder)
• & (bit-by-bit AND)
• |
(bit-by-bit OR)
• ^ (bit-by-bit XOR)
All of these operators can be used in either floating-point or integer statements. Integer division rounds
the result to the nearest integer; in the case where the fraction is exactly 0.5, it will round to the next more
positive integer (e.g. -7.5 to -7, and 7.5 to 8).
Comparators
As with any Turbo PMAC user program, Open Servo can utilize the following comparators in conditional
statements:
• = (equal to)
• > (greater than)
• < (less than)
• ~ (approximately equal to [within 0.5])
• != (not equal to)
• !> (not greater than, less than or equal to)
• !< (not less than, greater than or equal to)
• !~ (not approximately equal to [not within 0.5])
The ~ and !~ comparators can only be used in floating-point statements. Note that the <>, >=, and <=
comparators, which can be used in some programming languages, cannot be used in the Open Servo or
other Turbo PMAC programs.
Functions
As with any Turbo PMAC user program, Open Servo can utilize the following mathematical functions.
Note that these functions can only be used in floating-point statements within the Open Servo:
• SIN
(trigonometric sine)
• COS
(trigonometric cosine)
• TAN
(trigonometric tangent)
172
Motor Compensation Tables and Constants
Turbo PMAC User Manual
•
•
•
•
•
•
•
•
•
ASIN
(trigonometric arc sine
ACOS
(trigonometric arc cosine)
ATAN
(trigonometric arc tangent)
ATAN2
(special 2-argument, 4-quadrant arc tangent)*
ABS
(absolute value)
INT
(greatest integer within)
EXP
(exponentiation)
LN
(natural logarithm)
SQRT
(square root)
The trigonometric functions use degrees if Turbo PMAC variable I15 is set to the default value of 0; they
use radians if I15 is set to 1. The present value of I15 is evaluated each time a trigonometric function is
executed.
Note:
The ATAN2 function uses Q0 as its second argument – the cosine argument. The
first argument – the sine argument – is inside the parentheses immediately
following ATAN2. The Q0 used in Open Servo is always Coordinate System 1’s
Q0, no matter which coordinate system the executing motor has been assigned to.
Special Saturation-Check Function
The Open Servo has a special function to check a signed quantity against an unsigned limit magnitude
and saturate the quantity at the magnitude of the limit. The function FLIMIT({value},{limit})
compares the signed quantity {value} against the positive value {limit} and the negative value –
{limit}. If {value} is greater than {limit}, the function returns the value of {limit}; if
{value} is less than –{limit}, the function returns the value of -{limit}. Otherwise, the function
simply returns the value of {value}. Both quantities – {value} and {limit} – must be floatingpoint expressions (the expressions can be simply variables or constants).
For example, the statement:
P2=FLIMIT(P1,20000)
is equivalent to:
IF (P1>20000)
P2=20000
ELSE
IF (P1<-20000)
P2=-20000
ELSE
P2=P1
ENDIF
ENDIF
The FLIMIT function reduces the size of both the source code and the resulting compiled code.
Special Access to Motor Values and Registers
The Open Servo algorithms have several special statements to support useful features for servo loop
execution.
MTRNUM Function: The MTRNUM function returns the number of the motor (1 – 32) for which the
algorithm is presently executing. The number is returned as a fixed-point (integer) value. For example,
the statement L10=MTRNUM*100 would set fixed-point variable L10 to 200 when the Open Servo
algorithm is executing for Motor 2, or to 1800 when executing for Motor 18. Similarly, the statement
P10=ITOF(MTRNUM*100+30) would set floating-point variable P10 to 230 when the Open Servo
algorithm is executing for Motor 2, or to 1830 when executing for Motor 18.
Motor Compensation Tables and Constants
173
Turbo PMAC User Manual
COPYREG Command: The COPYREG command copies five key registers for the executing motor into
five consecutive P-variables, where they can easily be used for calculations. The user does not have to
know the addresses of these registers. In doing this copying, Turbo PMAC automatically converts the
data to 48-bit floating-point format.
The syntax of this command is COPYREG {P-variable name}, where {P-variable name}
specifies the number of the first variable into which data will be copied. The five registers to be copied
by this command are:
• Actual Velocity (1/[Ixx09*32] counts / [Ixx60+1] servo cycles)
• Desired Velocity (1/[Ixx08*32] counts / [Ixx60+1] servo cycles)
• Following Error (1/[Ixx08*32] counts)
• Actual Position (1/[Ixx08*32] counts)
• Desired Position (1/[Ixx08*32] counts)
The actual position value is derived from the register selected by Ixx03 for the motor (Position-Loop
Feedback Address), with the source value multiplied by the Ixx08 scale factor and extended into a 48-bit
long word. The actual velocity value is derived from the position value selected by Ixx04 for the motor
(Velocity-Loop Feedback Address), taking this cycle’s actual velocity-loop position value minus the
value at the previous loop closure and multiplying the difference by the Ixx09 scale factor. Note that this
scale factor is not necessarily the same as for the desired velocity.
For the desired position value, Turbo PMAC adds the trajectory commanded position and the master
position (from the position following, or electronic gearing function), then subtracts the compensation
position (from the position, or leadscrew compensation tables), creating a net desired position value. The
desired velocity value is simply this cycle’s desired position value minus the value at the previous loop
closure.
The following error value is the desired position the actual position. The subtraction is done using 48-bit
fixed-point values; then the difference is converted to floating-point format. There are several advantages
to using the following error value directly. First, it saves some computational time. Second, when the
commanded and actual positions get very large, it preserves fractional position data better.
If the command COPYREG P5 were used, the Actual-Velocity value would be copied into P5, Desired
Velocity into P6, Following Error into P7, Actual Position into P8, and Desired Position into P9. Note
the differing units between the actual and desired velocity registers. (The desired velocity value is not
typically used in actual servo loop closure. Turbo PMAC uses this register in the numerical integration
process to compute the desired position value each servo cycle.)
Offsets from Registers of Executing Motor: The compiler’s L-variables and F-variables can be
declared by address offset to specific registers of the executing motor. In this way, they automatically
index properly from motor to motor, permitting the same variables and code to be used for multiple
motors. These variables can be declared by offset to the motor’s R0 register, which is the motor’s
command output register ($BF for Motor 1), or by offset to the motor’s R1 register, which is the motor’s
status register ($B0 for Motor 1). L-variables can be declared to 24-bit X or Y registers this way; Fvariables can be declared to 48-bit fixed-point or floating-point registers this way. Some examples:
L220->X:(R1-$27)
L270->Y:(R1+0)
F392->D:(R1-$24)
F34->L:(R0+11)
;
;
;
;
Ixx08
Motor
Motor
Ixx16
scale factor register
status register
master position register
maximum commanded speed
The offset must be in the range –64 <= {offset} <= 63 (-$40 <= {offset} <= $3F).
Returned Value: The RETURN command takes the integer value inside the following parentheses and
places it in a 24-bit signed integer register where Turbo PMAC’s standard firmware will take it and use it as
the servo command. Typically, the commanded value will be computed as a floating-point value, so must
be converted to an integer with the ITOF function. Typical uses of the RETURN command could be:
174
Motor Compensation Tables and Constants
Turbo PMAC User Manual
RETURN(FTOI(P345))
L10=FTOI(P92/65536)
RETURN(L10)
The RETURN command will typically be the last line of an Open Servo algorithm. Putting it earlier in the
algorithm will not cause the command data to be used any sooner by the Turbo PMAC. If the Open
Servo program is used for a task other than servo-loop closure, there is no need to use the RETURN
command. In this case, when the Open Servo algorithm reaches the CLOSE statement that is required at
the end of the program, it will write a 0 to this holding register automatically.
Turbo PMAC will take the resulting value and add the contents of the torque compensation register
(usually from the motor’s TCOMP torque compensation table) to it. If Turbo PMAC is not performing
commutation for this motor (Ixx01 bit 0 = 0), it will take this sum and copy it to the register specified by
Ixx02. If you do not use the RETURN command in this case, Turbo PMAC will still copy the zero value
that it has placed in the holding register that would have been used by the RETURN command into the
register specified by Ixx02.
If Turbo PMAC is performing commutation for this motor (Ixx01 bit 0 = 1), it will use the resulting sum
as the “quadrature current” (torque) command input to the commutation algorithm. In this case, Ixx02
specifies the multiple output registers from the commutation algorithm.
The Ixx29 and Ixx79 offset terms are added automatically by Turbo PMAC, just as if the built-in servo
algorithms were used.
The returned value must be an integer value in the range –8,388,608 to +8,388,607. Most of the
command output ranges associated with Turbo PMAC’s automatic servo loops are expressed as 16-bit
values, with a range of –32,768 to +32,767. The values associated with RETURN are therefore 256 times
larger. The actual command output device will not necessarily have this full 24-bit resolution (and
probably will not). In general, however, an n-bit output device uses the high “n” bits of the 24-bit
returned value.
Variable Value Assignments
Mathematical operations in an Open Servo algorithm are performed with variable value assignment
statements, just as in other PMAC programs. The syntactical rules for these statements are the same as in
other PMAC interpreted and compiled programs. Any I, P, Q, M, L, or F-variable can be assigned a
value, whether referenced directly or as part of any array.
Logical Control
Logical branching and looping control in Open Servo algorithms is performed with IF / [ELSE] / ENDIF
branching constructs, and WHILE / ENDWHILE looping constructs, just as in other PMAC programs. The
syntactical rules for these statements are the same as in PMAC PLC programs; they do not support a few
features possible in motion programs (such as an action on the same line as a condition), and they do
support a few features not possible in motion programs (such as multiple-line conditions). Refer to the
Program Command section of the Software Reference manual for details (see IF, ELSE, ENDIF,
WHILE, ENDWHILE, AND, OR).
If WHILE / ENDWHILE loops are used in an Open Servo, it is the user’s responsibility to make sure that
the algorithm never gets stuck in a loop so long that other tasks are compromised. Turbo PMAC will not
release automatically from a loop in an Open Servo for any other task of equal or lower priority. Failure
to release from a loop in a timely fashion can result in servo error (failure to complete one cycle’s servointerrupt tasks by the next servo interrupt), run-time error (failure to compute commanded move
equations in time for that move to start, causing the motion program to abort), or watchdog timer error
(failure to cycle through all required tasks in a timely fashion, causing the Turbo PMAC to shut down
completely).
Motor Compensation Tables and Constants
175
Turbo PMAC User Manual
Processor Utilization
Servo algorithms are one of the most important tasks executed by the Turbo PMAC’s processor, but far
from the only one. While Turbo PMAC’s DSP processor is very efficient, it is still possible to overload
the processor, particularly with floating-point algorithms executing in compiled code from the Open
Servo. These do not run nearly as efficiently as the standard servo algorithms, which have been written in
assembly language and use fixed-point mathematics.
Generally, it is recommended that the portion of processor time devoted to phase and servo tasks not
exceed 50%, in order to allow sufficient time for lower-priority tasks, such as motion program and PLC
program calculations. To find out how much processor time an Open Servo algorithm occupies, refer to
the Evaluating the Turbo PMAC’s Computational Load section. This can be found in the Turbo PMAC
Computational Features section of this manual.
Memory Utilization
The DSP563xx CPU for the Turbo PMAC employs Harvard architecture, with separate areas of program,
or instruction, memory, and data memory. The actual instructions of the Open Servo are loaded into P
program memory; all of the data registers it uses are in X and Y data memory.
If an Option 5Cx 80 MHz CPU configuration is ordered, employing the DSP56303, only 3K words of
program memory are available for the Open Servo (P:$040000 through P:$040BFF). This may not be
enough for large algorithms, so it is recommended for the user to order a 100 MHz Option 5Dx or 160
MHz 5Ex CPU configuration, utilizing the DSP56309 and DSP56311 CPUs, respectively, for any
complex algorithms. With these processors, 19K words of program memory are available for Open Servo
code (P:$040000 through P:$044BFF). For reference, the Open Servo example below that mimics the
basic PID algorithm occupies 330 words of program memory.
The downloader will indicate how many words of program memory the program has just compiled
occupies. This can be found also by reading memory location P:$040014 (use an RHP:$040014 on-line
command), which contains the location of the end of the Open Servo algorithm in the Turbo PMAC
program memory; subtract the starting location $040000 (remember that these are hexadecimal values) to
get the program length.
For general data memory, most users will utilize some of Turbo PMAC’s 8192 P-variables to store
values. It is the user’s responsibility to keep track of which P-variables are used for Open Servo
algorithms and which are employed for other user tasks. Q-variables may also be used for Open Servo
algorithms; they are always accessed according to Coordinate System 1’s numbering. It is the user’s
responsibility to make sure that these P or Q-variables are not used for other tasks as well.
Many users will find it advantageous to utilize motor-specific registers that are not otherwise being used
because the Open Servo is executing for that motor. Registers listed as being gains or intermediate values
for either the PID or the ESA servo algorithms may safely be used by the Open Servo. The registers in
the range $000092 – $0000AC (for Motor 1; equivalent registers for other motors) with the exception of
$0000A5 (previous net desired position) may be used for this. Note that the integrated position error
register at $00009E and the previous net desired velocity register at $00009A are automatically zeroed
when the loop is opened. It is not necessary to use these registers in the same format as the automatic
servo algorithms would.
Note that any of these registers representing an I-variable for either the PID or ESA algorithms would be
overwritten by any command writing to that I-variable; that the values in such a register are copied into
flash memory on a SAVE command; that the last saved value from such a register is copied from flash
memory on a board reset. Some users may want to utilize these I-variable registers as gains for their own
servo algorithms – it is not necessary to use them for the same purpose, or with the same scaling, as the
built-in algorithms would.
176
Motor Compensation Tables and Constants
Turbo PMAC User Manual
For large amounts of extra data memory, it is recommended to use the User Buffer set up with the on-line
DEFINE UBUFFER command. The User Buffer occupies a number of registers at the high end of X/Y
data memory. With an Option 5x0 standard memory configuration, the end of data memory is at
X/Y:$0107FF; if DEFINE UBUFFER 2048 is declared, all data memory from $010000 through
$0107FF is reserved for the user’s own purposes. With an Option 5x3 extended memory configuration,
the end of data memory is at X/Y:$03FFFF; there is by default a User Buffer of 65,536 words, reserving
all memory registers from X/Y:$030000 to X/Y:$03FFFF for user use. It is the user’s responsibility to
make sure that registers in the UBUFFER utilized for Open Servo data storage are not used for other
purposes as well.
Writing the Open Servo Program
The Open Servo program should be written in a plain-text editor such as the editor in the new
PEWIN32PRO PMAC Executive program. While the program can be written in any plain-text editor, it
must be compiled by the PEWIN32PRO editor’s download function. In this program, released in October
2001, the download routine will recognize Open Servo routines automatically, compile them, and
download the resulting machine code. (Older versions of the PMAC Executive program are not capable
of doing this.)
In the file containing the Open Servo preceding the actual program must be all L-variable and F-variable
pointer definitions, and all #define macro substitutions, or #include references to accessible files
that contain these definitions and substitutions. Remember that the built-in compiler does not download
these definitions and substitutions to the Turbo PMAC; it uses them to do the compilation properly.
The OPEN SERVO command is a signal to the compiler that the statements following up to the (required)
CLOSE command are to be compiled into DSP machine code before downloading. The CLEAR
command that is used following the OPEN command on interpreted buffers is not required for Open Servo
algorithms, because downloading the newly compiled code automatically clears older code, but it may
still be used here.
Example 1: Proportional Control
The following algorithm shows one of the simplest possible Open Servo algorithms, implementing a
simple proportional control law using the motor’s Ixx30 parameter as the proportional gain.
OPEN SERVO
; Following lines to be compiled
CLEAR
; Not necessary, but acceptable
COPYREG P30
; Copy following error into P32
P35=P32*I(ITOF(MTRNUM*100+30))/65536
; Multiply by gain, scale
RETURN(FTOI(P35))
; Make an integer and output
CLOSE
Example 2: Bi-Quad Filter
The next example shows an implementation of a bi-quad filter capable of running on multiple motors,
storing values from cycle to cycle for each motor. It uses #define substitution macros to keep the code
readable, and the MTRNUM function for variable and register arrays to separate stored values for each
motor. Variable arrays (which are easier for the user to access) are used for user-set “gains”, and register
arrays (which are quicker for the algorithm to access) are used for algorithm-calculated stored values.
This Open Servo program implements the following transfer function:
U( z ) = K p
( z + a )( z + c )
E( z )
( z + b )( z + d )
It implements this as the following difference equation:
u k = K p [ek + (a + c )ek −1 + acek −2 ] − (b + d )u k −1 − bdu k −2
Motor Compensation Tables and Constants
177
Turbo PMAC User Manual
In order to start the algorithm correctly, it reads the servo cycle counter and compares it to the counter the
last time the last time this algorithm was executed. If the algorithm was not executed the previous cycle,
it zeros out the “history” values for the algorithm. It also does a saturation check on the commanded
output. This algorithm assumes a standard memory option (5x0) whose data memory ends at
X/Y:$0107FF and a UBUFFER defined of at least 2048 words.
Definitions and Substitutions
#define
Kp
P(ITOF(MTRNUM*100+30))
; Gain term is Pxx30
#define
A
P(ITOF(MTRNUM*100+31))
; A zero is Pxx31
#define
B
P(ITOF(MTRNUM*100+32))
; B zero is Pxx32
#define
C
P(ITOF(MTRNUM*100+33))
; C pole is Pxx33
#define
D
P(ITOF(MTRNUM*100+34))
; D pole is Pxx34
#define
E
P42
; Error term e(k) (not saved)
#define
Temp1
P45
; Temporary value
#define
Temp2
P46
; Temporary value
#define
Temp3
P47
; Temporary value
#define
U
P48
; Output term u(k) (not saved)
#define
LastE
F1[MTRNUM-1]
; e(k-1) is F1[#-1]
F1->L:$010000[32]
; Float reg array in UBUFFER
#define
PrevE
F2[MTRNUM-1]
; e(k-2) is F2[#-1]
F2->L:$010020[32]
; Float reg array in UBUFFER
#define
LastU
F3[MTRNUM-1]
; u(k-1) is F3[#-1]
F3->L:$010040[32]
; Float reg array in UBUFFER
#define
PrevU
F4[MTRNUM-1]
; u(k-2) is F4[#-1]
F4->L:$010060[32]
; Float reg array in UBUFFER
#define
ServoCycle L0
L0->X:$0,0,24,S
; Servo cycle counter
#define
LastServoCycle
L1[MTRNUM-1]
L1->X:$010080[32]
; Register array in UBUFFER
#define
ServoExtension L2
L2->Y:(R1-$21)
; Ixx60 register as integer
#define
OutputLimit 8388607
Start of Actual Algorithm
OPEN SERVO CLEAR
COPYREG P40
; Following error into P42
; If loop was not closed last cycle, zero out stored values
IF (ServoCycle-LastServoCycle!=ServoExtension+1)
LastU=0
PrevU=0
LastE=0
PrevE=0
ENDIF
LastServoCycle=ServoCycle
; Store for next cycle
Temp1=Kp*(E+(A+C)*LastE+A*C*PrevE) ; Compute TF numerator
Temp2=(B+D)*LastU+B*D*PrevU
; Compute TF denominator
Temp3=Temp1-Temp2
; Combine for net command
U=FLIMIT(Temp3,OutputLimit)
; Saturation check
PrevE=LastE
; Store values for next cycle
LastE=E
PrevU=LastU
LastU=U
RETURN(FTOI(U))
; Return command value as integer
CLOSE
178
Motor Compensation Tables and Constants
Turbo PMAC User Manual
Example 3: PMAC’s PID Filter
The following example mimics the action of Turbo PMAC’s basic PID loop (without the notch, deadband
compensation, position-error limiting, or friction feedforward terms), but with floating-point calculations.
It uses the PID’s own I-variables, accessed as L-variables for speed, then converted to floating-point
values.
This program is a user written servo that replicates the PMAC PID loop. It uses the following equations
defined in the PMAC User Manual but modified for PMAC passed terms:
K16 = 2**-16
K128 = 1/128
K23 = 2**-23
FE = DPOS - APOS - From PMAC in Ix08 * 32 counts
AVEL, DVEL - From PMAC in Ix09 * 32 counts/servo period
IPOS = ITOF(Ix33) * FE * K23 + IPOS
DACEL = DVEL - PDVEL
PDVEL = DVEL
DACOUT = ITOF(Ix30)*K16 * ( FE + K128 *(ITOF(Ix32) * DVEL +
ITOF(Ix35) * DACEL - ITOF(Ix31) *AVEL) + IPOS)
#define K16 0.000015259
#define K128 0.0078125
#define K23 0.000000119
; Constant of 2^-16
; Constant of 1/128
; Constant of 2^-23
Use L-variables for Quick Access to PID I-variable Registers.
#define Ix08 L1
L1->Y:(R1-$27)
#define Ix09 L2
L2->Y:(R1-$14)
#define Ix30 L3
L3->Y:(R1-$17)
#define Ix31 L4
L4->X:(R1-$1E)
#define Ix32 L5
L5->X:(R1-$21)
#define Ix33 L6
L6->X:(R1-$11)
#define Ix35 L7
L7->Y:(R1-$1D)
#define Ix63 L8
L8->Y:(R1-$11)
#define Ix69 L9
; Note that this is 24-bit value, not 16
L9->Y:(R1-2)
#define STATUS L10
L10->X:(R1+0)
#define IPOS F1
; Integrated position error register
F1->L:(R1-$12)
; Automatically zeroed on motor open loop
#define PDVEL F2
; Previous desired velocity register
F2->L:(R1-$29)
; Automatically zeroed on motor open loop
#define AVEL P0
; Floating-point actual velocity
#define DVEL P1
; Floating-point net desired velocity
#define FE P2
; Floating-point following error
#define APOS P3
; Floating-point actual position
#define DPOS P4
; Floating-point net desired position
#define DACOUT P5
; Floating-point commanded output
OPEN SERVO CLEAR
COPYREG P0
; Copy Motor AVEL,DVEL,FE,APOS,DPOS to float P0..4
Motor Compensation Tables and Constants
179
Turbo PMAC User Manual
; FE, APOS, & DPOS in 1 / [Ix08* 32] counts
; AVEL in 1 / [Ix09*32] counts/servo period
; DVEL in 1 / [Ix08*32] counts/servo period;
; if( Ix34 = 1 && Des_Vel0 = 1 OR Ix34 = 0)
; then integrate IPOS = IPOS + FE * Ix33 && Limit to Ix63
If (STATUS&$12000 = $12000 Or STATUS&$10000 = 0) ; Test Ix34 mode
IPOS = FLIMIT(ITOF(Ix33)*FE*K23+IPOS, ITOF(Ix63)*ITOF(Ix08)*2)
; Scale Ix63 to include Ix08
EndIf
DACOUT = FLIMIT(ITOF(Ix30)*K16 * ( FE + K128 *(ITOF(Ix32) * DVEL + ITOF(Ix35)
* (DVEL - PDVEL) - ITOF(Ix31) *AVEL) + IPOS), ITOF(Ix69))
PDVEL =DVEL
; Store for next cycle
RETURN(FTOI(DACOUT))
CLOSE
Downloading the Program
Each time the Executive program’s downloader sends the compiled code for an Open Servo algorithm to
the Turbo PMAC, it first completely erases the existing user-written servo code in the Turbo PMAC.
Therefore, always re-compile the entire Open Servo, even for the slightest change. The CLEAR command
used at the start of these examples is not therefore needed, but it can be included for consistency of style
with interpreted programs.
Remember that the download process only puts the Open Servo algorithm into the volatile active RAM
memory. To keep this algorithm loaded in the Turbo PMAC through a reset or power-cycling, copy it to
non-volatile flash memory with the SAVE command first.
Executing the Open Servo Program
For the Open Servo algorithm to execute for a given motor, Ixx00 for the motor must be equal to 1 so that
motor calculations are activated. Bit 0 of Ixx59 must be set to 1 (Ixx59 = 1 or 3) so it will choose the
user-written servo algorithm generated from the Open Servo instead of a built-in servo algorithm (PID or
ESA). Finally, the servo loop must be closed for this motor. The Open Servo algorithm will not executed
if the motor is either in the open-loop enabled state, or the killed (open-loop disabled) state.
The Open Servo algorithm obeys the Ixx60 servo-cycle extension parameter for each Motor xx. As with
the built-in PID and ESA servo algorithms, it executes every [Ixx60+1] servo interrupts. With the default
Ixx60 value of 0, it executes every servo interrupt.
Those users who are employing the Open Servo algorithm for tasks other than actually closing servo
loops must be careful not to disable the algorithm unintentionally. The algorithm will not be running
immediately on power-up/reset unless bit 0 of Ixx80 for the motor is set to 1. A <CONTROL-K> (kill all)
command disables the servo loops of all motors, including a pseudo-motor running an Open Servo
algorithm. With the default settings of bits 21 and 22 of Ixx24 an amplifier fault or fatal following-error
fault on any motor disables the servo loops of all motors, including a pseudo-motor running an Open
Servo algorithm.
The Open Servo replaces only the actual servo-loop closure algorithm for the selected motors. Other
tasks done as part of the servo interrupt, including encoder-conversion-table processing, commanded
trajectory generation, position following, and time-base control, are executed by the built-in firmware,
whether or not the Open Servo is selected for any motor. Also, related tasks done outside of the servo
interrupt, such as checking against a fatal following error limit, are executed by the built-in firmware.
180
Motor Compensation Tables and Constants
Turbo PMAC User Manual
Assembled User-Written Servo Algorithms
Highly efficient user-written servo algorithms may be written in the assembly language for the DSP56300
family of processors used in the Turbo PMAC. This requires the use of a cross-assembler from Motorola,
obtainable at no cost from their website. It also requires a linking program from Delta Tau, called
CODET.EXE and running under Microsoft Windows operating systems, available at no cost from the
Delta Tau website.
Writing the Algorithm
The algorithm is written in Motorola DSP56300 assembly language using any standard text editor. The
code written is subject to the following restrictions.
Program Memory Space
The program must start at memory location P:$040000, so the first line of code must be:
ORG P:$40000
For a DSP56303 processor (80 MHz CPU Option 5Cx), the resulting code must end by memory location
P:$040BFF, providing a 3-Kword buffer for the program. For a DSP56309 processor (100 MHz CPU
Option 5Dx) or a DSP56311 processor (160 MHz CPU Option 5Ex), the resulting code must end by
memory location P:$044BFF, providing a 19-Kword buffer for the program. However, in all cases, if a
user-written phase program is used, that program starts at memory location P:$040800, limiting the userwritten servo program to a 2-Kword buffer.
Conditions on Entry
On entry into the user-written servo, the program can expect the following data for the executing motor in
internal DSP registers:
• The A10 register contains a 48-bit integer representing the desired velocity in units of 1/(Ixx08*32)
counts per servo cycle
• The B10 register contains a 48-bit integer representing the desired position in units of 1/(Ixx08*32)
counts
• The Y register contains a 48-bit integer representing the actual position in units of 1/(Ixx08*32)
counts
• The R0 register contains the address of the command output and command bias registers for the
executing motor (e.g. $0000BF for Motor 1). This must not be changed.
• The R1 register contains the address of the first status word for the executing motor (e.g. $0000B0 for
Motor 1). This must not be changed.
• The N1 register contains the block length of the motor servo registers ($80 presently), which may be
useful in incrementing from motor to motor. This must not be changed.
Conditions on Exit
On exit from the user-written servo, the Turbo PMAC firmware expects to find the control effort value in
the A register as a 24-bit signed integer. If Turbo PMAC commutation is enabled for this motor (Ixx01
bit 0 = 1), this value is used as the “quadrature current” (torque) input to the commutation algorithm. If
Turbo PMAC commutation is disabled for this motor (Ixx01 bit 0 =0), this value is copied by Turbo
PMAC firmware to the register specified by Ixx02. In this case, not all of the register may be used. DAC
registers for a PMAC-style Servo IC use only the top 16 bits. DAC registers for a PMAC2-style Servo IC
use only the top 18 bits.
The last line of the user-written servo must be:
JMP
<$001
Motor Compensation Tables and Constants
181
Turbo PMAC User Manual
Available Registers
The following data registers may be used by the user-written servo:
• Internal DSP registers R4, N4, R5, and N5 may be used, and do not need to be restored when done.
• Internal DSP registers M0, M4, and M5 may be used, but must be restored to previous values when
done.
• Motor intermediate value registers X:$000x93/13 through X:$000x9A/1A may be used to hold values
from cycle to cycle. They are not used by any Turbo PMAC firmware as long as the user-written
servo is activated.
• Global registers X/Y:$0010F0 – $0010FF may be used. They are not used by any Turbo PMAC
firmware tasks, other than being set to 0 on power-up/reset.
• Registers in the user buffer established by the DEFINE UBUF command may be used. They are not
used by any Turbo PMAC firmware tasks.
• Other registers may be used as well, but it is possible for certain tasks of Turbo PMAC firmware to
overwrite these. For example, it is possible to use the registers for some P or Q-variables for the userwritten servo, but assigning a value to one of these variables will overwrite the register. It is also
possible to use the I-variables for Turbo PMAC’s standard servo algorithms as gains for the userwritten servo.
Programming Restrictions
Do not use any levels of the DSP’s stack, so no DO or JSR instructions are permitted.
Do not use internal DSP address registers R2, R3, R6, and R7; modifier registers M2, M3, M6, and M7;
offset registers N2, N3, N6 and N7.
Assembling the Algorithm
The assembly language algorithm must be assembled into DSP56300 machine code using Motorola’s
cross assembler for the computing platform. Follow the instructions from Motorola to do this.
Linking the Algorithm
Use the Delta Tau applet CODET.EXE, available on the Delta Tau website to convert the file that results
from the Motorola assembler into a format that can be directly downloaded to the Turbo PMAC. This file
should be archived on the computer or network.
Downloading the Algorithm
Use any version of the PMAC Executive program to download this resulting file into Turbo PMAC’s
program memory. Remember that it is downloaded into volatile RAM memory. To have the Turbo PMAC
retain this algorithm, issue a SAVE command before the controller is reset or power removed from it.
Executing the Algorithm
Set bit 0 of Motor xx variable Ixx59 to 1 (Ixx59 = 1 or 3) to select the user-written servo algorithm. As
with the PID or the ESA, the servo loop for the motor must be closed in order for this algorithm to
execute. It will not execute if the motor is either open-loop enabled or killed. If using the user-written
servo algorithm for non-servo tasks, be aware that certain commands (e.g. <CTRL-K>) or fault
conditions on other motors (fatal following error or amplifier fault if their Ixx24 specifies killing other
motors on their fault) can disable your algorithm.
182
Motor Compensation Tables and Constants
Turbo PMAC User Manual
MOTOR COMPENSATION TABLES AND CONSTANTS
Turbo PMAC has the capability to perform sophisticated table-based corrections for both position and
torque on its motors. These permit compensating for imperfections in the system that cannot be measured
with the sensors used in the actual application (although reference sensors that can measure the
imperfections must be used to characterize the errors).
Note:
Deadband compensation, controlled by motor parameters Ixx64 and Ixx65, is part
of the servo feedback algorithm. It is covered in the Setting Up the Servo Loop
section of this manual.
Note:
Cutter-radius compensation is a coordinate-system function, not a motor function.
It is covered in the Writing and Executing Motion Programs section of this manual.
Position Compensation Tables
Turbo PMAC is capable of performing table-based position correction, commonly called leadscrew
compensation. This technique, which also goes by other names, allows for a table of corrections to be
entered into Turbo PMAC as a function of motor position. Turbo PMAC can store up to 32 of these
compensation tables.
These tables are most often used to compensate for imperfections in the mechanics between the position
sensor (often on the back of the motor) and the load whose position is to be controlled. In many systems,
the leadscrew that moves the load linearly as the motor rotates is the largest source of positioning error, so
traditionally these tables are called leadscrew compensation tables. However, these tables can also be
used to compensate for imperfections in the sensor itself.
Source, Target, and Assigned Motors
Each motor can have one table that belongs to it; that is, the DEFINE COMP command that creates the
table assigns it to the presently addressed motor, and each motor can only have one table assigned to it.
Unless otherwise specified, the table uses position information from this motor (source data) to determine
the location in the table, and also adds its correction to this motor (target data). However, the source
motors or both the source and the target motors may be specified to be motors other than the motor to
which the table belongs. (If both motors are different, the concept of the table belonging to a motor is
useful only for Turbo PMAC’s own bookkeeping purposes.)
Standard Leadscrew Compensation
A position compensation table with a single source motor, and the target motor the same as the source
motor, forms a standard leadscrew compensation table. This is the most commonly used type of table, as
the errors in the direction of travel as function of that travel tend to be the largest errors.
Setting Up the Servo Loop
183
Turbo PMAC User Manual
PMAC Compensation Tables
Standard leadscrew compensation
e.g. ∆x = f(x)
y
y
y
Get linear encoder accuracy (almost!) with rotary encoder
Characterize system with linear sensor
Enter errors in PMAC
∆x
E
∆x
∆x
∆x
∆x
M
DEFINE COMP 200, #1, #1, 100000
Table Length
Source Motor
Table Span in Counts
Target Motor
Uses of Cross-Axis Compensation
The ability to have separate source and target motors for a table has several uses. The first is the
traditional compensation for imperfect geometry, as in a bowed leadscrew. For instance, on an XY table,
if the X-axis leadscrew is bowed, the Y-axis position should receive a correction as a function of X-axis
position. If motor #1 is the X-axis, and motor #2 is the Y-axis, the table holding this correction would
have motor #1 as the source motor, and motor #2 as the target motor.
PMAC Compensation Tables
Cross-axis compensation
e.g. ∆y = f(x)
y
y
Useful for bowed leadscrews
Can be used to build electronic cam tables
∆y
∆y
∆y
∆y
∆y
∆y
M
E
DEFINE COMP 100, #1, #2, 100000
Table Length
Source Motor
Table Span in Counts
Target Motor
A second use for cross-axis compensation is what is often known as the electronic cam. In this case, the
entire movement of the target motor is caused by the entries in the compensation table, not just the
corrections. This method of implementing electronic cam operation has two significant advantages over
Turbo PMAC’s time-base following, the other method of creating electronic cams: the compensation
table is bidirectional – the master can turn in either direction – and it is absolute, so the phasing in is
simply a matter of homing the axes.
The time-base method, in which the motion program of the slave motors defines the motion, retains the
advantage of being able to change on the fly through math and logic in the program, and of second or
third order interpolation between points, rather than the compensation table’s 1st-order interpolation.
Refer to the Synchronizing Turbo PMAC to External Events section of this manual for details.
Dimension of the Table
Turbo PMAC presently supports one-dimensional (1D) and two-dimensional (2D) position compensation
tables. Three-dimensional (3D) tables may be supported in the future. 1D tables have a single source
motor; 2D tables have two source motors. Note that a table with a single source motor that is different
from the target motor (a cross-axis compensation table) is still a 1D table.
184
Setting Up the Servo Loop
Turbo PMAC User Manual
PMAC Compensation Tables
2D (Planar) compensation tables
e.g. ∆z = f(x,y)
DEFINE COMP 15.20, #1, #2, #3, 20000, 15000
Table # of rows
Table # of columns
Row motor
Column motor
Column motor span in counts
Row motor span in counts
Target motor
Note:
3D compensation may be achieved in Turbo PMAC through the use of the
kinematic subroutines, which can be used to compute the corrections
algorithmically. The correction equations would be fit to the observed
measurements, probably through a least-squares fit on polynomial equations.
Refer to the section on kinematics algorithms in the Setting up a Coordinate
System section of this manual for details. Note that both the correction equations
(the inverse kinematics) and their inverse (the forward kinematics) must be
entered.
Kinematic equations can also be used for 1D and 2D compensation if algorithm,
rather than table-based, compensation is desired. Corrections for all motors in the
coordinate system must be done in the same kinematic algorithm. By
parameterizing the algorithm coefficients, the corrections can be made dynamically
adjustable (as a function of temperature, for example).
Using Desired vs. Actual Position
The position compensation tables can use either the desired position or the actual position of the source
motors to compute their corrections. In most applications it does not matter which is used, but if the
source and target motors are the same, the gain of the motor is very high, and there are significant
corrections, use of actual position can affect the servo loop performance, effectively changing loop gains
as a function of position. Use of desired position is recommended in these cases. See below for an
explanation of how to specify use of desired position.
Note that in either case, the table is a function of raw (uncorrected) motor position. Since the entire
purpose of the table is to permit command moves to be made to corrected positions, if the target motor is
the same as the source motor, at a certain commanded numerical position value, the correction will not in
general be the same as at the raw position of the same numerical value.
Multiple Tables per Motor
A motor may provide the source data for any of the position compensation tables; it may also be the target
of any of the position compensation tables, with the correction of each table to the target motor’s being
additive. For example, it is possible to have both a repeating fine compensation table for a motor for
cyclic errors such as sensor eccentricity, and a non-repeating coarse table. Also, corrections may be
applied to a motor both as functions of its own position and another motor’s position.
Setting Up the Servo Loop
185
Turbo PMAC User Manual
Table Range
The compensation is defined directly for a range of source motor positions starting at zero counts (the
most recent home or power- up/reset position) and going in the positive direction. The size of this range
is declared as the last argument of the DEFINE COMP command. This argument has units of counts of
the source motor. The spacing between entries is the total range divided by the number of entries (which
is the first argument of the DEFINE COMP command). The first entry in the table defines the correction
at one spacing from the zero position of the source motor, the second entry at two spacings, and so on.
Rollover
Outside of this range, the uncorrected position is rolled over to within this range – essentially a modulo
(remainder) operation – before the compensation is done. For 2D tables, this rollover occurs in both
dimensions. The rollover permits compensation of rotary axes over several revolutions, and simple
compensation for encoder eccentricity. Of course, if the table is made big enough to cover the entire
source motor travel, the rollover feature will never be used.
If the motor has a travel range to the negative side of zero, and compensation is desired here, these entries
should be made as if they were past the positive end of the motor range. For instance, if the motor travel
were +/- 50,000 counts and a table entry was to be made every 500 counts (so 200 entries total), the table
would be set up with a DEFINE COMP 200,100000 command. The first 100 entries would cover the
500 to +50,000 count range, and the last 100 entries would cover the -50,000 to 0 count range. (Usually
the table is referenced so there is a zero correction at the source motor zero position, so the last entry in
the table should be 0.) Essentially, the -50,000 to 0 range would be mapped into the +50,000 to +100,000
range.
If global variable I30 (new in V1.939 firmware) is set to the default value of 0 when the table is
downloaded to Turbo PMAC, the correction value at 0 counts of the source motor is always 0. In this
case, the last entry of the table must be set to 0, or there will be a discontinuity in the correction (and
therefore a position jump) as the source motor passes either end of the table, rolling over the correction.
Most often, the correction is defined to be 0 at the zero position of the motor.
If I30 is set to 1 when the table is downloaded to Turbo PMAC, the correction value at 0 counts of the
source motor is set equal to the last entry of the table, guaranteeing smooth rollover of the table. If the
last entry for the table is 0, the result will be the same regardless of the setting of I30.
Determining Compensation Values
The values that will be entered into the compensation table are determined by comparing the raw
measurements of the sensor that will be used in the application against a reference sensor that is installed
only for the calibration process. Move the axis (or the axes) to the raw position at which you want to
make an entry. Read the reference sensor at this position. The entry in the table will be the raw position
minus the reference position, with both values scaled to 1/16 of a Turbo PMAC software count, both
measured from the motor home (zero) position.
For standard leadscrew compensation tables (1D tables with the same motor as both source and target),
Delta Tau provides a PC software package called Flycal that can perform these measurements on the fly
and automatically generate these tables quickly.
Entering Tables
Position compensation tables are entered into the Turbo PMAC with on-line commands. First there is the
DEFINE COMP command, which defines the size and span of the table, and which motors it uses as its
source and target motors. Following this is a series of numerical constants, separated by spaces and/or
carriage-return characters, which are entered sequentially into the table. (If there is no table to be filled, a
numerical constant sent to PMAC is assigned to variable P0.)
Position compensation tables must be defined in order from those assigned to higher-numbered motors to
those assigned to lower-numbered motors.
186
Setting Up the Servo Loop
Turbo PMAC User Manual
Entering 1D Tables
If the position compensation table to be entered has both the source and target motors equivalent to the
addressed (assigned) motor, and the table uses the actual position of this motor to calculate the
corrections, then there is no need to specify the motors in the DEFINE COMP command. The command
would look something like:
#2 DEFINE COMP 25, 500000
This command establishes a table using Motor 2 as both the source and target motor, using the actual
position to compute the corrections. The next 25 numerical constants sent to Turbo PMAC would be
entered into the table. The first value would be the correction at 500,000/25 = 20,000 counts. (20,000
here represents the uncorrected position.) The second value would be the correction at 40,000
uncorrected counts, and so on. The 25th entry would be the correction at 500,000 counts; if I30 were set
to 1 at the time of entry, this would be the correction at 0 counts as well. The units of the correction itself
are 1/16 count.
If you desire that either the source or target motors be different from the addressed motor, or you wish to
make the table a function of the source motor’s desired position, you must explicitly declare the motors in
the DEFINE COMP command. In this case, the command would look something like:
#1 DEFINE COMP 25, #2D, #2, 500000
This command establishes a table using Motor 2 both as the source, working from its net desired position,
and as the target to which the corrections are applied, even though the table is assigned to Motor 1. As in
the above case, the next 25 numerical constants would be entered into the table, with the first value being
the entry at 20,000 counts the second at 40,000 counts, and so on.
1D Table Example
Below is a simple example of the entry of a 1D table
#1
DEFINE COMP 8,4000
-160
80
120
96
20
-56
-12
0
;
;
;
;
;
;
;
;
;
;
;
Table of 8 entries over 4000 cts belonging to
motor 1 ; Uses motor 1 for source (actual pos)
& target because no other motors specified
Correction at 4000/8 (500) cts is -160/16 = -10 cts
Correction at 1000 counts is 5 counts
Correction at 1500 counts is 7.5 counts
Correction at 2000 counts is 6 counts
Correction at 2500 counts is 1.25 counts
Correction at 3000 counts is -4.5 counts
Correction at 3500 counts is -0.75 cts
Correction at 4000 (and 0) cts is zero
In this example, the correction at a raw position of 1300 counts would be linearly interpolated between
the corrections at 1000 counts and 1500 counts, as follows:
Correction = 5 + ( 7.5−5 )*
1300 − 1000
= +6.5counts
500
Entering 2D Tables
If the position-compensation table has two source motors, establishing a 2D, or planar table, the entry is a
little more complex. Declare the length of the table (in number of entries) in both dimensions, and the
span of the table in both dimensions. Declare both source motors, and usually the target motor (although
the default is the addressed motor). The command that establishes the table will look something like:
#2 DEFINE COMP 15.20, #1D, #2D, #3, 20000, 15000
This command specifies that the table assigned to Motor 2 will have 15 rows and 20 columns. Therefore,
each row has 20 entries, and each column has 15 entries.
Setting Up the Servo Loop
187
Turbo PMAC User Manual
Motor 1 is the first source motor (the row motor), using its desired position; each row represents a row
span of counts of positions of Motor 1, and each column represents a constant raw position of Motor 1.
Motor 2 is the second source motor (the column motor), using its desired position; each column represents a
column span of counts of Motor 2, and each row represents a constant raw position of Motor 2.
Motor 3 is the target motor; the corrections are applied to Motor 3.
The span of each row is 20,000 counts, so there is a spacing of 20,000/20 = 1000 counts (of Motor 1)
between entries along the row dimension. The span of each column is 15,000 counts, so there is a spacing
of 15,000/15 = 1000 counts (of Motor 2) between entries along the column dimension.
The next (15+1)*(20+1)-1=335 numerical constants sent to Turbo PMAC are entered into this table.
Why this number of entries? Because there are entries in each row and column at both the zero position
and the maximum position (hence the n+1 terms), but there is no explicit entry at the origin of both source
motors (hence the final -1).
In this example, the first entry would be the correction at Motor 1 (raw) position 1000, Motor 2 (raw)
position 0 – i.e. at (1000, 0). The second entry would be at (2000, 0). The 20th entry would be at (20,000,
0). The 21st entry would be at (0, 1000), the 22nd at (1000, 1000), and so on. The 335th and last entry of
the table would be the correction at (20,000, 15,000). If I30 were set to 1, this value would also be the
correction at (0, 0). Typically the correction at the origin is made zero by definition, to serve as the
reference point for the other corrections.
If there is any possibility of motion going past the declared span of the table, whether for purposeful
rollover or not, the entries at both ends of each row should be the same; likewise for each column.
Otherwise, there will be a discontinuity in the correction at the edge of the table.
Note three things to be careful about in the entry of a 2D table. First, the number of rows and number of
columns is separated by a period, not a comma. Second, the number of rows (15 in the above example) is
entered first, before the number of columns, but the spacing (in counts) between rows is determined by
the span of a column (15,000 in the above example), which is entered after the span of a row. Finally, to
permit efficient computation in Turbo PMAC, both Row and Column 0 as well as Row and Column n
must be entered.
2D Table Example
The following example shows the entry of a simple 2D table, shown in a form that makes for easy reading
by a user. (Turbo PMAC does not require that the table be entered with each row on a separate line, but
this is recommended for readability.) Note that with an implied correction value of 0 for the zeroth entry,
Rows 0 and 4 are identical, as are Columns 0 and 5.
#3 DEFINE COMP 4.5, #1D, #2D, #3, 50000, 40000
38 45 –22 –35 0
; Row 0, Columns 1-5
24 56 13 –34 –8 24
; Row 1, Columns 0-5
18 43 –9 –65 32 18
; Row 2, Columns 0-5
-6 28 22 –38 12 –6
; Row 3, Columns 0-5
0 38 45 –22 –35 0
; Row 4, Columns 0-5
In this example each row covers the span of Motor 1 (0 – 50,000 counts) at a constant position of Motor
2; each column covers the span of Motor 2 (0 – 40,000 counts) at a constant position Motor 1.
Enabling and Disabling Tables
All position compensation tables (as well as backlash and torque compensation tables) are enabled when
I51 is set to 1. All of these tables are disabled when I51 is set to 0.
188
Setting Up the Servo Loop
Turbo PMAC User Manual
Active Calculation of Corrections
The position compensation is performed inside the servo loop (every servo cycle) to obtain the maximum
speed and accuracy. Turbo PMAC takes the position of the source motor and finds the matching position
in the table. Typically this is between two entries in a 1D table, or four entries in a 2D table, so Turbo
PMAC linearly interpolates (weighted average) between these entries to obtain the correction for the
current servo cycle. It then adds this correction to the desired position of the target motor. The entries of
corrections in the table must be integer values, with units of 1/16 count (so, for an example, an entry of 48
represents 3 counts) of the target motor.
The net correction for a target motor is stored each servo cycle in a specific register for the motor. For
Motor 1, this is the register at D:$000090. The suggested M-variable for this register for Motor xx is
Mxx69. The units of this register are 1/(Ixx08*32) counts, the same as other motor position registers. If
I51 is set to 1, this register is zeroed every servo cycle, and then the result of each table whose target
motor is this motor is added to this register. If I51 is set to 0, it is permissible to write values directly to
this register. Regardless of the setting of I51, the value in this register is added into the net desired
position value for the motor every servo cycle.
It is important to understand that the table corrections are calculated as a function of the uncorrected
motor position, whether using desired or actual position. For a table that uses the same motor as source
and target, the corrected position is by definition different from the uncorrected position. Since the
compensations vary with position, the compensation at a given corrected position will not be exactly the
same in general as the compensation at the uncorrected position of the same numerical value. The
differences are small, but may be noticeable if you are looking at the net compensation register. This is
important to understand if you are verifying the resulting correction of a table.
Reporting Table Information
The header information for a position compensation table – entered with the DEFINE COMP command –
can be queried with the LIST COMP DEF command. The contents of the table can be queried with the
LIST COMP command
Deleting Tables
The DELETE COMP command erases the position compensation table assigned to the presently addressed
motor (whether or not that motor is a source or target of the table). Position compensation tables must be
deleted in order from those assigned to lower-numbered motors to those assigned to higher-numbered
motors.
Setting Up the Servo Loop
189
Turbo PMAC User Manual
Backlash Compensation
Turbo PMAC can perform sophisticated backlash compensation for all motors. If the position feedback
utilizes a sensor on the motor and there is physical backlash in the coupling to the load (as in a typical
gear train), the physical position of the load for a given sensor-reported position will differ depending on
the direction of motion. Unless this is compensated for, significant position errors can result in the
application.
With Turbo PMAC’s backlash compensation, on reversal of the direction of the commanded velocity, a
pre-programmed backlash distance is added to or subtracted from the commanded position. This
backlash distance can be constant over the travel of the motor, or it can be a function of motor position.
The rate at which the backlash is introduced or removed is programmable, as is the magnitude of the
reversal required for backlash to be introduced or removed. The backlash amount is hidden from any
position reporting.
Constant Backlash Parameter
Variable Ixx86 for Motor xx is the constant backlash distance parameter. When the direction of the
motor’s commanded movement changes from positive to negative, this value is introduced into the active
backlash compensation register, which is subtracted from the nominal commanded position. When the
direction of the motor’s commanded movement changes from negative to positive, the value of the
backlash compensation register is reduced to zero.
Note that a positive value of Ixx86 adds extra distance to the travel of the motor on reversal, which is
what is desired to compensate for true physical backlash. (The only known practical use of a negative
backlash parameter is when the motor is electronically geared as a slave to an axis with physically greater
backlash.) The units of Ixx86 are 1/16 of a count, so the value should be 16 times the number of counts
of backlash compensation required (e.g. Ixx86=72 specifies 4.5 counts of backlash).
Backlash Take-Up Rate
Variable Ixx85 controls the rate at which backlash is introduced or removed upon reversal for Motor xx.
This permits the user to optimize for swift but smooth backlash compensation. When reversal is detected,
each background cycle (between each scan of each PLC) an amount equal to Ixx85 is added to or subtracted
from the active backlash compensation register, as appropriate, until a value Ixx86 or 0 in that register is
reached. In general, the highest value of Ixx85 that produces smooth transitions should be used.
Backlash Hysteresis
Variable Ixx87 controls for Motor xx the number of counts in the new direction of the net commanded
position that must be seen before PMAC determines that a reversal has occurred and the backlash must be
changed. As such, it acts as a “hysteresis” term. It is particularly important if a master encoder is used to
drive the motor, so slight dithering in the master encoder does not cause repeated introduction and
removal of backlash. Ixx87 has units of 1/16 count, so the default value of 64 provides a 4-count
hysteresis.
Backlash Compensation Tables
A backlash compensation table created with the DEFINE BLCOMP command can be used to create
backlash distances that vary with the position of the addressed motor. Most often this is used in
conjunction with a leadscrew compensation table to create the effect of a bi-directional leadscrew
compensation table. In this case, the backlash table (added to the constant backlash parameter) contains
the difference between the positive-going compensation and the negative-going compensation. Delta
Tau’s Flycal calibration program for the PC can generate these tables automatically.
190
Setting Up the Servo Loop
Turbo PMAC User Manual
Entering the Table
The backlash compensation tables are entered and operated much like the position leadscrew
compensation tables. However, there are no cross-axis or multi-axis backlash compensation tables. The
table belonging to a motor provides a backlash correction to that motor as a function of that motor’s
position. Backlash compensation tables must be defined in order from those belonging to highernumbered motors to those belonging to lower-numbered motors.
The backlash compensation table for a motor is declared with the on-line command DEFINE
BLCOMP{entries},{count length} for the addressed motor. {entries} defines the number
of points in the table, and {count length} defines the span of the table in counts of the motor. The
spacing between entries in the table is therefore {count length} / {entries}.
The first entry in the table defines the table’s backlash at one spacing from the zero position of the motor,
the second entry at two spacings, and so on. Typically, the last entry in the table is 0.
Calculating the Backlash
The value of backlash distance for a given motor position derived from the backlash table is added onto
the Ixx86 “constant” backlash parameter. The backlash distance from the table at motor position 0 (home
position) is zero by definition, so if a backlash table is used, Ixx86 should contain the amount of backlash
at the home position. The table then should hold the differences from this amount. The contribution from
the backlash table for a given motor position is calculated by linearly interpolating between adjacent
entries.
The backlash table for a motor is active only if the most recent commanded direction of movement is
negative; it is still active if the motor is currently commanded to stand still but reached this position by
traveling in the negative direction. In operation, the table reads the present nominal motor position and
computes a weighted average of the two closest table entries, creating a first-order interpolation between
table points.
The backlash compensation is defined directly for a range of motor position starting a zero counts and
going in the position direction to the count length declared by the last argument in the DEFINE BLCOMP
command. The spacing between entries is this length divided by the number of entries (which is the first
argument in the command). The first entry in the table defines the correction at one spacing from the zero
position of the motor, the second entry at two spacings, and so on.
Outside this range, the uncorrected position is “rolled over” to within this range before the compensation
is done. This rollover occurs exactly as for leadscrew compensation tables; refer to that description for
details.
Reporting Table Information
The header information for a backlash compensation table – entered with the DEFINE BLCOMP
command – can be queried with the LIST BLCOMP DEF command. The contents of the table can be
queried with the LIST BLCOMP command
Deleting Tables
The DELETE BLCOMP command erases the backlash table for the presently addressed motor. Backlash
compensation tables must be deleted in order from those assigned to the lower-numbered motors to those
assigned to higher numbered motors.
Enabling and Disabling Backlash
The constant backlash parameter Ixx86 is always (potentially) active; that is, there will be backlash
compensation anytime Ixx86 is a non-zero value. Backlash tables are active if I51 is set to 1; they are
inactive if I51 is set to 0.
Setting Up the Servo Loop
191
Turbo PMAC User Manual
Backlash Table Example
Imagine the calibration of an axis assigned to Motor 3 had been performed against an accurate linear
measurement device on the load, working in both directions, and the following readings of the linear
reference device for set positions of the motor encoder (expressed in units of the motor encoder):
0
500
0*
510
5
516
* Reference point; zero by definition
Motor Pos. (cts)
Load Pos.+ (cts)
Load Pos.- (cts)
1000
995
998.5
1500
1492.5
1494
2000
1994
2000
2500
2497.5
2501
3000
3003.5
3010.5
3500
3500.5
3508.5
Only the compensation table works in the positive direction, so the entries in the compensation table
should be the negative of the difference between positive-going load position and motor position,
expressed in 1/16 counts:
0
0*
0*
* Reference point; zero by definition
Motor Pos. (cts)
Load - Motor (cts)
Motor - Load (1/16 cts)
500
+10
-160
1000
-5
+80
1500
-7.5
+120
2000
-6
+96
2500
-2.5
+40
3000
+3.5
-56
3500
+0.5
-8
The position (leadscrew) compensation table definition to create these corrections would be:
#3 DEFINE COMP 8,4000
-160 80 120 96 40 -56 -8 0
Note that the first entry is for the correction at 500 counts, and the added last entry is 0, for the correction
at 4000 counts and 0 counts.
There is a 5-count backlash at motor position 0, so Ixx86 should be set to 5*16, or 80.
Next, the backlash compensation table should contain the differences between negative-going load
position and positive-going load position, minus Ixx86:
Motor Pos. (cts)
Load(-) - Load(+) (cts)
Load(-) - Load(+)Ix86 (cts)
Load(-) - Load(+)Ix86 (1/16 cts)
0
5
0*
500
6
1
1000
3.5
-1.5
1500
1.5
-3.5
2000
6
1
2500
3.5
-1.5
3000
7
2
3500
8
3
0*
16
-24
-56
16
-24
32
48
* Reference point; zero by definition
The backlash table definition to create these corrections would be:
#3 DEFINE BLCOMP 8,4000
16 -24 -56 16 -24 32 48 0
Note that the first entry is for the correction at 500 counts, and the added last entry is 0, for the correction
at 4000 counts and 0 counts.
Note:
While the range and spacing of a backlash table will typically be the same as for
the leadscrew compensation table for the same motor, this is not required. Even
the presence of a leadscrew compensation table for a motor is not required to have
a backlash table for that motor.
192
Setting Up the Servo Loop
Turbo PMAC User Manual
Torque Compensation Tables
Turbo PMAC provides the capability to create a table of corrections as a function of motor position to the
output of the servo loop. Typically, this feature will be used with the servo loop in torque mode (whether
or not Turbo PMAC is also performing motor commutation), so this function is called torque
compensation table.
The torque compensation tables are entered and operated much like the leadscrew compensation tables,
which provide a position correction. However, there are no cross-axis or multi-axis torque compensation
tables. The table belonging to a motor provides a torque correction to that motor as a function of that
motor’s position.
If the motor’s servo loop is outputting a velocity command, the corrections from the torque compensation
tables would actually be velocity corrections.
Entering Tables
The torque compensation table for a motor is declared with the on-line command DEFINE
TCOMP{entries},{count length} for the addressed motor. {entries} defines the number of
points in the table, and {count length} defines the span of the table in counts of the motor. The
spacing between entries in the table is therefore {count length} / {entries}.
The first entry in the table defines the correction at one spacing from the zero position of the motor, the
second entry at two spacings, and so on. If I30 is set to 1 to enable automatic table rollover (highly
recommended), the last entry in the table provides for both the correction at {count length} and at
zero position. Otherwise, the last entry in the table must be 0 if a continuous correction at the rollover
point is desired.
The correction is defined directly for the range of motor positions 0 to {count length}. For motor
positions outside this range, the position is “rolled over” to within this range before the correction is
applied. In this way, cyclic disturbances such as motor cogging torque can be compensated for. The
correction at the end of the table is equivalent to the correction at zero position; if you want this
correction to be non-zero, you must have I30 set to 1 when you enter the table, so Turbo PMAC forces the
correction at zero to this value.
If you enter the table with I30 set to 0, the last entry in the table must be set to 0 for continuity of
correction through table rollover. Because the desired torque correction at the motor’s zero position is
generally not zero, in this case, the entries in the table should contain the difference between the desired
torque correction at that point and the desired torque correction at position zero. The resulting table will
yield a constant torque offset; the integral gain term in the servo loop will then compensate for this offset.
After the table definition command, the next {entries} constants sent to PMAC are put into the table
as table entries. The units of the entries in the table are signed 24-bit values, with a full range of –
8,388,608 to +8,388,607. These values are 256 times larger than the signed 16-bit values of the Ivariables affecting the output, such as Ixx69, Ixx57, Ixx29, and Ixx79, which imply a range of –32,768 to
+32,767. Corrections at points in between entries of the table are linearly interpolated from the adjacent
values in the table.
If the following table were entered:
#1 DEFINE TCOMP 8,
32000
; Corr
-12800
; Corr
21248
; Corr
-24832
; Corr
15360
; Corr
-11008
; Corr
33024
; Corr
-25600
; Corr
Setting Up the Servo Loop
2000 ; Table of 8 entries over 2000 counts for Motor 1
at 2000/8=250 cts is 32000/256=125 16-bit DAC bits
at 500 cts is –12800/256=-50 16-bit DAC bits
at 750 cts is 21248/256=83 16-bit DAC bits
at 1000 cts is –24832/256=-97 16-bit DAC bits
at 1250 cts is 15360/256=60 16-bit DAC bits
at 1500 cts is –11008/256=-43 16-bit DAC bits
at 1750 cts is 33024/256=129 16-bit DAC bits
at 2000 cts (& 0) is –25600/256=-100 16-bit DAC bits
193
Turbo PMAC User Manual
then the correction applied to a 16-bit DAC at 600 counts would be:
Correction = −50 +
600 − 500
(83 − [− 50 ]) = 3bits
750 − 500
How to Calculate Table Entry Values
Torque compensation tables are most commonly used to correct for torque ripple in motors due to
“cogging torque” effects. The simplest way of determining the entries in the table is to command a move
to each point you desire in the table and using integral gain to drive the error to zero, wait for the motor to
settle at this position. Then read the command output of the servo loop as a 24-bit value. This value is
best read in the motor’s quadrature/torque command value (X:$0000BF for Motor 1). Read as a 24-bit
value (e.g. use RX:$BF in the Executive program’s Watch window), this can be the entry into the table
for that position.
Reporting Table Information
The header information for a torque compensation table – entered with the DEFINE TCOMP command –
can be queried with the LIST TCOMP DEF command. The contents of the table can be queried with the
LIST TCOMP command
Enabling and Disabling Tables
All torque compensation tables (as well as position and backlash compensation tables) are enabled when
I51 is set to 1. All of these tables are disabled when I51 is set to 0.
Active Calculation of Corrections
The torque compensation is performed inside the servo loop (every servo cycle) to obtain the maximum
speed and accuracy. Turbo PMAC takes the position of the motor and finds the matching position in the
table. Typically this is between two entries in a table, so Turbo PMAC linearly interpolates (weighted
average) between these entries to obtain the correction for the current servo cycle. It then adds this
correction to the desired output from the position/velocity servo loop of the motor. Both the loop output
and the correction are 24-bit values at this point.
The torque correction for a motor is stored each servo cycle in a specific register for the motor. For
Motor 1, this is the register at Y:$0000BF, and the net output with correction is stored at X:$0000BF. If
I51 is set to 0, it is permissible to write values directly to the correction register. Regardless of the setting
of I51, the value in this register is added into the net desired output value for the motor every servo cycle.
Deleting Tables
The DELETE TCOMP command erases the torque compensation table for the presently addressed motor.
Torque compensation tables must be deleted in order from those assigned to the lower-numbered motors
to those assigned to higher numbered motors.
194
Setting Up the Servo Loop
Turbo PMAC User Manual
TURBO PMAC GENERAL PURPOSE I/O USE
Turbo PMAC controllers have substantial input/output capabilities that are not directly related to servo
operation. I/O points are both digital and analog, both input and output. Board-level Turbo PMAC
controllers have some on-board general-purpose I/O, and more can be added with accessory boards. With
the modular UMAC systems, I/O boards can be added according to the needs of the particular application.
This section summarizes the general-purpose I/O capabilities of the Turbo PMAC family. More details
can be found in the hardware reference manuals for the particular Turbo PMACs, or the manuals for the
particular accessory boards.
Turbo PMAC General-Purpose I/O (JOPTO) Port
The JOPTO port on a Turbo PMAC (J5 on Turbo PMAC PC, PCI, and VME) provides eight generalpurpose digital inputs and eight general-purpose digital outputs. Each input and each output has its own
corresponding ground pin in the opposite row. The 34-pin connector was designed for easy interface to
OPTO-22 or equivalent optically isolated I/O modules. Delta Tau’s Acc-21F is a 180-cm (six-foot) cable
for this purpose. Note that these I/O points are not optically isolated on the Turbo PMAC itself.
Hardware Characteristics
A Turbo PMAC is shipped standard with a ULN2803A sinking (open-collector) output IC for the eight
outputs. These outputs can sink up to 100 mA each, but must have a pull-up resistor to go high.
Do not connect these outputs directly to the supply voltage, or damage to the Turbo PMAC will result
from excessive current draw.
CAUTION:
Having Jumpers E1 and E2 set wrong can damage the IC.
A high-side voltage (+5 to +24V) can be provided into Pin 33 of the JOPTO connector and allow this to
pull up the outputs by connecting pins 1 and 2 of Jumper E1. Jumper E2 must also connect pins 1 and 2
for a ULN2803A sinking output.
CAUTION:
Having Jumpers E1 and E2 set wrong can damage the IC.
It is possible for these outputs to be sourcing drivers by substituting a UDN2981A IC for the ULN2803A.
This IC (U3 on the Turbo PMAC PC, U26 on the Turbo PMAC Lite, U33 on the Turbo PMAC VME) is
socketed, and so may be replaced easily. For this driver, pull-down resistors should be used. With a
UDN2981A driver IC, Jumper E1 must connect pins 2 and 3, and Jumper E2 must connect pins 2 and 3.
Jumper E7 controls the configuration of the eight inputs. If it connects pins 1 and 2 (the default setting),
the inputs are biased to high-side voltage (+5V to +24V) for the OFF state and they must be pulled low
for the ON state. These inputs are best thought of as 24V-tolerant 5V logic, with a switching point at 23V. If E7 connects pins 2 and 3, the inputs are biased to ground for the OFF state and must be pulled
high for the ON state. In either case, a high voltage is interpreted as a 0 by the Turbo PMAC software,
and a low voltage is interpreted as a 1.
Software Access
Usually these inputs and outputs are accessed in software through the use of M-variables. In the
suggested set of M-variable definitions, variables M1 through M8 are used to access outputs 1 through 8,
respectively, and M11 through M18 to access inputs 1 through 8, respectively. This port maps into Turbo
PMAC’s memory space at Y-address $078802.
Turbo PMAC General Purpose I/O Use
195
Turbo PMAC User Manual
Turbo PMAC Multiplexed I/O (JTHW) Port
The Multiplexer port on the JTHW (J3) connector of a Turbo PMAC has eight input lines and eight
output lines. The output lines can be used to multiplex large numbers of inputs and outputs on the port,
and Delta Tau provides accessory boards and software structures (special M-variable definitions TWB,
TWD, TWR, and TWS) to capitalize on this feature. Up to 32 of the multiplexed I/O boards may be
daisy-chained on the port, in any combination.
Alternately, the eight inputs and eight outputs on the port can be used directly by assigning M-variables to
the I/O points themselves. This port maps into Turbo PMAC’s memory space at Y-address $078801.
The suggested M-variable definitions for this use are M40 to M47 for the eight outputs, and M50 to M57
for the eight inputs.
Turbo PMAC Control Panel Port
The control-panel port on the JPAN connector (J2 on Turbo PMAC PC, PCI, VME) of a Turbo PMAC is
a 26-pin connector with dedicated control inputs, dedicated indicator outputs, a quadrature encoder input,
and an analog input. The control inputs are active low in their automatic function with internal pull-up
resistors. They have predefined functions unless the control-panel-disable I-variable (I2) has been set to
1. If this is the case, they may be used as general-purpose inputs by assigning M-variables to their
corresponding memory-map locations (bits of Y address $078800).
Control-Panel Inputs
The JOG-/, JOG+/, PREJ/ (return to pre-jog position), and HOME/ inputs affect the motor selected by the
FDPn/ lines (see below). The STRT/ (run), STEP/, STOP/ (abort), and HOLD/ (feed hold) inputs affect
the coordinate system selected by the FDPn/ lines.
WARNING
Do not change the selector inputs while holding one of the jog inputs low.
Releasing the jog input will then not stop the previously selected motor. This can
lead to a dangerous situation.
The four low-true BCD-coded input lines FDP0/ (LSBit), FDP1/, FDP2/, and FDP3/ (MSBit) form a lowtrue BCD-coded nibble that selects the active motor and coordinate system (simultaneously). Usually
these are controlled from a single 4-bit motor/coordinate-system selector switch. Variable I59 bank
selects which group of eight motors and coordinate systems can be specified by the switch. The motor
selected with these input lines will respond to the motor-specific inputs. It will also have its position
following function turned on (Ixx06 bit 0 is set to 1 automatically); the motor just de-selected has its
position following function turned off (Ixx06 bit 0 is automatically set to 0).
Control-Panel Outputs
There are five dedicated low-true outputs on the JPAN connector, usually used to light LEDs. They are
BRLD/ (buffer-request LED), IPLD/ (in-position LED), EROR/ (fatal following error LED), F1LD/ (first
warning – following error LED), and F2LD/ (which goes true when the watchdog timer trips). BRLD/,
ERLD/, and F2LD/ are global status lines. IPLD/ and F1LD/ are coordinate-system specific status lines.
If I2=0, they refer to the panel-selected coordinate system (by FDPn/ and I59). If I1=1, they refer to the
host-selected coordinate system (&n).
If I2=0 but no coordinate system is selected (all FPDn/ inputs are floating or pulled high), these lines can
be used as general purpose outputs, addressed as bits 20-23 of Y:$078802.
196
Turbo PMAC General Purpose I/O Use
Turbo PMAC User Manual
Turbo PMAC2 General-Purpose I/O (JIO) Port
The JIO port on a Turbo PMAC2 (on Acc-5E for a UMAC Turbo) has 32 discrete digital I/O lines for
general-purpose use. The lines are configurable by byte for input or output (on the DSPGATE2 I/O IC,
the lines are individually configurable for input or output, but the buffer ICs are only byte-configurable),
and individually configurable for inverting or non-inverting format.
Hardware Characteristics
When configured as an output, each line has a 5V CMOS totem-pole driver. This driver can sink or
source up to 20 mA. There is a 10 kΩ pull-up resistor to 5V on each line for input purposes, but the
driver IC can hold the line high or low despite this resistor. When configured as an input, the buffer IC
presents a high-impedance input either sinking or sourcing; no significant current will flow. The pull-up
resistor on the line will bias the line high in the absence of anything actively pulling the line low at
significantly lower impedance.
Note:
Because all of these lines default to inputs at power-up/reset, any lines to be used
as outputs will pull to +5V at power-up/reset until software configures them as
outputs.
Suggested M-Variables
Note:
In a UMAC Turbo system, it is possible to have the DSPGATE2 IC driving the
JIO port at a base address other than the standard $078400. However, this is
unlikely, so the following discussion assumes the standard base address of
$078400 for this IC.
The 32 I/O lines are memory-mapped into PMAC’s address space in registers Y:$078400 and Y:$078401.
Typically these I/O lines are accessed individually with M-variables. A complete list of the suggested Mvariables is shown in the Software Reference; a few are shown here:
M0->Y:$078400,0
M1->Y:$078400,1
…
M23->Y:$078400,23
M24->Y:$078401,0
M25->Y:$078401,1
…
M31->Y:$078401,7
; I/O00 Data Line; J3 Pin 1
; I/O01 Data Line; J3 Pin 2
; I/O23 Data Line; J3 Pin 24
; I/O24 Data Line; J3 Pin 25
; I/O25 Data Line; J3 Pin 26
; I/O31 Data Line; J3 Pin 32
Direction Control
The direction control bit for each of these I/O bits is in the corresponding bit in the matching X register.
For example, the direction control bit for I/O03 is located at X:$078400,3; the direction control bit for
I/O30 is located at X:$078401,6. Because the buffer ICs can be switched only by byte, it is best to define
8-bit M-variables for the direction control. Suggested definitions are:
M32->X:$078400,0,8
M34->X:$078400,8,8
M36->X:$078400,16,8
M38->X:$078401,0,8
;
;
;
;
Direction
Direction
Direction
Direction
control
control
control
control
for
for
for
for
I/O00
I/O08
I/O16
I/O24
to
to
to
to
I/O07
I/O15
I/O23
I/O31
These M-variables should take values of 0 or 255 ($FF) only; 0 sets the byte to input, 255 sets the byte to
output.
Turbo PMAC General Purpose I/O Use
197
Turbo PMAC User Manual
In addition, the bi-directional buffer IC for each byte has a direction control line accessible as a software
control bit. These control lines and bits must match the ASIC direction bits. The buffer direction control
bits are at PMAC address Y:$070800, with bits 0 to 3 controlling the four bytes of the JIO port. A bit
value of 0 specifies input; 1 specifies output. Suggested M-variable definitions are:
M33->Y:$070800,0
M35->Y:$070800,1
M37->Y:$070800,2
M39->Y:$070800,3
;
;
;
;
Buffer
Buffer
Buffer
Buffer
direction
direction
direction
direction
control
control
control
control
for
for
for
for
I/O00
I/O08
I/O16
I/O24
to
to
to
to
I/O07
I/O15
I/O23
I/O31
In the default configuration automatically set at power-up/reset, I/O00 to I/O31 are set up as inputs (M32
through M39 = 0). This is done for maximum safety; no lines can be forced into an undesirable high or
low state. Any of these lines that are to be used as outputs must be changed to outputs by user programs
Usually this is done in PLC 1 acting as a reset PLC, scanning through once on power-up/reset, then
disabling itself.
Inversion Control
Each line on the JIO port is individually controllable as to whether it is an inverting I/O point (0=+5V;
1=0V) or a non-inverting I/O point (0=0V; 1=+5V). Registers X:$078404 and X:$078405 contain the
inversion control bits:
• X:$078404 bits 0 to 23 control I/O00 to I/O23, respectively
• X:$078405 bits 0 to 7 control I/O24 to I/O31, respectively
A value of 0 in the control bit sets the corresponding I/O point as non-inverting. A value of 1 in the
control bits sets the corresponding I/O point as inverting. At power-up/reset, PMAC automatically sets
all of the I/O points on the JIO port as non-inverting.
Alternate Uses
Each general-purpose I/O point on the JIO port has an alternate use as a supplemental fixed-use I/O point
on a supplemental machine interface channel (1* or 2*). The points are individually controllable as to
general-purpose use or fixed use by control registers Y:$078404 and Y:$078405. Refer to these registers
in the memory-I/O map to see the alternate uses of each point. At power-up/reset, Turbo PMAC
automatically sets up all of the I/O points on the port for general-purpose use.
Note:
The byte-wide direction control of the buffer ICs must be set properly for the
alternate uses of the I/O points, just as for the general-purpose I/O uses.
Turbo PMAC2 Multiplexed I/O Port (JTHW)
The JTHW multiplexer port has 16 discrete digital I/O lines for general-purpose use. Most people will
use them in the default configuration of eight inputs and eight outputs, and to support multiplexed I/O
accessories from Delta Tau. When used in this manner, no special setup of the I/O points is required.
However, if it is desired to use these I/O points directly, the following discussion explains their use.
The lines are configurable by byte for input or output (on the DSPGATE2 I/O IC, the lines are
individually configurable for input or output, but the buffer ICs are only byte-configurable), and
individually configurable for inverting or non-inverting format.
Note:
Variable I20 for a Turbo PMAC2 specifies the base address for the first MACRO
IC in the system, the DSPGATE2 IC that controls the JTHW multiplexer port.
This variable must be set correctly for the automatic multiplexer port functions to
work correctly. Usually, this base address is $078400, and the following
discussions assume that the system is set up this way.
198
Turbo PMAC General Purpose I/O Use
Turbo PMAC User Manual
Multiplexer Port Accessories
Delta Tau provides several accessories that can connect to the JTHW multiplexer port, with automatic
firmware support for accessing the I/O on these boards. These accessories provide a direct flat-cable
connection to the JTHW port, and the port is configured automatically at power-up/reset to work with any
of these boards. These accessories include:
• Acc-8D Opt. 7 Resolver-to-Digital Converter Board
• Acc-8D Opt. 9 Yaskawa Absolute Encoder Interface Board
• Acc-18 Thumbwheel Board (obsolete)
• Acc-34x 32-Input 32-Output Boards
For the Acc-8D Option 7 and Acc-8D Option 9 boards, special settings of Ixx91 and Ixx95 support
absolute position reads from these devices. There is also special M-variable formats for accessing I/O
data on most of these boards:
• Acc-8D Opt. 7: TWS-format M-variables for absolute position data
• Acc-18: TWD-format M-variables for binary-coded decimal reads of digits
• Acc-34x: TWS-format M-variables for serial access of 32-bit input or output word
Hardware Characteristics
When configured as an output, each line on the multiplexer port has a 5V CMOS totem-pole driver. This
driver can sink or source up to 20 mA. There is a 10 kΩ pull-up resistor to 5V on each line for input
purposes, but the driver IC can hold the line high or low despite this resistor. When configured as an
input, the buffer IC presents a high-impedance input either sinking or sourcing; no significant current will
flow. The pull-up resistor on the line will bias the line high in the absence of anything actively pulling
the line low at significantly lower impedance.
Suggested M-Variables
The 16 I/O lines are memory-mapped into PMAC’s address space in register Y:$078402. Typically these
lines are used as a unit with specially designed multiplexing I/O accessories and appropriate multiplexing
M-variables (TWB, TWD, TWR, and TWS formats), in which case Turbo PMAC handles the direct control
of these I/O lines automatically. However, these lines can also be accessed individually with M-variables.
The complete list of M-variables is shown in the Software Reference Manual; a few are shown here:
M40->Y:$078402,8
…
M47->Y:$078402,15
M48->Y:$078402,8,8,U
M50->Y:$078402,0
…
M57->Y:$078402,7
M58->Y:$078402,0,8,U
; SEL0 Line; J2 Pin 4
; SEL7 Line; J2 Pin 18
; SEL0-7 Lines treated as a byte
; DAT0 Line; J2 Pin 3
; DAT7 Line; J2 Pin 17
; DAT0-7 Lines treated as a byte
Direction Control
In the default configuration automatically set at power-up/reset, DAT0 to DAT7 are set up as noninverting inputs; SEL0 to SEL7 are set up as non-inverting outputs with a zero (low voltage) value. If
any of the multiplexer port accessories are to be used, this configuration must not be changed.
The direction control bit for each of these I/O bits is in the corresponding bit in the matching X register.
For example, the direction control bit for DAT3 is located at X:$078402,3; the direction control bit for
SEL6 is located at X:$078402,14. Because the buffer ICs can be switched only by byte, it is best to
define 8-bit M-variables for the direction control. Suggested definitions are:
M60->X:$078402,0,8
M62->X:$078402,8,8
Turbo PMAC General Purpose I/O Use
; Direction control for DAT0 to DAT7
; Direction control for SEL0 to SEL7
199
Turbo PMAC User Manual
These M-variables should take values of 0 or 255 ($FF) only; 0 sets the byte to input, 255 sets the byte to
output.
In addition, the bi-directional buffer IC for each byte has a direction control line accessible as a software
control bit. These control lines and bits must match the ASIC direction bits. In the ISA and PCI-bus
versions of the Turbo PMAC, the buffer direction control bits are at Turbo PMAC address Y:$070800,
with bits 4 and 5 controlling the two bytes of the JTHW port. A bit value of 0 specifies input; 1 specifies
output. Suggested M-variable definitions are:
M61->Y:$070800,4
M63->Y:$070800,5
; Buffer direction control for DAT0 to DAT7
; Buffer direction control for SEL0 to SEL7
In the VME-bus versions of the Turbo PMAC, the buffer direction control bits are at Turbo PMAC
address Y:$070802, with bits 0 and 1 controlling the two bytes of the JTHW port. A bit value of 0
specifies input; 1 specifies output. Suggested M-variable definitions are:
M61->Y:$070802,0
M63->Y:$070802,1
; Buffer direction control for DAT0 to DAT7
; Buffer direction control for SEL0 to SEL7
If it is desired to change either of these I/O bytes, it must be done by user programs. Usually this is done
in PLC 1 acting as a reset PLC, scanning through once on power-up/reset, and then disabling itself.
Inversion Control
Each line on the JTHW port is individually controllable as to whether it is an inverting I/O point (0=+5V;
1=0V) or a non-inverting I/O point (0=0V; 1=+5V). Register X:$078406 contains the inversion control bits:
• X:$078406 bits 0 to 7 control DAT0 to DAT7, respectively
• X:$078406 bits 8 to 15 control SEL0 to SEL7, respectively
A value of 0 in the control bit sets the corresponding I/O point as non-inverting. A value of 1 in the
control bits sets the corresponding I/O point as inverting. At power-up/reset, PMAC automatically sets
all of the I/O points on the JTHW port as non-inverting. To use any of the multiplexed I/O accessory
boards on the JTHW port, all I/O points on the port must be left non-inverting.
Alternate Uses
Each general-purpose I/O point on the JTHW port has an alternate use as a supplemental fixed-use I/O
point on a supplemental machine interface channel (1* or 2*). The points are individually controllable as
to general-purpose use or fixed use by control register Y:$078406. Refer to this register in the memoryI/O map to see the alternate uses of each point. At power-up/reset, Turbo PMAC automatically sets up all
of the I/O points on the port for general purpose use.
Note:
Because of the byte-wide direction-control buffer ICs, it is not possible to use all
of the I/O points on the JTHW in their alternate uses.
Turbo PMAC Analog Input (JANA) Port
The analog input (JANA) port is present only if Option 12 is ordered for the Turbo PMAC2 or a Turbo
PMAC PCI board. Option 12 provides 8 12-bit analog inputs (ANAI00-ANAI07). Option 12A provides
8 additional 12-bit analog inputs (ANA08-ANAI15) for a total of 16 inputs.
Hardware Characteristics
The analog inputs can be used as unipolar inputs in the 0V to +5V range, or bipolar inputs in the -2.5V to
+2.5V range. Each input has a 470Ω input resistor in-line, and a 0.033 µF resistor to ground. This
provides a 16 µsec time constant on each input line.
200
Turbo PMAC General Purpose I/O Use
Turbo PMAC User Manual
The analog-to-digital converters on Turbo PMAC require +5V and -12V supplies. These supplies are not
isolated from digital +5V circuitry on PMAC. If the Turbo PMAC2 is plugged into the bus (ISA, PCI, or
VME), these supplies are taken from the bus power supply. In a standalone application, these supplies
must be brought in on terminal block TB1.
The -12V and matching +12V supply voltages are available on the J1 connector to supply the analog
circuitry providing the signals. The +12V supply is not used by Turbo PMAC; it is merely passed
through to the J1 connector for convenience. If use of this supply is desired, it must come either from the
bus supply through Turbo PMAC’s bus connector, or from TB1.
Multiplexing Principle
Only one pair of analog-to-digital converter registers is available to the Turbo PMAC processor at any
given time. The data appears to the processor at address Y:$078800 on a Turbo PMAC2; it appears at
Y:$078808 on a Turbo PMAC PCI. The data from the selected analog input 0 to 7 (ANAI00-ANAI07)
appears in the low 12 bits; the data from the selected analog input 8 to 15 (ANAI08-ANAI15) appears in
the high 12 bits. This data is present only if Option 12A has been ordered.
The input is selected and the conversion is started by writing to this same word address Y:$078800. A
value of 0 to 7 written into the low 12 bits selects the analog input channel of that number (ANAI00ANAI07) to be converted in unipolar mode (0V to +5V). A value of 0 to 7 written into the high 12 bits
selects the analog input channel numbered 8 greater (ANAI08-ANAI15) in unipolar mode. If the value
written into either the low 12 bits or the high 12 bits is 8 higher (8 to 15), the same input channel is
selected, but the conversion is in bipolar mode (-2.5V to +2.5V).
De-multiplexing I-Variables
Turbo PMAC I-variables I5060 – I5096 permit the automatic de-multiplexing of these multiplexed A/D
converters (and of multiplexed A/D converters on external Acc-36 and Acc-59 boards as well).
I5060 controls the number of A/D converter pairs accessed in the de-multiplexing ring, up to 16 pairs.
Variables starting at I5061, and possibly up to I5076, specify the Turbo PMAC address of each ADC pair
to be read. The addresses of all eight pairs on the JANA port are located at $078800.
Variables starting at I5081, and possibly up to I5096, specify which pair of ADCs at the address specified
by the I-variable numbered 20 lower is read (e.g. I5081 for I5061), and how it is to be converted. I5081 –
I5096 are 24-bit values, represented by 6 hexadecimal digits. Legitimate values are of the format
$00m00n, where m and n can take any hex value from 0 through F.
For the on-board Option 12 and 12A ADCs on a Turbo PMAC2, the m value determines which of the
inputs ANAI08 to ANAI15 that come with Option 12A is to be read, and how it is to be converted,
according to the following formulas:
•
m = ANAI # −8
; 0 to +5V unipolar input
•
m = ANAI # ; -2.5V to +2.5V bipolar input
For the on-board Option 12 and 12A ADCs on a Turbo PMAC2, the n value determines which of the
inputs ANAI00 to ANAI07 that come with Option 12A is to be read, and how it is to be converted,
according to the following formulas:
•
n = ANAI #
; 0V to +5V unipolar input
•
n = ANAI # +8
; -2.5V to +2.5V bipolar input
The results of this A/D de-multiplexing are placed in registers at addresses Y:$003400 to Y:$00341F,
using bits 12 to 23 of these registers. The value of the A/D converter found in the low 12 bits of the
source register is placed in the register with the even-numbered address; the value of the A/D converter
found in the high 12 bits of the source register is placed in the register with the odd-numbered address.
Refer to the Turbo PMAC memory map or I5061 – I5076 description for details. Suggested M-variables
for the result registers are M5061 – M5076.
Turbo PMAC General Purpose I/O Use
201
Turbo PMAC User Manual
In operation, Turbo PMAC reads one ADC pair each phase cycle and copies it into the appropriate
memory registers. Therefore, it reads each ADC pair every I5060 phase cycles. If these values are used
as feedback for a servo loop, the loop should not be closed more often than the ADC is read.
UMAC Digital I/O Boards
The UMAC has an extensive family of digital I/O boards. The following table summarizes these boards
and their properties:
ACC
# of
Inputs
Input
Range
Drivers for
Inputs
# of
Outputs
Output
Range
Output
Drivers
Notes
9E
48
–
–
Isolated I/O
0
Sink/Source by
wiring
–
0
10E
12V24V
–
48
5V – 24V
Isolated I/O
11E
24
14E
0 - 48
12V24V
5V
Sink/Source by
wiring
TTL/Sinking
48 - 0
5V
65E
24
Sourcing only
24
5V – 24V
66E
48
Sourcing only
0
–
–
67E
0
12V24V
12V24V
–
Sink/Source by
factory option
Sink/Source by
factory option
Sinking with
pull-ups
Sourcing only
–
48
5V – 24V
Sourcing only
5V – 24V
Isolated I/O
TTL, each
reversible
Isolated,
protected I/O
Isolated,
protected I/O
Isolated,
protected I/O
Addressing UMAC I/O Boards
All of these boards utilize Delta Tau’s memory-mapped IOGATE ASIC, which supports 48 I/O points
mapped as 6 consecutively addressed 8-bit registers. In the next two 8-bit registers are setup and control
bits. The base address of the IOGATE IC is set by jumpers on the Acc- 9E, 10, and 11E boards, set by
DIP-switches on the 14E, 65E, 66E, and 67E boards.
The Acc-14E, 65E, 66E, and 67E boards support automatic identification by the UMAC CPU. Status
variables I4950 – I4965 indicate which of these boards are present, and at what addresses. The Turbo
Setup program can identify and display the results of these variables automatically, indicating which
boards are present. It can also identify the presence of Acc-9E, 10E, and 11E boards, but because these
older boards lack self-identification features, it cannot tell which type of board each one is.
Boards With Jumper-Set Addresses
For the Acc-9E, 10E, and 11E boards, the base address of the board is set by putting a jumper on one and
only one of the E-points E1 – 4. The registers on the board are found at eight consecutive addresses:
{Base} through {Base + 7}. The byte – low (bits 0 – 7), middle (bits 8 – 15), or high (bits 16 – 23) – of
Turbo PMAC’s 24-bit word in which the registers are found is determined by the setting of the jumper
bank E6A – H. The settings for the base addresses for these boards are summarized in the following
table:
202
Address
Jumper
E6A H Pins 1&2
E6A H Pins 2&3
E6A H Pins 1&2
E1
E2
E3
E4
Y:$078C00 Bits 0 - 7
Y:$078D00 Bits 0 - 7
Y:$078E00 Bits 0 - 7
Y:$078F00 Bits 0 - 7
Y:$078C00 Bits 8 - 15
Y:$078D00 Bits 8 - 15
Y:$078E00 Bits 8 - 15
Y:$078F00 Bits 8 - 15
Y:$078C00 Bits 16 - 23
Y:$078D00 Bits 16 - 23
Y:$078E00 Bits 16 - 23
Y:$078F00 Bits 16 - 23
Turbo PMAC General Purpose I/O Use
Turbo PMAC User Manual
Boards With Switch-Set Addresses
For the Acc-14E, 65E, 66E, and 67E boards, the base address of the board is determined by the settings of
DIP switches SW1-1 through SW1-4. When these boards are used with a UMAC Turbo CPU, SW1-5
and SW1-6 must always be ON. These boards always appear in the low byte (bits 0 – 7) of the 24-bit
word.
Switch Settings
SW1-3 ON
SW1-4 ON
SW1-3 OFF
SW1-4 ON
SW1-3 ON
SW1-4 OFF
SW1-3 OFF
SW1-4 ON
SW1-1 ON
SW1-2 ON
SW1-1 OFF
SW1-2 ON
SW1-1 ON
SW1-2 OFF
SW1-1 OFF
SW1-2 OFF
Y:$078C00
Y:$079C00
Y:$07AC00
Y:$07BC00
Y:$078D00
Y:$079D00
Y:$07AD00
Y:$07BD00
Y:$078E00
Y:$079E00
Y:$07AE00
Y:$07BE00
Y:$078F00
Y:$079F00
Y:$07AF00
Y:$07BF00
Note that both types of boards can be set up to the same addresses in some cases. It is, of course, very
important not to have any addressing conflicts.
Setting up UMAC I/O Boards
I/O points on the IOGATE IC itself are selectable by byte for input or output. However, only the Acc14E TTL-level I/O board gives you a choice as to which I/O points will be inputs and which will be
outputs. On all the other of these boards, the surrounding buffer/driver circuitry determines how each I/O
point must be used. The IOGATE IC must be set up each power-on/reset to determine the direction of
each I/O point; typically this is done in a “one-shot” PLC program. The manual for each board shows
example program code that could be used to do this.
In typical applications, very little setup of the IOGATE IC is required for operation with the UMAC I/O
boards. However, the IOGATE IC has special features that are useful in unusual applications. The
following section details how the control register and the setup registers of the IOGATE IC can be used to
provide great flexibility
Control Register
The control register at address {Base + 7} permits the configuration of the IOGATE IC to a variety of
applications. The control register consists of 8 write/read-back bits – Bits 0 - 7.
Direction Control Bits
Bits 0 to 5 of the control register simply control the direction of the I/O for the matching numbered data
register. That is, Bit n controls the direction of the I/O at {Base + n}. A value of 0 in the control bit (the
default) permits a write operation to the data register, enabling the output function for each line in the
register. Enabling the output function does not prevent the use of any or all of the lines as inputs, as long
as the outputs are off (non-conducting). A value of 1 in the control bit does not permit a write operation
to the data register, disabling the output, reserving the register for inputs.
For example, a value of 1 in Bit 3 disables the write function into the data register at address {Base + 3},
ensuring that lines IO24 - IO31 can always be used as inputs.
Register Select Control Bits
Bits 6 and 7 of the control register together select which of 4 possible registers can be accessed at each of
the addresses {Base + 0} through {Base + 5}. They also select which of two possible registers can be
selected at {Base + 6}.
Turbo PMAC General Purpose I/O Use
203
Turbo PMAC User Manual
The following table explains how these bits select registers:
Bit 7
Bit 6
Combined
Value
0
0
0
0
1
1
1
0
2
1
1
3
* With bits 0 to 5 set to 0
Byte
Value*
{Base + 0} to {Base + 5}
Register Selected
{Base + 6} Register
Selected
$00
$40
$80
$C0
Data Register
Setup Register 1
Setup Register 2
Setup Register 3
Data Register
Setup Register
n. a.
In a typical application, non-zero combined values of Bits 6 and 7 are only used for initial configuration
of the IC. These values are used to access the setup registers at the other addresses. After the
configuration is finished, zeros are written to both Bits 6 and 7, so the data registers at the other registers
can be accessed.
Setup Registers
There are a total of four registers accessible at each of the IC addresses {Base + 0} to {Base + 5}: three 8bit setup registers and an 8-bit data register. The setup registers control how data is written to and read
from the data registers.
Setup Register 1: Inversion Control
Setup Register 1 at each address {Base + 0} through {Base + 5}, which is selected by writing a 1 to Bit 6
of the Control Word at {Base + 7} and a 0 to Bit 7, is the inversion control register for the Data Register
at the same address. Each bit of Setup Register 1 controls the inversion of the matching bit of the Data
Register at the same address.
A value of 0 in a bit of Setup Register 1 specifies an inverting I/O point for the matching bit of the Data
Register at the same address. That is, for an output, a value of 0 produces a low (conducting) output, and
a value of 1 produces a high (non-conducting) output. For an input, a line pulled low produces a 1 value,
and a line pulled high or permitted to float high produces a 0 value.
A value of 1 in a bit of Setup Register 1 specifies a non-inverting I/O point for the matching bit of the
Data Register at the same address. That is, for an output, a value of 0 produces a high (non-conducting)
output, and a value of 1 produces a low (conducting) output. For an input, a line pulled low produces a 0
value, and a line pulled high or permitted to float high produces a 1 value.
Setup Register 2: Read Control
Setup Register 2 at each address {Base + 0} through {Base + 5}, which is selected by writing a 0 to Bit 6
of the Control Word at {Base + 7} and a 1 to Bit 7, is the read control register for the Data Register at the
same address. Each bit of Setup Register 2 controls what data is read from the matching bit of the Data
Register at the same address.
The action of a bit of Setup Register 2 is dependent on the setting of the matching bit of Setup Register 3
for the same address. If the matching bit of Setup Register 3 is 0, selecting unlatched inputs, the bit of
Setup Register 2 controls whether the pin value is read, or the value in the writeable register is read. A
value of 0 in the bit of Setup Register 2 selects the pin value to be read from the matching bit of the Data
Register at the same address; a value of 1 in the bit selects the writeable register value.
If the matching bit of Setup Register 3 is 1, selecting latched inputs, the bit of Setup Register 2 controls
whether the directly latched data is read, or the value that is the result of a Gray-code-to-binary
conversion. A value of 0 in the bit of Setup Register 2 selects the directly latched value to be read from
the matching bit of the Data Register at the same address; a value of 1 in the bit selects the value that is
the result of a Gray-code-to-binary conversion.
204
Turbo PMAC General Purpose I/O Use
Turbo PMAC User Manual
Setup Register 3: Latch Control
Setup Register 3 at each address {Base + 0} through {Base + 5}, which is selected by writing a 1 to Bit 6
of the Control Word at {Base + 7} and a 1 to Bit 7, is the latch control register for the Data Register at the
same address. Each bit of Setup Register 3 controls whether latched or unlatched data is read from the
matching bit of the Data Register at the same address.
A value of 0 in the bit of Setup Register 3 selects unlatched data to be read from the matching bit of the
Data Register at the same address; a value of 1 in the bit selects latched data to be read.
For both the latched and unlatched settings, the matching bit of Setup Register 2 controls exactly what
type of data is read from the Data Register.
Data Registers
The Data Register at each address {Base + 0} through {Base + 5}, which is selected by writing a 0 to Bit
6 of the Control Register at {Base + 7} and a 0 to Bit 7, provides the working interface for the 8
input/output lines matched to that address. The processor reads from or writes to the data register to
access the input/output lines.
If there is a value of 1 in Bit n (n = 0 to 5) of the Control Word, a write operation to the Data Register at
address {Base + n} has no effect on the I/O line, effectively disabling the output function for all 8 lines
associated with the register.
A read operation from a Data Register can access one of 4 types of data for each I/O line associated with
the register (individually selectable), depending on how the Setup Registers 2 and 3 at the same address
have been configured.
The following table summarizes how the Setup Register bits control what data is read in the matching bit
of the Data Register:
Setup Register 3 Bit Value
Setup Register 2 Bit Value
0
0
1
1
0
1
0
1
Data Type Read
Pin Value Read
Writeable Register Read
Latched Input Read
Converted Gray Code Read
Pin Data Read
If the pin value has been selected to be read, the line voltages for the eight I/O lines is read as an input,
even if it is being driven as an output.
Writeable Register Data Read
If the writeable register value has been selected to be read, the value written by the processor into the data
register is read back, even if this does not match the voltage state of the pin.
Latched Input Read
If the latched input has been selected to be read, the input value last latched by the falling edge of line En
for address {Base + n} is read, even if the input value has changed since then.
Converted Gray Code Read
If the converted Gray code input has been selected to be read, the input value last latched by line En for
address {Base + n} and processed through the Gray code conversion circuitry is read.
Turbo PMAC General Purpose I/O Use
205
Turbo PMAC User Manual
206
Turbo PMAC General Purpose I/O Use
Turbo PMAC User Manual
MAKING AN APPLICATION SAFE
Delta Tau Data Systems has provided many safety features on the Turbo PMAC controller, and invested
many resources to make Turbo PMAC a safe product. However, the ultimate responsibility for the safety
of a control system using Turbo PMAC must lie with the system designer, utilizing the safety features on
Turbo PMAC and in other parts of the system.
Following Error Limits
Turbo PMAC has three following error limits for each motor. Following error is the difference between
the commanded position and the actual position at any time. The following error limit is an important
protection against serious system faults such as loss of feedback, which can cause dangerous conditions
like full speed runaway.
Fatal Following Error Limit
One of these limits (Ixx11) is a fatal limit which causes a shutdown of the motor: servo loop opened, zero
output commanded, amplifier disabled (i.e. the motor is killed), moves for the motor and programs for the
motor’s coordinate system aborted.
Which motors get killed – only the offending motor, all in the coordinate system, or all in Turbo PMAC –
is determined by bits 21 and 22 of Ixx24 for the faulted motor. This limit is intended for conditions in
which something has gone seriously wrong (e.g. loss of feedback or power stage) and all operation should
cease.
After the motor or motors have been killed due to the fatal following error limit, closed-loop control can
be re-established with the J/ command (single motor), the A command (coordinate system), or the
<CTRL-A> command (entire card).
WARNING
Although this limit may be disabled by setting Ixx11 to zero, but this is strongly
discouraged in any application that has the potential to kill or injure people, or
even to cause property damage. Disabling the fatal limit removes an important
protection against serious fault conditions that can cause runaway situations,
bringing the system to full power output faster than anybody could react.
Good tuning of the motor’s servo loop is important for safety reasons as well as performance reasons.
The smaller the true following errors are made during proper operation, the tighter the Ixx11 limits can be
set without getting nuisance trips. Particularly important in this regard are the feedforward terms that can
dramatically reduce the errors at high speeds and accelerations.
Warning Following Error Limit
The second limit (Ixx12) is a warning limit – when exceeded, Turbo PMAC sets status bits for the motor
and the motor’s coordinate system, and can set output lines on the control panel connector, the machine
connectors, and through the programmable interrupt controller (for ISA and PCI bus Turbo PMACs).
This permits special action to be taken, either by Turbo PMAC itself through a PLC program, by the host,
which can find out through an interrupt or by polling the card, or by an operator notified with one of the
external signals.
The warning following error status bit for the motor can be used as the trigger condition for any of Turbo
PMAC’s automatic triggered moves (homing-search move, jog-until-trigger, programmed RAPID-mode
move-until-trigger) if bit 0 of Ixx97 is set to 1. This permits easy implementation of tasks such as homing
into a hard stop, torque-limited screwdriving, etc.
Making your Application Safe
207
Turbo PMAC User Manual
Integrated Following Error Protection
In addition to the normal following error protection provided by the Ixx11 variable for each motor, Turbo
PMAC can shut down the motor if the time-integrated value of the following error exceeds a preset value.
This integrated error feature can protect against those cases in which the magnitude of the measured
following error never gets very large – for example, a loss of feedback followed by a very short
commanded move.
Turbo PMAC performs the integrated following error check only if the Ixx63 integration limit parameter
is less than zero. When this is the case, the magnitude of Ixx11 is used for the normal unintegrated
following error check. In addition, the value of the PID integrator is compared against the Ixx63
integration limit magnitude. If the integrator value has saturated at +/-Ixx63 (the limiting function in the
PID loop will not let it exceed this value), then Turbo PMAC will trip (kill) this motor on an integrated
following error fault, just as it would for a normal following error fault.
For the integrated following error limit to be effective, the Ixx33 integral gain must be greater than zero,
and preferably set as high as can be tolerated. Also, the Ixx34 integration mode parameter must be set to
0, so that the integrator is on during programmed moves.
Remember that the integrator stops increasing in magnitude if the command output has saturated at Ixx69.
The magnitude of Ixx63 must be small enough that it will trip before the output saturates. The magnitude
of Ixx63 that would cause output saturation at Ixx69 from the integrator alone is:
 Ixx69 ∗ 2 23 

Ixx63 = 
 Ixx08 ∗ Ixx30 
The magnitude of Ixx63 must be less than this value for the shutdown function to be effective.
Remember that there will be other components to the output, for instance from the proportional gain.
With a bare motor, test to see that this limit can trip the motor reliably.
When a motor is killed due to integrated following error fault, the standard following error fault motor
status bit is set. In addition, a separate integrated following error fault motor status bit is set. Both bits
are cleared when the motor is re-enabled.
Position (Overtravel) Limits
Turbo PMAC has both hardware and software overtravel position limit features. These are intended to
prevent motion accidentally commanded out of the legal range of positions.
Hardware Overtravel Limit Switches
The axis-interface circuitry associated with each servo interface channel in a Turbo PMAC system has
positive and negative hardware overtravel limit switch inputs. The exact nature of this input circuitry and
instructions for connecting the limit switches are described in the Hardware Reference Manual for each
Turbo PMAC and axis-interface accessory.
Generally, these inputs are optically isolated, with a failsafe circuit design. The limit switches must be
normally closed, conducting current through the opto-isolator when the axis is not in the limit. This
conducting condition produces a zero state in the flag register for the channel in the Servo IC; the
processor must read this zero to permit motion in that direction.
Anything that stops current from flowing through the opto-isolator, whether from actually hitting the
limit, from cable disconnection, or from loss of power supply for the limit circuit, produces a one state in
the Servo IC. When the processor sees this, it will not permit motion in that direction.
208
Making your Application Safe
Turbo PMAC User Manual
Variable Ixx25 for the motor must contain the address of the flag register for the channel into which these
limit switches are wired. Bit 17 of Ixx24 must be set to the default value of 0 to use these limit inputs. If
this bit is set to 1, Turbo PMAC will not monitor these limit inputs. This bit can be set permanently to 1
if the motor does not have limit switches; it can be set to 1 temporarily for operations such as homing into
a limit.
On hitting a limit, Turbo PMAC decelerates the offending motor at a user-programmed rate as defined by
the motor’s Ixx15. If the motor is in a coordinate system that is running a motion program at the time, all
motors in the coordinate system are decelerated to a stop at their own Ixx15 rate. The effect is equivalent
to issuing an A (abort) command. If the coordinate system has been executing a path move, this
deceleration will not necessarily be along that path.
Note:
Turbo PMAC brings the commanded trajectory for the motor to a stop at the Ixx15
rate as soon as it detects a limit condition. If there is significant following error at
the time, the actual position can try to catch up to the commanded position for a
long period of time. With a large enough following error, it is possible that the
commanded position would be well past the limit and into the hard stop. It is
important to set a reasonable fatal following error limit and to allow sufficient
room past the limit switch to absorb errors up to that following error limit.
The limit input pins are direction sensitive: the positive-end limit pin stops positive direction moves only
(those coming at it from the negative side), and the negative-end limit pin stops negative direction moves
only (those coming at it from the positive side). This makes it possible to command a move out of the
limit that you have run into. However, this also makes it essential to have your limit switches wired into
the proper inputs, or they will be useless.
Software Overtravel Limit Variables
Turbo PMAC also has positive and negative software limits for each motor to complement or replace the
hardware limits. These limits can use the motor’s desired as well as actual positions. Motor variables
Ixx13 and Ixx14 define the positive and negative actual position limits, respectively, in counts, for Motor
xx. These limits are referenced to the motor’s zero (home) position, and do not change if the
programming origin for the associated axis is offset.
Turbo PMAC continually compares the motor’s actual position to these limits. The behavior on
exceeding one of these limits is the same as hitting a hardware limit. A value of zero in one of these
parameters disables that software limit.
If bit 15 of Ixx24 for the motor is set to 1, Turbo PMAC will compare the motor’s desired position also,
calculated ahead of time for programmed moves, to the software limits. This permits the motor to come
to a full stop within the limits, not just begin to decelerate at the limits. This in turn can provide an extra
useful range of motion at the perimeter of a machine.
Note:
Turbo PMAC cannot pre-calculate the desired position for indefinite jog moves J+
and J-, and so will not begin to decelerate on these moves until the present desired
position exceeds a software limit. If it is desired that jog moves stop within the
limits, a J+ command should be replaced by a definite jog to the positive limit; a
J- command should be replaced by a definite jog to the negative limit.
Making your Application Safe
209
Turbo PMAC User Manual
For a LINEAR or CIRCLE-mode move executed with the special lookahead buffer active and exceeding
a desired position limit, the move will come to a stop along the programmed path exactly at the limit,
decelerating as controlled by the Ixx17 maximum-acceleration parameters for the motors in the
coordinate system. This is the equivalent of the \ quick-stop command. In this case, it is possible to
resume motion along the path after changing the offending limit parameter. (If bit 14 of Ixx24 is set to 1,
motion of these moves in lookahead mode will not stop on hitting a limit; instead, the commanded
position of the offending motor will saturate at the limit value.)
Motor variable Ixx41 defines the difference (in counts) between the motor’s actual and desired position
limits, permitting the desired position limits to be set inside the actual position limits. The positive
desired position limit is set at (Ixx13 - Ixx41); the negative desired position limit is set at (Ixx14 + Ixx41).
This can be important to ensure that the limit deceleration of path moves within lookahead stay on the
path, and to enable the program to be resumed if desired. Ixx41 should be set large enough to make sure
that the actual position during the deceleration from exceeding a desired position limit does not exceed
the actual position limit.
The software position limits are disabled automatically during homing search moves, until the homing
trigger is found. As soon as the trigger is found, the software limits are re-activated, using the new home
position as the reference. The software position limits are always referenced to the motor’s zero position,
whether established by a homing search move, an absolute position read, or just at power-up/reset. If the
programming origin of the axis assigned to the motor is offset, the software position limits are not
automatically offset with the programming origin.
Velocity Limits
Vector Velocity Limit
Turbo PMAC has a vector velocity limit parameter (Isx98), known as maximum feedrate for each
coordinate system. For programs run by this coordinate system, the value of any F (feedrate) command
in the program is compared to Isx98. If the value is greater than Isx98, the value of Isx98 is used instead
as the feedrate command. Isx98 is expressed in the same user velocity units (axis length units divided by
Isx90 milliseconds) as the feedrate command itself.
Motor Velocity Limit
Turbo PMAC has a programmable velocity limit parameter for each motor (Ixx16), in units of counts per
millisecond, which has several functions. First, it serves as the commanded velocity for the motor in
RAPID-mode moves if the motor’s rapid velocity-select parameter Ixx90 for the motor is set to the
default value of 1.
Second, for simple LINEAR-mode moves with move segmentation disabled (Isx13=0), Ixx16 serves as
the maximum velocity permitted. If the commanded velocity requested of a motor exceeds the limit for
the motor, the move is slowed so that the velocity limit is not exceeded. In a multi-axis programmed
move, all axes in the coordinate system are slowed proportionally so that no change in path occurs.
In addition, for LINEAR and CIRCLE-mode moves executed with segmentation enabled (Isx13>0) and
the special lookahead buffer active, it serves as the maximum velocity for each segment of the motion.
This can be particularly valuable for non-Cartesian systems programmed with Turbo PMAC’s kinematic
equations; very high motor velocities can inadvertently be commanded near “singularities.” The
lookahead algorithm can detect these problems beforehand, and slow the motion down along the path into
the problem point, observing the Ixx17 motor acceleration limits.
Velocities are compared to these limits assuming no feedrate override (% value of 100); if feedrate
override (a.k.a. time-base control) is used, the velocity limits scale with the override.
210
Making your Application Safe
Turbo PMAC User Manual
Acceleration Limits
Turbo PMAC has two programmable acceleration limits for each motor, one for jogging, homing, and
RAPID-mode moves (Ixx19), and one for LINEAR and CIRCLE-mode programmed moves (Ixx17).
Both parameters are in units of counts per (millisecond-squared). PVT and SPLINE-mode moves do not
observe either of these limits.
If the commanded acceleration requested of a motor by the change in velocity and the acceleration time
parameters exceeds the limit for the motor, the acceleration time is extended so that the acceleration limit
is not exceeded. In a multi-axis programmed move, all axes in the coordinate system are slowed
proportionally so that no change in path occurs. Accelerations are compared to these limits assuming no
feedrate override (% value of 100); if feedrate override (a.k.a. time-base control) is used, the acceleration
limits scale with the square of the override percentage.
Without the special lookahead buffer enabled, the Ixx17 acceleration limit works only on LINEAR-mode
moves with segmentation disabled (Isx13=0). In this mode of operation, the acceleration time can be
extended only equal to the move time of the incoming move. If this is not enough of an extension to
observe the acceleration limit, the limit will be violated.
The Ixx17 limit works on LINEAR and CIRCLE-mode moves executed with segmentation enabled
(Isx13>0) and the special lookahead buffer active. Ensuring that all these moves observe this acceleration
limit is the most important feature of the special lookahead buffer.
Command Output Limits
Turbo PMAC has a programmable output limit (on the command Turbo PMAC sends to the amplifier or
the internal commutation algorithm) for each axis (Ixx69) that acts as a torque/force limit for currentmode, sine-wave, or direct-PWM amplifiers, or a speed limit for velocity-mode amplifiers. If this limit is
engaged to change what the servo loop commands, Turbo PMAC’s anti-windup protection activates to
prevent oscillation when coming out of the limiting condition. In addition, there is a limit on the size of
the error that the feedback filter is permitted to see (the Ixx67 “position error” limit), which has the effect
of slowing down too sudden a move in a controlled fashion.
Integrated Current (I2T) Protection
Turbo PMAC can be set up to fault a motor if the time-integrated current levels exceed a certain
threshold. This can protect the amplifier and/or motor from damage due to overheating. It can either
2
integrate the square of current over time – commonly known as I T (eye-squared-tee) protection, or
integrate the absolute value of current over time – usually called |I|T (eye-tee) protection. I2T protection
is used when the most thermally sensitive components are resistive in nature (e.g. motor windings or FET
transistors), because their power dissipation is proportional to the square of current. |I|T protection is used
when the most thermally sensitive components have a constant voltage drop (e.g. bipolar transistors),
because their power dissipation is proportional to the magnitude of current. Because the use of the square
of current is more common, this protection is generically referred to as “I2T.”
Some amplifiers have their own internal integrated-current protection, but many others do not. Turbo
PMAC’s integrated-current protection can be used in either case. It can be used with any amplifier for
which Turbo PMAC computes current commands, whether or not Turbo PMAC also performs the
commutation and/or digital current loop functions. If Turbo PMAC is closing the current loop for the
motor, this function uses the measured current values; otherwise it uses the commanded current values.
This protection is not suitable for use in systems in which Turbo PMAC outputs a velocity command,
either as an analog voltage or a pulse frequency.
Making your Application Safe
211
Turbo PMAC User Manual
2
Two I-variables control the functioning of the I T protection for each motor. Ixx57 is the continuous
current limit magnitude. It has the same units as the Ixx69 instantaneous output limit, bits of a 16-bit
DAC (even if some other output device is used). Both have a maximum magnitude of 32,767, which is
the size of Turbo PMAC’s maximum possible output. If Ixx57 is a positive value, I2T protection will be
used; if Ixx57 is a negative number, |I|T protection will be used. Generally Ixx57 will be 1/4 to 1/2 of the
magnitude of Ixx69.
Ixx58 is the integrated current limit parameter. If Ixx58 is set to 0, this function is disabled. If Ix58 is
greater than 0, Turbo PMAC will compare the integrated current value to Ixx58. When the integrated
current value exceeds this value, Turbo PMAC will fault this motor as if an amplifier fault had occurred.
The offending motor is killed; if it was in a coordinate system running a motion program, that motion
program aborted; other motors are killed according to the setting of bits 21 and 22 of Ixx24.
2
Turbo PMAC’s I T function works according to the following equation:
2
2
 I
2
 q 
 Id 
 Ixx57  





+
Sum = Sum +
 −  32768   ∆t
 32768 
 32768 



Turbo PMAC’s |I|T function works according to the following equation:

 I q   I


  d   Ixx57 
Sum = Sum + 
∆t
+
−
 32768   32768   32768 



 


where:
Iq
Id
∆t
(quadrature current) is the commanded torque-producing output of the PID filter in units of a 16bit DAC;
(direct current) is the magnetization current command as set by Ixx77. This is usually zero
except when Turbo PMAC is doing vector control of induction motors.
is the time since the last sample in servo cycles
If Sum exceeds Ixx58, an integrated-current fault will occur. When commanded current levels are below
Ixx57, Sum will decrease, but it will never go below zero.
PMAC I2T Protection Feature
Normalized
I2
I2
32768 2
1.0 2
Ix692
 Ix69 


 32768 
2
Saturation
Integrator
Charge-Up
~
I2T Fault
2
~
 Ix57 


 32768 
~
Ix57 2
~
Integrator
Discharge
time
(servo cycles)
Example: With command output limit Ixx69=32767 (maximum), integrated current limit Ixx57=16384
(half of maximum), and magnetization current Ixx77=0, the motor hits an obstruction, and the command
output saturates at 32767. The integrated-current protection function will calculate during this time:
212
Making your Application Safe
Turbo PMAC User Manual
[
]
Sum = Sum + 1 2 + 0 2 − 0.5 2 ∆t = Sum + 0.75 ∆t
Sum will increase at a rate of 0.75 per servo cycle. At the default servo cycle update rate of 2.25 kHz,
Sum will increase at a rate of 2250*0.75=1688 per second. If you want the motor to trip after 3 seconds
of this condition, you should set Ixx58 to 1688*3 = 5064.
When an integrated-current fault occurs on a motor, Turbo PMAC reacts just as for an amplifier fault
error. The offending motor is killed and possibly other motors as set by bits 21 and 22 of Ixx24. Turbo
PMAC sets the amplifier fault motor status bit. For an integrated-current fault, Turbo PMAC also sets a
separate integrated-current fault motor status bit. Both bits are cleared when the motor is re-enabled.
Note:
2
When Turbo PMAC is not commutating a motor with I T protection, make sure
magnetization current parameter Ixx77 is still set to 0. In this setup, Ixx77 will not
affect operation, but it will affect integrated-current calculations.
Amplifier Enable and Fault Lines
The use of the amplifier-enable (AENAn) output and the amplifier-fault (FAULTn) input lines for each
motor are important for safe operation. Without the use of the enable line, disabling the amplifier relies
on precise zero offsets in Turbo PMAC’s analog output and the amplifier’s analog input.
Without the use of the fault line, Turbo PMAC may not know when an amplifier has shut down and may
not take appropriate action.
Note:
With the default sinking drivers for the amplifier enable signals, using the low-true
enable polarity (low voltage – conducting – is enable; high voltage – nonconducting – is disabled) provides better failsafe protection against loss of powersupply. If either the +5V supply for Turbo PMAC’s computational section, or the
+15V analog supply is lost, the amplifier will be disabled automatically, because
the output transistor will go into its non-conducting state. If you desire this failsafe
protection but cannot connect a signal of this polarity directly to the amplifier, you
must use intermediate circuitry to change the signal format. With the alternate
sourcing drivers, the high-true enable polarity provides better failsafe protection.
Encoder-Loss Detection
Most Turbo PMAC controllers have encoder-loss detection circuitry for each encoder input. Designed for
use with encoders with differential line-driver outputs, the circuitry monitors each input pair with an
exclusive-or (XOR) gate. If the encoder is working properly and connected to the Turbo PMAC, the two
inputs of the pair should be in opposite logical states – one high and one low – yielding a true output from
the XOR gate.
If the input circuits for the encoder have been configured so that both lines of the pair have pull-up
resistors (this is not the default – either an E-point jumper must be changed or a SIP resistor pack reversed
in its socket), then an encoder or cabling failure will cause both inputs into the same logical state, yielding
a false output from the XOR gate, and setting an error status bit. Note that in this setting, a single-ended
encoder cannot be used on the channel.
Making your Application Safe
213
Turbo PMAC User Manual
The following table shows the resistor pack for each channel for Turbo PMACs and accessories with this
feature. To enable the encoder-loss feature, pin 1 of the resistor pack (marked by a dot on the package)
should be placed at the opposite end of the socket from pin 1 of the socket (marked by a white-ink square
on the circuit board). For the 4-channel accessories (Acc-24E2x and 24C2x), the first four channels
shown are for accessories addressed as an even-numbered Servo IC (2, 4, 6, or 8); the second four
channels shown are for accessories addressed as an odd-numbered Servo IC (3, 5, 7, or 9).
Resistor Packs for Encoder Loss Circuitry
Ch. 3
Ch. 4
Ch. 5/1
Ch. 6/2
Ch. 7/3
Ch. 8/4
PMAC-PCI
RP60
RP62
RP66
PMAC2-PCI
RP43
RP48
RP44
QMAC
RP55
RP57
RP66
Acc-24P2
RP74
RP75
RP85
Acc-24E2
RP22
RP24
RP22*
Acc-24E2A
RP22
RP24
RP22*
Acc-24E2S
RP19
RP21
RP27
Acc-24C2A
RP33
RP34
RP63
* On the daughter board for the accessory module
Device
Ch. 1
Ch. 2
RP68
RP49
RP68
RP86
RP24*
RP24*
RP29
RP64
RP97
RP104
RP139
RP22
RP22
RP19
RP33
RP99
RP109
RP140
RP24
RP24
RP21
RP34
RP103
RP105
RP150
RP22*
RP22*
RP27
RP63
RP105
RP110
RP151
RP24*
RP24*
RP29
RP64
The following table shows the address of the encoder-loss status bit for each channel of each device. The
address is always for a Y-register. The x shown in some of the addresses represents the hex digit 8, 9, A,
or B, the same as the base address of the card itself. The bit value is 1 for a valid encoder signal; 0 to
signify encoder loss.
Device
Ch. 1
Ch. 2
Ch. 3
Ch. 4
Ch. 5/1
Ch. 6/2
Ch. 7/3
Ch. 8/4
PMAC-PCI
PMAC2-PCI
QMAC
Acc-24P2
Acc-24E2
Acc-24E2A
Acc-24E2S
Acc-24C2A
$70801,1
$78403,8
$78403,8
$7xF00,0
$7xF08,5
$7xF08,5
$7xF08,5
$7xF08,5
$70801,2
$78403,9
$78403,9
$7xF00,1
$7xF09,5
$7xF09,5
$7xF09,5
$7xF09,5
$70801,3
$78403,10
$78403,10
$7xF00,2
$7xF0A,5
$7xF0A,5
$7xF0A,5
$7xF0A,5
$70801,4
$78403,11
$78403,11
$7xF00,3
$7xF0B,5
$7xF0B,5
$7xF0B,5
$7xF0B,5
$70801,5
$78403,8
$7xF00,4
$7xF0C,5
$7xF0C,5
$7xF0C,5
$7xF0C,5
$70801,6
$78403,9
$7xF00,5
$7xF0D,5
$7xF0D,5
$7xF0D,5
$7xF0D,5
$70801,6
$78403,10
$7xF00,6
$7xF0E,5
$7xF0E,5
$7xF0E,5
$7xF0E,5
$70801,7
$78403,11
$7xF00,7
$7xF0F,5
$7xF0F,5
$7xF0F,5
$7xF0F,5
As of this writing, there is no automatic action taken on detection of encoder loss. Users who want to
take action on detecting encoder loss should write a PLC program to look for a change in the encoder loss
bit and take the appropriate action. Generally, the only appropriate response is to kill (open loop, zero
output, disabled) the motor with lost encoder feedback; other motors may be killed or aborted as well.
The following example shows how all motors can be killed on detection of the loss of signal for Encoder
1, used as feedback for Motor 1, on a Turbo PMAC.
#define Mtr1OpenLoop
M138
#define Enc1LossIn
M180
#define Mtr1EncLossStatus
P180
#define Lost
0
; Low-true fault
#define OK
1
Mtr1OpenLoop->Y:$0000B0,18,1 ; Standard definition
Enc1LossIn->Y:$078403,8,1
; CTRL0 input
OPEN PLC 18 CLEAR
; Logic to disable and set fault status
IF (Mtr1OpenLoop=0 AND Enc1LossIn=Lost)
; Closed loop, no enc
CMD^K
Mtr1EncLossStatus=1
ENDIF
214
Making your Application Safe
Turbo PMAC User Manual
; Logic to clear fault status
IF (Mtr1OpenLoop=0 AND Enc1LossIn=OK AND Mtr1EncLossStatus=0)
Mtr1EncLossStatus=0
ENDIF
CLOSE
Refer to the individual hardware reference manuals for more details of the implementation of this
function.
User-Written Safety Algorithms
You can write your own safety-checking algorithms easily in a PLC program. These algorithms are best
implemented in a compiled background PLC program, which scans at the same rate as Turbo PMAC’s
own built-in safety checks – once per background cycle. The above program of monitoring encoder loss
is a good example of this type of program.
Watchdog Timer
Turbo PMAC has an on-board watchdog timer. This subsystem provides a fail-safe shutdown to guard
against software and hardware malfunction. To keep it from tripping the hardware circuit for the
watchdog timer requires that two basic conditions be met. First, it must see a DC voltage greater than
approximately 4.75V. If the supply voltage is below this value, the circuit’s relay will trip and the card
will shut down. This prevents corruption of registers due to insufficient voltage.
The second necessary condition is that the timer must see a square wave input (provided by the Turbo
PMAC software) of a frequency greater than approximately 25 Hz. In the foreground, the servo-interrupt
routine decrements a counter (as long as the counter is greater than zero), causing the least significant bit of
the timer to toggle. This bit is fed to the timer itself. At the end of each background cycle, the CPU resets
the counter value to a maximum value set by variable I40 (or to 4096 if I40 is set to the default of 0).
If the card, for whatever reason, due either to hardware or software problems, cannot set and clear this bit
repeatedly at 25 Hz or greater, the timer will trip and the Turbo PMAC system will shut down.
Actions on Watchdog Timer Trip
When the timer trips due to either under-voltage or under-frequency, the system is latched into a reset
state, with a red LED indicating watchdog failure. The processor stops operating and will not
communicate. All Servo, MACRO, and I/O ICs are forced into their reset states, which force discrete
outputs off, and proportional outputs (DAC, PWM, PFM) to zero-level.
Turbo PMAC systems have discrete outputs indicating the state of the watchdog timer. On Turbo PMAC
boards, there is a solid-state open-collector output called “FEFCO/” that is turned on when the timer trips.
In Turbo PMAC2 systems there is a hard-contact relay with both normally open and normally closed
contacts. In a system, these outputs should be used to drop power to the amplifiers and other key circuitry
if the card fails.
Once the watchdog timer has tripped, power to the Turbo PMAC must be cycled off and on, or the INIT/
hardware reset line must be taken low, then high, to restore normal functioning.
Diagnosing Cause of Watchdog Timer Trip
Because the watchdog timer is designed to trip on a variety of hardware and software failures, and the trip
makes it impossible to query the card, it can be difficult to determine the cause of the trip. The following
procedure is recommended to figure out the cause:
1. Reset the Turbo PMAC normally (with the re-initialization jumper OFF). If it does not trip again
immediately, there is an intermittent software or hardware problem. Check for the following:
• Software events that overload the processor at times (e.g. additional servo-interrupt tasks,
intensive lookahead) or possible erroneous instruction (look for firmware or program checksum).
Review the Evaluating the Turbo PMAC’s Computational Load section of this manual.
Making your Application Safe
215
Turbo PMAC User Manual
•
•
5V power-supply disturbances
Loose connections
2. If there is an immediate watchdog timer trip in Step 1, power up with the re-initialization jumper ON.
If it does not trip now, there is a problem in the servo/phase task loading for the frequency, or an
immediate software problem on the board. Check for the following:
• Phase and servo clock frequencies vs. the number of motors used by Turbo PMAC. These
frequencies may need to be reduced.
• A PLC 0 or PLCC 0 program running immediately on power-up (I5 saved at 1 or 3) and taking
too much time.
• User-written servo or phase program not returning properly.
3. If there is an immediate watchdog timer trip in Step 2, check for hardware issues:
• Correct settings of clock-frequency jumpers (Turbo PMAC)
• Jumpers set for external servo and phase clock, but none provided
• Disconnect any accessories and repeat to see if they are causing the problem
• Check for adequate 5V power supply levels (check at the Turbo PMAC CPU, not at the supply)
• Inspect for hardware damage
4. If nothing is found in Step 3, power up with the firmware reload (bootstrap) jumper ON. If there is
no watchdog timer trip here and you can do basic communications (??? should cause a BOOTSTRAP
ROM reply), there is a problem with your operational firmware and it must be reloaded.
Hardware Stop Command Inputs
Turbo PMAC controllers have hardware inputs that can stop a move or a program with user-set
decelerations. The Abort input stops motion of all axes in the selected coordinate systems, as determined
by the motor/system select inputs, starting immediately, and with each motor decelerating at a rate set by
Ix15. The Hold input performs the same function, except that the axes are decelerated at rates such that
the desired multi-axis path is maintained during deceleration.
These dedicated inputs are on Turbo PMAC’s control panel connector (JPAN; J2). Which coordinate
system they act on is determined by the binary number produced by the four low-true input lines FPD0/
(LSBit), FPD1/, FPD2/, and FPD3/ (MSBit). A value of zero (all high) disables the functions; values of 1
through 8 select the numbered coordinate system.
Host-Generated Stop Commands
These functions and several others can also be performed from the host with one- or two-character
commands. For instance, <CTRL-A> performs the same function as the Abort input with all coordinate
systems selected, and A aborts the software-addressed coordinate system. <CTRL-O> holds all
coordinate systems, and H holds the software-addressed coordinate system. In addition <CTRL-Q> stops
all programs at the end of the upcoming move, and Q stops the program of the software-addressed
coordinate system. <CTRL-K> disables all motors immediately and K disables the software-addressed
motor (if the motor is in a coordinate system that is running a motion program, an Abort command should
be issued before the K command.
Any of these commands may be issued from within a Turbo PMAC program, using the
COMMAND"{command}" or the COMMAND^{letter} syntax. However, a motor-kill (K) command
for a motor in the coordinate system will be rejected automatically when issued from within a motion
program running in that coordinate system.
216
Making your Application Safe
Turbo PMAC User Manual
The following table summarizes the different commands that can be used to stop motion and their
attributes:
Command
Scope
Begin
Immed?
Stop on
Path?
Stop at
Prog Pt?
Decel
Rate
J/
Motor
Yes
–-
No
A
<CTRL-A>
Q
<CTRL-Q>
/
C.S.
Global
C.S.
Global
C.S.
Yes
Yes
No
No
No
No
No
Yes
Yes
Yes
No
No
Yes
Yes
Yes
Ixx1921
Ixx15
Ixx15
TA, TS
TA, TS
TA, TS
H
<CTRL-O>
\
C.S.
Global
C.S.
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
Isx95
Isx95
Ixx17/
Isx95
K
<CTRL-K>
Motor
Global
Yes
Yes
No
No
No
No
—
—
Notes
No easy restart
No easy restart
Finishes calculated moves
Finishes calculated moves
End of currently executing
move*, undoes blending
Similar to %0
Similar to %0
Acts as H command if
outside of lookahead
Must stop program first
Must stop program first
Program Checksums
Firmware Checksum
Turbo PMAC continually computes the checksum of its internal program (firmware) as a background
task. Each time it has computed the checksum, it compares this value to a reference register in memory
(obtainable with the CHECKSUM command) that has been manually entered with the correct value. Turbo
PMACs shipped from the factory are preloaded with the correct reference value for that firmware version
at the factory.
If Turbo PMAC detects a mismatch between its calculated checksum and the reference checksum, it sets
global status bits (bits 12 and 13 of X:$000006 – accessible with the ??? command) and stops
performing any checksum operations. This leaves the calculated value frozen in the running checksum
register X:$001080. Turbo PMAC does take any other action in the event of a firmware checksum error;
it is up to the host or a Turbo PMAC PLC program to decide what action to take.
When a Turbo PMAC is upgraded to new firmware by replacement of the PROM IC in standard CPU
sections or downloading of new firmware into the flash EEPROM IC in Option CPU sections, the
reference checksum value will be updated automatically for the new firmware.
User-Program Checksum
Turbo PMAC continually computes the checksum of the fixed user program buffers as a background task.
Each time it has computed the checksum, it compares this value to the checksum value that was computed
the last time one of these buffers was closed, stored in X:$001090.
If Turbo PMAC detects a mismatch between these two checksums, it sets a global status bit (bit 13 of
X:$000006 – accessible with ???) and stops performing any program or firmware checksum operations
(communications checksum is independent), freezing the running checksum value in X:$001080. It does
not shut down operation automatically. It is up to the host or a Turbo PMAC PLC program to decide
what action to take if there is a checksum error.
Making your Application Safe
217
Turbo PMAC User Manual
Communications Data Integrity
Turbo PMAC provides a variety of techniques for ensuring valid transmission of data, including serial
parity checking, framing error checking, serial full-duplex communications, and bidirectional checksum
computation on both serial and bus communications. For more details on how these techniques work,
refer to the Writing Host Communications Programs section of this manual.
218
Making your Application Safe
Turbo PMAC User Manual
EXECUTING INDIVIDUAL MOTOR MOVES
Once the motor is defined and basically working with reasonable gains, command some basic moves for
the motor. Jogging commands make simple moves for the motor, independent of other motors, without
writing a motion program. Use these moves for development, diagnostics, and debugging, but they may
also be used in an actual application.
Another type of simple motor move is the homing search move. This is basically a jog-until-trigger type
of move, where Turbo PMAC commands the motor to move until it sees a pre-defined trigger. It then
brings the motor to a stop and returns to the trigger position (possibly with an offset), and sets the motor
position to zero.
A homing search move should be performed when it is not known where home position is. If there is an
incremental position sensor, and it is not known the location on power-up; therefore, typically the homing
search move is the first move done in this type of system. However, if the home position known, but just
wish to return to that position, there is no need to do a homing search move; simply command a move to
the zero position (e.g. J=0 or X0)
The trajectories for jogging and homing moves are the same as for rapid-mode program moves.
However, these moves are specified directly to the motor, specified by number, rather than the axis,
specified by letter. The moves are described in unscaled units (all based on counts and milliseconds).
Jogging Move Control
The velocity and acceleration for jogging moves is controlled by I-variables for the motor jogged. The
destination is controlled by jog commands. These are described below.
Jog Acceleration
Variable Ixx20 for Motor xx specifies the acceleration time for jogging, homing, and programmed
RAPID-mode moves, in milliseconds. Ixx21 specifies the time in each half of the S-curve acceleration
profile for these moves, also in milliseconds. If Ixx20 is less than two times Ixx21, the acceleration time
used will be twice Ixx21.
The acceleration rate limit for jog/home/RAPID moves is set by Ixx19 (in counts/msec2). If Ixx20 and
Ixx21 are so small that Ixx19 would be exceeded, the acceleration time is extended so that Ixx19 is not
exceeded. To specify the acceleration by rate instead of time, set the acceleration time parameters small
enough that the limiting acceleration rate parameter is always used.
Note:
To specify the acceleration by rate, do not set both acceleration time parameters
Ixx20 and Ixx21 to zero. This will cause a division-by-zero error in the move
calculations that could cause erratic movement. The minimum acceleration time
setting should be Ixx20=1 and Ixx21=0.
Jog Speed
Jogging speed is specified by Ixx22, which is a magnitude of the velocity, in counts per millisecond.
Direction is specified by the jog command itself.
Executing Individual Motor Moves
219
Turbo PMAC User Manual
MOTOR x MOTION VARIABLES
Ix20 ACCELERATION TIME (JOG, HOME)
(Units: msec); integer
Ix21 S-CURVE TIME (JOG, HOME)
(Units: msec); integer
V
Ix20 > 2 * Ix21
Ix21
Ix21
Ix21
Ix20
V
Ix21
T
Ix20
Ix20 < 2 * Ix21
Ix21
Ix21
Ix21
2*Ix21
Ix21
T
2*Ix21
V
Ix21=0
Ix20
Ix20
T
Jog Move Trajectory
Move Timer Active = 0
(if command was J+ or J-
Jog Stop given (J/)
Move Timer Active = 1
Vel
Move Timer Active = 1
Desired Velocity Zero = 0
In Position = 0
Max Accel = Ix19
Max Accel = Ix19
Ix22
Desired Velocity Zero = 1
Running a Program = 0
In Position = 1
(when FE in range)
Time
Ix21
Jog Start
given (J+)
Ix21
I12
[Ixx90]
Note: if the jog command is not of
the indefinite length (not J+ or J-) then
the “Running a Program” status bit will
be high for the entire jog move
220
Ix21
Ix20
Ix21
Ix20
I12
[Ixx90]
Note: Rate of acceleration
limited by Ix19 - can override
Ix20 and Ix21
Executing Individual Motor Moves
Turbo PMAC User Manual
Jog Commands
The commands to jog a motor are on-line (immediate) commands that are motor-specific; they act on the
currently addressed motor.
Note:
A jog command to a motor will be rejected if the motor is in a coordinate system
that is currently executing a motion program, even if the motion program is not
commanding that motor to move. PMAC will report ERR001 if I6 is set to 1 or 3.
Indefinite Jog Commands
J+ commands an indefinite positive jog for the addressed motor; J- commands an indefinite negative
jog; J/ commands an end to the jog, leaving the motor in position control after the deceleration. It is
possible for the J/ command to leave the commanded position at a fractional count, which can cause
dithering between the adjacent integer count values. If this is a problem, the J! command can be used to
force the commanded position to the nearest integer count value. (Alternatively, ½-count of deadband
created with Ixx64 and Ixx65 can prevent dithering at fractional count values.
Jogging to a Specified Position
Jog commands to a specified position, or of a specified distance, can be given. J= commands a jog to the
last pre-jog position; J={constant} commands a jog to the (unscaled) position specified in the
command; J=={constant} commands a jog to the (unscaled) position specified in the command and
makes that position the pre-jog position; J^{constant} commands a jog of the specified distance from
the actual position at the time of the command (J^0 can be useful to take up remaining following error);
J:{constant} commands a jog of the specified distance from the commanded position at the time of
the command.
Jog Moves Specified by a Variable
Jogging moves to a position or of a distance specified by a variable are possible. Each motor has a
specific register (L:$0000D7 for Motor 1, L:$000157 for Motor 2, etc., suggested M-variables Mxx72)
that holds the position or distance to move on the next variable jog command. This register contains a
floating-point value scaled in encoder counts. It should be accessed with an L-format (floating-point) Mvariable. The J=* command causes PMAC to use this value as a destination position. The J^*
command causes PMAC to use the value as a distance from the actual position at the time of the
command. The J:* command causes PMAC to use the value as a distance from the commanded position
at the time of the command.
Issuing Commands During Jog Moves
It is permissible to issue a jog command to a motor that is already jogging. On receipt of the new
command, Turbo PMAC will break into the already planned trajectory and create a smooth blend to the
trajectory of the new command as determined by the present acceleration and velocity commands. The
existing trajectory is extended out for Ixx92 milliseconds (default 10 msec) after receipt of the new
command; during this time the new trajectory is calculated. The calculations for the new move must be
able to complete within Ixx92 msec.
Parameters Used by Jog Commands
Each time one of these commands is given, the acceleration and velocity parameters at that time control
the response to the command. To change speed or acceleration parameters of an active jog move, change
the appropriate parameters, and then issue another jog command.
Executing Individual Motor Moves
221
Turbo PMAC User Manual
Triggered Motor Moves
Triggered moves in Turbo PMAC are double moves, with a pre-trigger portion and a post-trigger portion.
Upon the trigger event, Turbo PMAC will break into the pre-trigger move and calculate a post-trigger
move ending at a pre-specified distance from the trigger point.
Types of Triggered Moves
There are three types of triggered motor moves:
1. Homing search moves (on-line or motion-program)
2. On-line jog-until-trigger moves
3. Motion-program RAPID move-until trigger
These moves all work basically in the same manner, just in different contexts. Triggering and positioncapture functions are the same in all three types of moves. Each will be described in detail below.
Types of Trigger Conditions
There are fundamentally two types of triggers for these triggered moves: input triggers and followingerror triggers. Bit 1 of motor variable Ixx97 determines which of these is used for the motor’s triggered
moves.
Input Triggering
If bit 1 of Ixx97 is 0 (Ixx97 = 0 or 1), the trigger condition for the motor is an input trigger. In this, the
default case, the trigger is caused by one or two of the inputs mapped into the register whose address is
specified by Ixx25. Usually, Ixx25 contains the address of a flag register in a Servo IC, or a register in
memory containing the contents of flag data copied in from the MACRO ring. For the super-accurate
hardware-capture (see below), Ixx25 must specify the flags of the same hardware channel as the encoder
used for position-loop feedback. If bit 18 of Ixx24 is set to 1, Turbo PMAC expects the capture trigger
and position to come from over the MACRO ring; if bit 18 of Ixx24 is 0, Turbo PMAC expects these
from hardware channels directly connected to the controller’s CPU.
Trigger Signal(s) and Edge(s)
Two setup variables for the Servo IC channel determine which edges of which signals will cause the
trigger. I7mn2 for Servo IC m Channel n (node-specific variable MI912 on a MACRO Station) specifies
whether the index channel is used or not, whether an input flag is used or not, and which edges of the
index and/or flag will be cause the trigger. If both index and flag are selected, the two signals are
combined with Boolean hardware logic inside the IC to create the trigger.
If I7mn2 specifies the use of a flag, I7mn3 (node-specific variable MI913 on a MACRO Station) selects
which of the four input flags for the channel is used. For a PMAC-style Servo IC, the choices are HMFL,
+LIM, -LIM, and FAULT; for a PMAC2-style Servo IC, the choices are HOME, MLIM, PLIM, and
USER.
Edge vs. Level Triggering
There is a subtle difference between PMAC-style and PMAC2-style Servo ICs that lead to different
modes of failure if the trigger move starts with the inputs already in the trigger state (there is no
difference otherwise). PMAC-style Servo ICs are edge triggered, requiring a transition (edge) from nontrigger state to trigger state to generate the actual trigger. With these ICs, there will be no trigger
generated in this case, and the pre-trigger move will continue, perhaps indefinitely.
However, PMAC2-style Servo ICs are level triggered only requiring that the IC see the input(s) in the
trigger state to generate the actual trigger. With these ICs, a trigger will be generated immediately in this
case, and there will be little or no motion at all.
222
Executing Individual Motor Moves
Turbo PMAC User Manual
Merits of Dual Trigger
In homing-search moves, it is common practice to use a combination of a homing switch and the index
channel as the home trigger condition. The index channel of an encoder, while precise and repeatable, is
not unique in most applications, because the motor can travel more than one revolution. The homing
switch, while unique, is typically not extremely precise or repeatable. By using a logical combination of
the two, you can get uniqueness from the switch, and precision and repeatability from the index channel.
In this scheme, the homing switch is used to select which index channel pulse is used as the home trigger.
Although the homing switch does not need to be placed extremely accurately in this type of application, it
is important that its triggering edge remain safely between the same two index channel pulses. Also, the
homing switch pulse must be wide enough to always contain at least one index channel pulse.
Following-Error Triggering
Sometimes it is desired that a trigger occur when an obstruction such as a hard stop is encountered, as
when using an end stop for the homing reference. To support this type of functionality, Turbo PMAC
permits triggering on a warning following error condition instead of an input flag. This is sometimes
called “torque-mode” triggering, because it effectively triggers on a torque level (except for velocitymode amplifiers) because output torque command is proportional to following error. It is also called a
“torque-limited” mode, because it provides an easy way to create moves that are limited in torque, and
that stop when the torque limit is reached (as in torque-limited screw driving).
To enable this torque-mode triggering for Motor xx, set Ixx97 to 3, specifying both following-error
trigger and software capture (there is no hardware signal to create a hardware capture). In this mode, the
trigger for a move-until-trigger is a true state of the warning following-error status bit for the motor.
Variable Ixx12 sets the warning following-error magnitude for the motor, with units of 1/16 of a count.
When Turbo PMAC detects that the magnitude of the following error has exceeded this value, it will read
the present feedback position as the trigger position, then move relative to this position.
When using torque-mode triggering, it is a good idea to set the integral gain term Ixx33 to 0 to prevent a
large “charge-up” of the integrator when it hits the hard stop. It may also be desirable to set the Ixx69
output limit lower to limit the possible torque directly when the obstruction is reached.
Note that if the warning following error status bit is true at the start of the move, the trigger will occur
almost immediately.
Capturing the Trigger Position
Because the post-trigger move ends at a commanded position expressed relative to the position at the time
of the trigger, it is necessary for Turbo PMAC to “capture” the position at the time of the trigger.
Fundamentally, there are two ways of doing this: hardware capture and software capture.
Hardware Capture
The Servo ICs of a Turbo PMAC have dedicated registers to latch the encoder counter instantly upon
receipt of the pre-specified input trigger state. The latching action occurs entirely in the IC hardware,
requiring no software action, so the captured position is accurate to the exact count regardless of motor
speed. This means that there is no need to slow down the move to get an accurate capture.
Hardware capture is selected for Motor xx trigger moves by setting Ixx97 to the default value of 0,
specifying both hardware capture and input triggering. If hardware capture is selected, the position-loop
feedback as selected by Ixx03 must come through the encoder counter of a Servo IC. It must use the
same hardware channel as the flag set selected by Ixx25. This means that if you are using dual feedback
on the motor, the flag set specified by Ixx25 should be the same channel as your position-loop feedback,
not your velocity-loop feedback.
Executing Individual Motor Moves
223
Turbo PMAC User Manual
Hardware Capture with Acc-51 Interpolators
To utilize the hardware-capture feature on triggered moves using sinusoidal encoder feedback through an
Acc-51 high-resolution interpolator, several additional firmware features (introduced in firmware revision
V1.940) must be activated. Because the capture flags must be of the same Servo IC and channel as the
position loop feedback, Ixx25 must be set to the address of the channel on the interpolator board.
However, the interpolator board does not provide amplifier-enable and fault flags, or overtravel limit
flags, so variables Ixx42 and Ixx43, respectively, are used to specify different addresses for these flags
(probably on the axis card).
Next, because the Acc-51 produces 10 bits of fractional count data for each increment of the hardware
counter in the Servo IC, instead of the usual 5 bits, bit 11 of Ixx24 must be set to 1 for the captured
hardware count value to be scaled properly when converted to motor position.
Finally, starting with Revision D of the PMAC2-style “DSPGATE1” Servo IC (introduced early in 2002)
used on the Acc-51E and Acc-51C interpolator boards, hardware capture with sub-count resolution
became possible. Note that this capability is not required for hardware capture using the interpolator
board, and probably should not be used for homing. To activate this feature for the channel in the Servo
IC, variable I7mn9 must be set to 1. To use this fractional data in a move-until-trigger on Motor xx, bit
12 of Ixx24 must be set to 1.
Software Capture
If use of hardware capture for trigger position is not possible for some reason, software capture is
possible. Software capture is specified for Motor xx by setting Ixx97 to 1 (with input trigger) or 3 (error
trigger). If software capture has been selected, Turbo PMAC software uses the most recent servo cycle’s
motor actual position as the trigger position, regardless of the source, when the software notices that the
trigger has occurred.
When software capture is used, there is a potential delay between the actual trigger and Turbo PMAC’s
position capture of one background cycle, which could be several milliseconds. This delay can lead to
inaccuracies in the captured position; the speed of the motor at the time of the trigger must be kept low
enough to achieve an accurate enough capture. For homing, a two-step procedure can be used: a fast,
inaccurate capture followed by a slow, accurate capture.
Post-Trigger Move
On detection of the trigger, Turbo PMAC will break into the pre-trigger move trajectory and create a
smooth blend to the trajectory of the post-trigger move as determined by the present acceleration and
velocity commands. The post-trigger move ends at a pre-determined distance from the captured position
– the method for specifying this distance depends on the type of move (see in each section below). The
existing trajectory is extended out for Ixx92 milliseconds (default 10 msec) after detection of the trigger;
during this time the new trajectory is calculated. The calculations for the post-trigger move must be able
to complete within Ixx92 msec.
The blending from the pre-trigger to the post-trigger move, and from the post-trigger move to a stop, will
use the acceleration parameters Ixx19, Ixx20, and Ixx21 in force when the trigger is found. The
magnitude of the velocity of the post-trigger move will be the same as the pre-trigger move, if the posttrigger move is long enough to reach that velocity. Usually, the post-trigger move will involve a reversal
from the pre-trigger move, but this is not necessarily the case.
Homing Search Moves
The purpose of a homing search move is to establish an absolute position reference when an incremental
position feedback sensor is used. The move until trigger construct is ideal for finding the sensor that
establishes the home position and automatically returning to this position.
224
Executing Individual Motor Moves
Turbo PMAC User Manual
Homing Acceleration
The acceleration for homing search moves is controlled by the same parameters – Ixx19 (maximum
acceleration), Ixx20 (acceleration time), and Ixx21 (S-curve time) – as for jogging moves. These are
described in the above section on jogging moves.
Homing Speed
Ixx23 specifies the speed and direction of the homing-search move. If Ixx23 is greater than zero, the pretrigger homing-search move will be in positive direction. If it is less than zero, the pre-trigger move will
be in the negative direction. The magnitude of Ixx23 (expressed in counts/msec) controls the speed of
both the pre-trigger and post-trigger moves (should they be long enough to get to this speed).
Home Trigger Condition
The trigger condition for homing-search moves, as for other triggered moves, is specified by Ixx97,
Ixx24, and Ixx25, as described above in detail. If no trigger is found, the pre-trigger move will continue
indefinitely, or until stopped by an error condition such as hitting overtravel limits.
Post-Trigger Move
Variable Ixx26 specifies the (signed) distance from the trigger-captured position to the end of the posttrigger move. The units of Ixx26 are 1/16 count. The endpoint of the commanded post-trigger move is
the new motor position zero (the motor’s “home” position). The change of the motor’s reported position
reference occurs at the beginning of the post-trigger move. As soon as this is done, reported positions are
referenced to this new zero position (plus or minus any axis offset in the axis definition statement – if the
axis definition is #1->10000X+3000, the home position will be reported as 3000 counts). Also at this
point, the motor’s “home search in progress” status bit is cleared, and the “home complete” status bit is
set.
If the post-trigger move fails with an error condition, it is not necessary to re-home the motor, as the
home position is already known. A command such as J=0 can move the motor to the home position once
the source of the problem has been cleared up.
If software overtravel limits are used (Ixx13, Ixx14 not equal to zero), they are re-enabled at this time
after having been automatically disabled during the search for the trigger. The trajectory to this new zero
position is then calculated, including deceleration and reversal if necessary. Note that if a software limit
is too close to zero, the motor may not be able to stop and reverse before it hits the limit. In normal
termination, the motor will stop under position control with its commanded position equal to the home
position. If there is a following error, the actual position will be different by the amount of the following
error.
Failure to Find Trigger
The pre-trigger move of a homing search will continue indefinitely if it fails to find the trigger condition it
is looking for. Typically, it will be stopped by an overtravel position limit switch or fatal following error
limit at the end of travel in this case. If you want a programmed limit to the length of the pre-trigger
move, you should use an incremental jog-until-trigger command or programmed move-until-trigger, with
the first value specifying the distance to move in the absence of a trigger, and the second the distance
from the trigger to the end of the post-trigger move (replacing Ixx26). Once the post-trigger move is
finished, a HOMEZ command (see below) can be used to set this position to be the motor zero position.
Executing Individual Motor Moves
225
Turbo PMAC User Manual
Homing Search Move Trajectory
Vel
Trigger
Occurs
Home Complete=1
Home Search in Progress=0
Home Complete=0
Home Search
In Progress=1
Net distance from
trigger position
= Ix26
Ix23
Time
Ix21
Ix21
Ix21
Ix20
Note: Rate of acceleration
limited by Ix19 - can override
Ix20 and Ix21
Desired Velocity Zero=1
In Position=1
(when FE in range)
Ix21
Ix20
Ix21
Ix21
Ix20
Home Command
The homing search move can be executed either through an on-line command (which can be given from a
PLC program using the COMMAND"" syntax) or a motion program statement.
On-Line Command
A homing search move can be initiated with the on-line motor-specific command HOME (short form HM),
for example #1HM. This is simply a command to start the homing search; Turbo PMAC provides no
automatic indication that the move is completed, unless set up to recognize the in-position (IPOS)
interrupt.
Monitoring for Finish: If monitoring the motor from the host or from a PLC program to see if it has
finished the homing move, it is best to look at the home complete and desired velocity zero motor status
bits, accessed either with the ? command or with M-variables. The home complete bit is set to zero on
power-up/reset; it is also set to zero at the beginning of a homing search move, even if a previous homing
search move was completed successfully. It is set to one as soon as the trigger is found in a homing
search move, before the motor has come to a stop.
Note:
The home search in progress bit is the inverse of the home complete bit during the
move: it is 1 until the trigger is found, then 0 immediately after. Therefore the
monitoring should also look for the desired velocity zero status bit to become 1,
which will indicate the end of the post-trigger move.
Monitoring for Errors: A robust monitoring algorithm will also look for the possibility that the homing
search move could end in an error condition. Often this is just part of the general error monitoring that is
done at all times, looking for overtravel limits, fatal following errors, and amplifier faults. If an error
does occur during the homing move, it is important to distinguish between one that occurs before the
trigger has been found, and one that occurs after. If the error occurs after, Turbo PMAC knows where the
home position is, and the homing search does not need to be repeated. Once the error cause has been
fixed, the motor can simply be moved to the home position with a command such as J=0.
Buffered Motion-Program Command
The homing search move also can be commanded from within a motion program with the HOMEn
command, where n is the motor number. Note that this command specifies a motor; unlike other motion
program commands that specify an axis move. In a motion program, Turbo PMAC’s automatic program
sequencing routines monitor for the end of the move. When the move is successfully completed, program
execution continues with the next command.
226
Executing Individual Motor Moves
Turbo PMAC User Manual
Multiple homing moves can be started together by specifying a list or range of motor numbers with the
command (e.g. HOME1,3 or HOME2..6). Further program execution will wait for all of these motors to
finish their homing moves. Separate homing commands, even on the same line (e.g. HOME1 HOME2)
will be executed in sequence, with the first finishing before the second starts. It is not possible to execute
partially overlapping homing moves from a single motion program.
Note:
Note carefully the difference in syntax between the on-line command and the
buffered command. The on-line command is simply HOME or HM, and it acts on
the currently addressed motor, so the motor number must be specified in front of
the command (e.g. #1HM). In the buffered command, the motor number is part of
the command, following immediately after HOME or HM letters (e.g. HM1).
Homing from a PLC Program
Turbo PMAC PLC programs can command homing search moves by giving on-line commands with the
COMMAND"" statement (e.g. COMMAND"#1HM"). These commands simply start the homing search
move; code must be written to monitor for finishing if that is desired. The motor number must be
specified in the specific command string, or with the ADDRESS#n statement; without this statement,
motor addressing is not modal within PLC programs.
Motion vs. PLC Program Homing
The following table summarizes the differences between homing using motion programs and PLC
programs.
Motion Programs
PLC Programs
Program execution point stays on the line
containing the Home command until the
homing move is finished.
Home command can be combined with
programmed axis moves.
The C.S. must be ready to run a motion
program.
Can only home motors defined in the C.S.
running the program.
Motors can be homed simultaneously, one
after another, or any combination of the
two. All motors started together must
finish before next action can start.
The motion program must be started by an
on-line command, a PLC program, or
another motion program.
The PLC does not automatically monitor
for the start and end of the homing move.
Axis motion can only be performed
through Jog commands. .
The C.S. does not need to be ready to run
a motion program.
Can home any motor not defined in a C.S.
presently running a program.
Motors can be homed in any order. This
includes starting one motor in the middle
of another motor's home move.
The PLC can be started by an on-line
command, a PLC program, another
motion program, or automatically at
power-up or reset.
Zero-Move Homing
To declare the current position the home position without commanding any movement, use the HOMEZ
(on-line) or HOMEZn (motion program) command. These are like the HOME command, except that they
immediately take the current commanded position as the home position. The Ixx26 offset is not used
with the HOMEZ command. This no-move homing is useful both for early development, before the true
homing procedure is developed, and for specialized homing routines not supported by Turbo PMAC’s
automatic homing procedures.
Executing Individual Motor Moves
227
Turbo PMAC User Manual
Note:
If there is a following error when the HOMEZ command is given, the reported
actual position after the HOMEZ command will not be exactly zero; it will be equal
to the negative of the following error.
Homing Into a Limit Switch
It is possible to use a limit switch as a home switch. However, first disable the limit function of the limit
switch for the move to finish normally; if this is not done, the limit function will abort the homing search
move. Even so, the home position has been set; a J=0 command can then be used to move the motor to
the home position.
Note:
On PMAC-style servo channels, the polarity of the limit switches is the opposite of
what many people would expect. The -LIMn input should be connected to the
limit switch at the positive end of travel; the +LIMn input should be connected to
the limit switch at the negative end of travel.
To disable the limit function of the switch, set bit 17 of variable Ixx24 (bit value $20000) for the motor to
1. It is a good idea to use the home offset parameter Ixx26 to bring the home position out of the limit
switch, so the limits can be re-enabled immediately after the homing search move, without being in the
limit.
The following examples show two quick routines to do this type of homing. One uses a motion program
and the other a PLC program. The same function could also be done with on-line commands.
;*********** Motion Program Set-up Variables (to be saved) *************
CLOSE
I123=-10
I124=$000000
I125=$78000
I126=32000
I7102=3
I7103=2
; Home speed 10 cts/msec negative
; PMAC-style flags, normal mode
; Use Servo IC 0 Channel 1 flags for Motor 1
; Home offset of +2000 counts
; (enough to take you out of the limit)
; Capture on rising flag and rising index
; Use +LIM1 as flag (negative end switch)
;*********** Motion program to execute routine *********************
OPEN PROG 101 CLEAR
I124=$20000
HOME1
I124=$0
CLOSE
; Disable +/-LIM as limits
; Home #1 into limit and offset out of it
; Re-enable +/-LIM as limits
; End of program
;*********** PLC Set-up Variables (to be saved) ************************
CLOSE
I123=-10
I124=$000000
I125=$78000
I126=32000
I902=3
I903=2
; Home speed 10 cts/msec negative
; PMAC-style flags, normal mode
; Use Servo IC 0 Channel 1 flags for Motor 1
; Home offset of +2000 counts
; (enough to take you out of the limit)
; Capture on rising flag and rising index
; Use +LIM1 as flag (negative end switch)
M133->X:$B0,13,1
M145->Y:$C0,10,1
; Desired Velocity Zero bit
; Home complete bit
228
Executing Individual Motor Moves
Turbo PMAC User Manual
;*********** PLC program to execute routine *********************
OPEN PLC 10 CLEAR
I124=$20000
CMD"#1HM"
WHILE (M145=1)
ENDWHILE
WHILE (M133=0)
ENDWHILE
I124=$0
DIS PLC10
CLOSE
; Disable +/-LIM as limits
; Home #1 into limit and offset out of it
; Waits for Home Search to start
; Waits for Home motion to complete
; Re-enable +/-LIM as limits
; Disables PLC once Home is found
; End of PLC
Multi-Step Homing Procedures
You may require a homing procedure that cannot be executed with a single Turbo PMAC homing move.
In this case, you will use two (or possibly more) homing search moves, changing the move parameters in
between. Although this can be done with a sequence of on-line commands, it is probably easier to create
a small motion program to execute the sequence.
Which Direction to Home? The most common of these situations is the case in which it is not known on
which side of the home trigger it is at power-up. In this case, move into one of the limit switches to make
sure it is at one end of travel (this can be done by homing into the limit, much as in the above example).
Then do a homing move the other direction into the real home trigger. A sample motion program routine
that does this is:
CLOSE OPEN PROG 102 CLEAR
I223=10
; Home speed 10 cts/msec positive direction
I224=$20000
; Disable hardware limits
I225=$78208
; Servo IC 2 Channel 2 for flags
I226=0
; No home offset
I7222=2
; Capture on rising edge of a flag
I7223=1
; Use PLIM2 as flag (positive end limit)
HOME2
; Home into limit
I223=-10
; Home speed 10 cts/msec negative direction
I224=$0
; Re-enable hardware limits
I7222=11
; Capture on flag low and index channel high
I7223=0
; Use HOME2 (home flag) as trigger flag
HOME2
; Do actual homing move
CLOSE
A sample PLC Program routine that does this is:
CLOSE
M233->X:$130,13,1
M245->Y:$140,10,1
OPEN PLC 11 CLEAR
I223=10
I224=$20000
I225=$78208
I226=0
I7222=2
I7223=1
CMD"#2HM"
WHILE (M245=1)
ENDWHILE
WHILE (M233=0)
; #2 Desired-velocity-zero bit
; #2 Home complete bit
; Home speed 10 cts/msec positive direction
; Disable hardware limits
; Servo IC 2 Channel 2 for flags
; No home offset
; Capture on rising edge of a flag
; Use PLIM2 as flag (positive end limit)
; Home into limit
; Waits for Home Search to start
; Waits for Home motion to complete
Executing Individual Motor Moves
229
Turbo PMAC User Manual
ENDWHILE
I223=-10
I224=$0
I7222=11
I7223=0
CMD"#2HM"
WHILE (M245=1)
ENDWHILE
WHILE (M233=0)
ENDWHILE
DIS PLC11
CLOSE
; Home speed 10 cts/msec negative direction
; Re-enable hardware limits
; Capture on flag low and index channel high
; Use HOME2 (home flag) as trigger flag
; Do actual homing move
; Waits for Home Search to start
; Waits for Home motion to complete
; Disables PLC once Home is found
; End of PLC
Already Into Home? A similar situation occurs when it is not known on power-up whether or not it is
already into the home trigger. Here, the easiest solution is to write a program that evaluates this
condition; if it is in the trigger, it moves out before doing the real homing.
;*************** Motion Program Set-up variables (to be saved) *********
CLOSE
M320->X:$078210,20,1
I325=$078210
; Variable for Servo IC 2 Ch. 3 home input
; Use Flags3 for Motor 3
;************** Motion program to execute routine *********************
OPEN PROG 103 CLEAR
IF (M320=1)
I323=10
I326=1600
I7232=11
I7233=0
HOME3
ENDIF
I323=-10
I326=0
I7232=3
I7233=0
HOME3
CLOSE
; Already in trigger?
; Home speed 10 cts/msec positive direction
; Home offset +100 counts (to make sure clear)
; Capture on falling flag and rising index
; Use Home3 as flag
; "Home" out of switch
; Home speed 10 cts/msec negative direction
; No home offset
; Capture on rising flag and rising index
; Use HMFL3 as flag
; Do actual homing move
; End of program
;***************PLC Set-up variables (to be saved) ********************
CLOSE
M320->X:$078210,20,1
I325=$078210
M333->X:$01B0,13,1
M345->Y:$01C0,10,1
; Variable for Servo IC 2 Ch. 3 home input
; Use Flags3 for Motor 3
; Desired Velocity Zero bit
; Home complete bit
;**************** PLC program to execute routine *********************
OPEN PLC 12 CLEAR
IF (M320=1)
I323=10
I326=1600
I7232=11
I7233=0
CMD"#3HM"
WHILE (M345=1)
230
; Already in trigger?
; Home speed 10 cts/msec positive direction
; Home offset +100 counts (to make sure clear)
; Capture on falling flag and rising index
; Use Home3 as flag
; “Home” out of switch
; Waits for home search to start
Executing Individual Motor Moves
Turbo PMAC User Manual
ENDWHILE
WHILE (M333=0)
ENDWHILE
ENDIF
I323=-10
I326=0
I7232=3
I7233=0
CMD"#3HM"
WHILE (M345=1)
ENDWHILE
WHILE (M333=0)
ENDWHILE
DIS PLC12
CLOSE
;Waits for home motion to complete
; Home speed 10 cts/msec negative direction
; No home offset
; Capture on rising flag and rising index
; Use Home3 as flag
; Do actual homing move
; Waits for home search to start
; Waits for home motion to complete
; Disables PLC once home is found
; End of program
Storing the Home Position
Turbo PMAC automatically stores the encoder position that was captured during the latest homing search
move for the motor. This value is kept in the Motor Encoder Home Capture register [Y:$CE (Motor 1),
Y:$14E (Motor 2), etc., suggested M-variable Mxx73], which is set to zero on power-up/reset for motors
without absolute power-on position. If the position reference is obtained by reading an absolute sensor
(Ixx10>0) such as from a resolver, this register holds the negative of the absolute position read. In either
case, it contains the difference between the encoder-counter zero position (power-on position) and the
motor zero (home) position, scaled in counts.
There are two main uses for this register. First, it provides a reference for using the encoder positioncapture and position-compare registers. These registers are referenced to the encoder zero position, which
is the power-up position, not the home (motor zero) position. This register holds the difference between
the two positions. This value should be subtracted from encoder position (usually from position capture)
to get motor position, or added to motor position to get encoder position (usually for position compare).
Example: To move an axis until a trigger is found, then convert the captured encoder position to a motor
position, you can use the following M-variable definitions:
M103->X:$078003,24,S
M117->X:$078000,17
M125->Y:$0000CE,24,S
; Servo IC 0 Channel 1 position-capture register
; Servo IC 0 Channel 1 position-capture flag
; Motor 1 encoder position offset register
Now you can use a motion program segment like the following:
INC
TM10 TA10
WHILE (M117=0)
X0.2
ENDWHILE
P103=M103-M125
; Incremental moves
; Move segment time 10 msec
; While no trigger to capture position
; Command next move segment
; Read captured position; subtract offset to
; get motor position at trigger
The second use for this register is to determine whether the encoder counter has lost any counts. This can
be done by performing a second homing search move after an operation, and comparing the contents of
the register after the second homing search move to the contents after the first homing search move.
Jog-Until-Trigger Moves
The jog-until-trigger function permits a jog move to be interrupted by a trigger and terminated by a move
relative to the position at the time of the trigger. It is similar to a homing search move, except that the
motor zero position is not altered, and there is a specific destination in the absence of a trigger.
Executing Individual Motor Moves
231
Turbo PMAC User Manual
The “jog-until-trigger” function for a motor is specified by adding a ^{constant} specifier to the end
of a regular “definite” jog command for the motor, where this {constant} is the distance to be traveled
relative to the trigger position before stopping, in encoder counts. It cannot be used with the J+ and Jindefinite jog commands.
This makes the jog command for a jog-until trigger something like J=10000^100 , J=*^-50 or
J:50000^0. The value before the ^ is the destination position or distance (depending on the type of jog
command) to be traveled in the absence of a trigger. If this first value is represented by a * symbol,
PMAC looks in a pre-defined motor register (suggested M-variable Mxx72) for the position or distance.
The second value is the distance to be traveled relative to the position at the time of the trigger. This
value is always expressed as a distance, regardless of the type of jog command. Both values are
expressed in encoder counts.
The trigger condition for the motor is set up just as for homing search moves, explained above.
Turbo PMAC will use the jog parameters Ixx19-Ixx22 in force at the time of the command for the pretrigger move, and the values of these parameters in force at the time of the trigger for the post-trigger
move.
The captured value of the sensor position at the trigger is stored in a dedicated register (Y:$D8 for Motor
1, Y:$158 for Motor 2, etc.) if later access is needed. The units are in counts; for incremental encoders,
they are relative to the power-up/reset position.
Turbo PMAC sets the motor home-search-in-progress status bit (bit 10 of the first motor status word
returned on a ? command) true (1) at the beginning of a jog-until-trigger move. The bit is set false (0)
either when the trigger is found, or at the end of the move.
Turbo PMAC also sets the motor trigger move status bit (bit 7 of the second motor status word returned
on a ? command) true at the beginning of a jog-until-trigger move, and keeps it true at least until the end
of the move. If a trigger is found during the move, this bit is set false at the end of the post-trigger move;
however, if the pre-trigger move finishes without finding a trigger, the bit is left true at the end of the
move. Therefore, this bit can be used at the end of the move to tell whether the trigger was found
successfully or not. The motor “desired-velocity-zero” status bit can be used to determine the end of the
move.
Motion Program Move-Until-Trigger
The move-until-trigger construct can be used from within a motion program. In this version it is a variant
of the RAPID move mode. These moves execute exactly like on-line jog-until-trigger moves, but they are
described a little differently.
A program move-until trigger is commanded with the {axis}{data}^{data} syntax. Basic
examples are X50^-2 and Y(P1)^(P2). The first value is the destination of the axis if no trigger is
found, expressed in the engineering units for the axis. This value can be a position or a distance,
depending on whether the axis is in absolute or incremental mode, respectively. The second value is the
distance from the trigger-captured position to the end of the post-trigger move, expressed in the
engineering units for the axis. The motion program must be in RAPID mode for the triggering to operate;
otherwise just the pre-trigger move will be executed to the specified endpoint.
The commanded acceleration for the move is specified by Ixx19, Ixx20, and Ixx21, as for other trigger
moves. The magnitude of the velocity for the move is specified by maximum velocity parameter Ixx16 if
Ixx90 is at the default value of 1, or by jog speed parameter Ixx22 if Ixx90 is 0. The trigger conditions
and capture methods are specified as for other triggered moves, as described above. Status bits are set as
for on-line jog-until-trigger moves.
232
Executing Individual Motor Moves
Turbo PMAC User Manual
For more information on these moves, look under RAPID-mode moves in the Writing and Executing
Motion Programs section and in the description of {axis}{data}^{data} motion-program
statements in the Software Reference manual.
Open-Loop Moves
Open-loop moves, as their name implies, do not do closed-loop position control. They open up the servo
loop and just put commands of the specified magnitude on the outputs. Typically, these are used for
diagnostic purposes, but they can also be used in the actual applications.
These moves are executed using the motor-specific O{constant} on-line command, where
{constant} represents the magnitude of the output as a percentage of Ixx69, the maximum output
parameter for the motor. This command may not be part of a motion program, and it may not be given to
a motor when that motor’s coordinate system is executing a motion program, even if it is not moving that
motor.
If Turbo PMAC does not commutate the motor, this command creates a constant signal on the single
output for the motor. If Turbo PMAC does commutate the motor, this command sets the magnitude of
the signal that is input into the sinusoidal commutation algorithm for the motor.
To do a variable O-command, define an M-variable to the filter result register (X:$AE for Motor 1, etc.),
command an O0 to the motor to put it in open-loop mode, then assign a variable value to the M-variable.
This technique will even work on PMAC-commutated motors.
The PMAC Executive Program tuning section uses the open-loop moves to allow you to diagnose and
tune amplifier response.
Executing Individual Motor Moves
233
Turbo PMAC User Manual
234
Executing Individual Motor Moves
Turbo PMAC User Manual
TURBO PMAC COMPUTATIONAL FEATURES
Turbo PMAC has advanced computational features that permit off-loading of many operations from a
host, or even stand-alone operation in ways that were not previously possible. Many arithmetic, logical,
and transcendental operations can be performed on variables and constants in user programs on board the
card.
Computational Priorities
As a multitasking, real-time computer, Turbo PMAC has an elaborate prioritization scheme to ensure that
vital tasks are accomplished when needed, and that all tasks are executed reasonably quickly. The
scheme was designed to hide its complexity as much as possible, but also to provide some flexibility in
optimizing the controller for particular needs. The tasks in order of priority are:
1. Single Character I/O
2. Commutation Update
3. Servo Update
4. Real-Time Interrupt Tasks
5. VME Mailbox Processing
6. Background Tasks
Single Character I/O
Bringing in a single character from, or sending out a single character to, the serial port or host port (from
ISA, PCI, USB, or Ethernet) is the highest priority in Turbo PMAC. This task takes only two instruction
cycles per character, but having it at this high priority ensures that Turbo PMAC cannot be outrun by the
host on a character-by-character basis. This task is never a significant portion of Turbo PMAC’s total
calculation time. Note that this task does not include processing a full command; that happens at a lower
priority (see below).
Commutation Update
The commutation (phasing) update is the second highest priority on Turbo PMAC. Every phase interrupt
(cycle of the phase clock), the MACRO ring (if present) is updated with new information, and if I5060 >
0, a pair multiplexed ADCs (e.g. Option 12, Acc-36) is read and the data copied into de-multiplexed
RAM registers.
Every (I7 + 1) phase interrupts, Turbo PMAC performs the commutation calculations on each motor for
which commutation is active (Ixx01 bit 0 = 1). For each motor commutated by Turbo PMAC, this task
takes 1 – 2 µsec per update cycle for an 80 MHz CPU.
The phase clock frequency is determined by:
• Jumpers E98 and E29 – E33 on a Turbo PMAC
• I7m00 and I7m01 (for clock-source Servo IC m as set by I19) on a (non-Ultralite) Turbo PMAC2
• I6800 and I6801 on a Turbo PMAC2 Ultralite
The default update frequency is 9 kHz (110 µsec cycle). At the default, the commutation of each motor
takes approximately 1 – 2 percent of Turbo PMAC’s computational power.
Servo Update
The servo update is the third highest priority on Turbo PMAC. Every servo interrupt (cycle of the servo
clock), Turbo PMAC processes each of the entries of the encoder conversion table to prepare the raw
feedback and master data for use by the servo algorithms. Then it performs the servo update calculations
for each active Motor xx (Ixx00 = 1). This update consists of the interpolation calculations to compute
the next instantaneous commanded position, and the servo-loop closure calculations that use this value,
the actual position value, and the servo gain terms to compute the commanded output. (The servo-loop
closure algorithms for Motor xx can skip interrupt cycles if Ixx60 is set greater than 0.) These servo
calculations take about 4 µsec per update cycle for an 80 MHz CPU.
Turbo PMAC Computational Features
235
Turbo PMAC User Manual
The servo clock frequency is determined by:
• Jumpers E98, E29 – E33, and E3 – E6 on a Turbo PMAC
• I7m00, I7m01, and I7m02 (for clock-source Servo IC m as set by I19) on a (non-Ultralite) Turbo PMAC2
• I6800, I6801, and I6802 on a Turbo PMAC2 Ultralite
The default update frequency is 2.25 kHz (440 µsec cycle). At the default, the servo update of each
motor takes approximately 1% of Turbo PMAC’s computational power. See Closing the Servo Loop for
information on optimizing this update rate.
Real-Time Interrupt Tasks
The real-time interrupt (RTI) tasks are the fourth highest priority on Turbo PMAC. They occur
immediate after the servo update tasks at a rate controlled by parameter I8 (every I8+1 servo update
cycles). There are two significant tasks occurring at this priority level: motion program move planning
and the PLC 0 programs (interpreted and compiled).
Motion Program Move Planning
Motion program move planning consists of working through the lines of a motion program until the next
move or dwell command is encountered, and computing the equations of motion for this next part of the
move sequence. Every time Turbo PMAC starts executing a new move, it sets an internal flag indicating
that it is time to plan the next move in the program. This planning occurs at the next RTI.
PLC Program 0, Compiled PLC Program 0
PLC 0 and PLCC 0 (compiled PLC 0) are special PLC programs that execute at a higher priority than the
other PLC programs. They are meant to be used for only a few tasks that must be done at a higher
frequency than the other PLC tasks. PLC 0 and PLCC 0 will execute every real-time interrupt as long as
the tasks from the previous RTI have been completed. PLC 0 and PLCC 0 are potentially the most
dangerous tasks on Turbo PMAC as far as disturbing the scheduling of tasks is concerned. If they are too
long, they will starve the background tasks for time. The first thing you will notice is that
communications and background PLC tasks will become sluggish. In the worst case, the watchdog timer
will trip, shutting down the card, because the housekeeping tasks in background did not have the time to
keep it updated.
VME Mailbox Processing
Reading or writing a block of up to 16 characters through the VME mailbox registers is the 5th highest
priority in Turbo PMAC. The rate at which this happens is controlled by the host. This never takes a
significant portion of Turbo PMAC’s computational power.
Background Tasks
In the time not taken by any of the higher-priority tasks, Turbo PMAC will be executing background tasks.
There are three basic background tasks: command processing, interpreted and compiled PLC programs 131, and housekeeping. The frequency of these background tasks is controlled by the computational load on
Turbo PMAC: the more high-priority tasks are executed, the slower the background tasks will cycle
through; and the more background tasks there are the slower they will cycle through.
Interpreted PLC Programs 1 – 31
Interpreted PLC programs 1 – 31 are executed in background. Each background cycle, Turbo PMAC
will execute one scan (to the end or to an ENDWHILE statement) of a single active interpreted background
PLC program uninterrupted by any other background task (although it can be interrupted by higher
priority tasks). In between each scan of each interpreted background PLC program, Turbo PMAC will do
its general housekeeping, and respond to a host command, if any.
236
Turbo PMAC Computational Features
Turbo PMAC User Manual
Compiled PLC Programs 1 – 31
Compiled PLC programs (PLCC programs) 1-31 are executed in background. Each background cycle,
Turbo PMAC will execute one scan (to the end or to an ENDWHILE statement) of all active compiled
background PLC programs, starting from lowest numbered to highest, uninterrupted by any other
background task (although it can be interrupted by higher priority tasks). At power-on/reset, PLCC
programs run after the first PLC program runs.
Host Command Response
The receipt of a control character from any port is a signal to Turbo PMAC that it must respond to a
command. The most common control character is the carriage return (<CR>), which tells Turbo PMAC
to treat all the preceding alphanumeric characters as a command line. Other control characters have their
own meanings, independent of any alphanumeric characters received. Here Turbo PMAC will take the
appropriate action to the command, or if it is an illegal command, it will report an error to the host.
General Housekeeping
Each background cycle, Turbo PMAC performs its housekeeping duties to keep itself properly updated.
The most important of these are the safety limit checks (following error, overtravel limit, fault, watchdog,
etc.); general status updates are done here as well. Although this happens at a low priority, a minimum
frequency is ensured because the watchdog will trip, shutting down the card, if this frequency gets too
low. You can set this required frequency by changing variable I40 from its default value of 0. If I40 is
set to a value greater than 0, this value is the number of servo cycles within which every background
cycle must complete, or the watchdog timer will trip. (If I40 is set to 0, this number is 4096 servo cycles.)
The following diagram shows a time-line example of Turbo PMAC’s multi-tasking, focusing on the
servo, real-time interrupt, and background tasks.
PMAC Multitasking Example
*
CT
#1
#2
#3
#4
PLC 0
Servo
CT
#1
#2
#3
#4
#1
#2
Background
PLC 1 (cont.)
Servo
CT
PLC 1...
RTI
HK
PLC 2...
Background
#3
#4
PLC 2 (cont.)
Servo
HK Comm PLC 3...
Background
*
CT
#1
#2
#3
#4
PLC 0
Servo
CT
#1
#2
#3
#4
Move Planning (cont)
PLC 3 (cont) ...
RTI
Background
Servo
CT
#1
#2
Move Planning ...
RTI
#3
#4
PLC 3 (cont)
Servo
HK
PLC 1 ...
Background
*
CT
#1
#2
#3
#4
Servo
CT
#1
#2
PLC 1 (cont) ...
RTI
#3
Servo
CT - Conversion Table
# n - Motor n Servo Update
HK - Housekeeping
Turbo PMAC Computational Features
PLC 0
#4
PLC 1 (cont)
HK
Comm
Background
Comm - Communications Line Processing
- Servo Interrupt
RTI - Real Time Interrupt Task
* - Start of RTI
237
Turbo PMAC User Manual
Priority Level Optimization
Usually, Turbo PMAC will have enough speed and calculation power to perform all of the tasks asked of
it. Some applications will put a large demand on a certain priority level, and to make Turbo PMAC run
more efficiently some priority level optimization should be done.
When Turbo PMAC begins to run out of time, problems such as sluggish communications, slow
PLC/PLCC scan rates, run-time errors, and even tripping the watchdog timer, can occur. The specific
solutions to the above symptoms are discussed in the sections of this manual dedicated to those subjects.
The general solution to such problems is two-fold.
First, high priority jobs could be slowed down or moved to a lower priority position. Jobs such as the
Encoder Conversion Table, PLC/PLCC0, and the Real Time Interrupt (RTI) should be evaluated. Check
to see if everything in these jobs is necessary or if some of it could be moved to a lower priority or slowed
down. For example; A 5-axis application might not need Encoder Conversion Table entries 6 to 9.
Perhaps PLC0 could be done as PLCC1, or the RTI could be done every 4th or 5th servo cycle.
Second, the jobs could be adjusted to a priority level that gives them less emphasis. Large PLC programs
can be split into a few shorter PLC programs. This increases the frequency of housekeeping and
communications by giving more breaks in PLC scans. Motion program WHILE(condition)WAIT
statements could be done as follows;
WHILE(condition)
DWELL20
ENDWHILE
This will give more time to other tasks of equal or lower priority such as PLC programs and
communications.
Evaluating the Turbo PMAC’s Computational Load
Turbo PMAC controllers offer facilities that permit you to calculate the computational loads you are
putting on the processor. There are several key timer registers to use in calculating these loads. These
registers are scaled so that one increment of the timer is two clock cycles of the DSP. So if the DSP were
running at a clock frequency of exactly 80 MHz – a clock period of 12.5 nsec – one increment of the
timer would be 25 nsec.
The DSP’s clock frequency is multiplied up from the crystal clock frequency of 19.66 MHz, using the
saved value of I52, according to the formula:
DSPfrequency =
19.66 MHz
* (I 52 + 1)
2
In terms of period, the timer increment – 2 DSP cycles – can be calculated as:
TimerIncrement (n sec ) =
203.4
I 52 + 1
Phase Interrupt Tasks
There are two key timer registers for evaluating the computational load of the phase-interrupt tasks such
as commutation, current-loop closure, and ADC de-multiplexing. The first is a hardware timer in the
DSP, at address X:$FFFF8C. This register holds the number of timer increments between the last two
phase interrupts, establishing the period of the phase interrupt. This can be used to verify the phase
period you think you have, and with other registers, computational duty cycles.
The second register, located at X:$000037, holds the number of timer increments from the beginning to
the end of the phase-interrupt tasks for the last interrupt. When divided by the time between phase
interrupts, this will give the duty cycle of the phase-interrupt tasks.
238
Turbo PMAC Computational Features
Turbo PMAC User Manual
Servo Interrupt Tasks
Another timer register can be used to evaluate the computation load of the servo-interrupt tasks such as
the conversion table, interpolation, position/velocity-loop closure, and data gathering. This register,
located at Y:$000037, holds the number of timer increments elapsed from the beginning to the end of the
servo-interrupt tasks for the last interrupt.
If this time plus the phase-task time is less than the time between phase interrupts (X:$37 + Y:$37 <
X:$FFFF8C), then this is the actual time the servo tasks took. However, if the sum of these times is
greater than one phase cycle (X:$37 + Y:$37 > X:$FFFF8C), then the servo tasks were interrupted (at
least once) by phase tasks, and the time for the interrupting phase tasks must be subtracted out (see
example below).
When the net time for the servo tasks is divided by the product of the phase-interrupt period and the
number of phase-interrupts per servo-interrupt, the result is the duty cycle of the servo-interrupt tasks.
Note that certain servo tasks, such as data gathering, foreground motor data reporting, and even servoloop closure if Ixx60 > 0, do not have to be executed every servo cycle, so the duty cycle can vary.
Real-Time Interrupt Tasks
Two timer registers provide information on the loading of real-time interrupt (RTI) tasks such as PLC 0,
PLCC0, and motion-program calculations. The first register, at X:$00000B, holds the number of timer
increments from the beginning to the end of the RTI tasks for the last interrupt. The second register, at
Y:$00000B, holds the largest number of timer increments from the beginning to the end of a set of RTI
tasks since the last power-up/reset.
If these times plus the phase and servo-task times are less than the time between phase interrupts (X:$37
+ Y:$37 + X/Y:$0B < X:$FFFF8C), then these are the actual times the RTI tasks took. However, if these
times are greater than one phase cycle (X:$37 + Y:$37 + X/Y:$0B > X:$FFFF8C), then the RTI tasks
were interrupted (at least once) by phase tasks, and the time for the interrupting phase tasks must be
subtracted out. Also, if these times are greater than one servo cycle, then the RTI tasks were also
interrupted by servo tasks (see example below).
Dividing the latest net time for the RTI tasks by the product of the phase interrupt period, the number of
phase interrupts per servo interrupt, and the number of servo interrupts per RTI yields the duty cycle of
the RTI tasks. The duty cycle for real-time interrupt tasks can vary widely within an application, so it is
advisable to compute a running average to compute general loading.
Total Interrupt Tasks
The total duty cycle for Turbo PMAC interrupt tasks can be calculated by summing the duty cycles for
the three types of interrupt tasks. In general, it is recommended that the duty cycle for phase and servo
tasks does not exceed 50%, and the duty cycle for all foreground tasks does not exceed 75%. These are
not strict limits – it is possible to exceed them, but the timing of all operations should be carefully
evaluated if these guidelines are exceeded.
Sample Monitoring Program
The following sample code can be used to monitor the total interrupt-task duty cycle:
M70->X:$FFFF8C,0,24
M71->X:$000037,0,24
M72->Y:$000037,0,24
M73->X:$00000B,0,24
P70=4
P76=16
;
;
;
;
;
;
Time between phase interrupts
Time for phase tasks
Time for servo tasks
Time for RTI tasks
4 phase interrupts per servo interrupt
Length of filter for averaging duty cycle
OPEN PLC 17 CLEAR
P71=M71/M70
P69=INT((M71+M72)/M70)
; Phase task duty cycle
; # of times phase interrupted servo
Turbo PMAC Computational Features
239
Turbo PMAC User Manual
P72=(M72-P69*M71)/(M70*P70)
; Servo task duty cycle
P68=INT((M71+M72+M73)/M70)
; # of times phase interrupted RTI
P67=INT((M71+M72+M73)/(M70*P70))
; # of times servo interrupted RTI
P73=(M73-P68*M71-P67*(M72-P69*M71))/(M70*P70*(I8+1))
; RTI task duty cycle
P74=P71+P72+P73
; Latest total foreground duty cycle
P75=(P75*(P76-1)+P74)/P76
; Averaged total foreground duty cycle
CLOSE
Background Cycle Time
There are two timer registers important to evaluate the time required to execute a background cycle.
These registers are involved in the operation of the watchdog timer. The register at address Y:$000025 is
set to the value of I40 at the end of every background cycle. (If I40 is 0, the register is set to 4095.) Until
the next background cycle is completed, this register is decremented every servo cycle. The data
gathering function is useful to establish how long background cycles take.
The register at X:$000025 contains the lowest value reached by Y:$000025 since the last power-up/reset
of the Turbo PMAC. If this is close to 0, the Turbo PMAC has come close to tripping its watchdog timer,
and background tasks such as PLC program execution, communications response, and safety checks have
been slow.
Numerical Values
Turbo PMAC can store and process numerical values in many forms, with both fixed-point and floatingpoint values. The Motorola 56300 DSP that acts as Turbo PMAC’s CPU is a fixed-point processor with
built-in 24-bit and 48-bit arithmetic capability (plus a 56-bit accumulator). However, Turbo PMAC’s
firmware implements a full set of floating-point routines.
Internal Formats
The internal servo, interpolation, and commutation routines all operate with fixed-point arithmetic, 24-bit
and 48-bit, for maximum speed. The user programs, motion and PLC, use 48-bit floating-point arithmetic
for maximum range and generality. Even when reading from and/or writing to fixed-point registers, the
intermediate formats are all floating-point values. This permits users to mix different variable types at
will, letting Turbo PMAC handle the type matching automatically
The only exception to this rule is the compiled PLC programs; in a statement containing only “Lvariables” and integer constants that can fit in a signed 24-bit range (+/-8M), the intermediate format is
signed 24-bit integer. Refer to the section on compiled PLCs in the Writing and Executing a PLC
Program section of this manual for more details.
The general floating-point format is 48 bits long, with a 36-bit mantissa and a 12-bit exponent. This
provides a range of +2+2047, or +3.233 x 10+616, which should provide sufficient range for any foreseeable
uses of the card.
240
Turbo PMAC Computational Features
Turbo PMAC User Manual
The internal format of 48-bit floating-point registers is shown in the following table:
X-word:
Bit:
Part:
Val:
Bit:
Part:
Val:
23
Mant
-20
11
Mant
+2-12
22
Mant
+2-1
10
Mant
+2-13
21
Mant
+2-2
9
Mant
+2-14
20
Mant
+2-3
8
Mant
+2-15
19
Mant
+2-4
7
Mant
+2-16
18
Mant
+2-5
6
Mant
+2-17
17
Mant
+2-6
5
Mant
+2-18
16
Mant
+2-7
4
Mant
+2-19
15
Mant
+2-8
3
Mant
+2-20
14
Mant
+2-9
2
Mant
+2-21
13
Mant
+2-10
1
Mant
+2-22
12
Mant
+2-11
0
Mant
+2-23
22
Mant
+2-25
10
Exp
+210
21
Mant
+2-26
9
Exp
+29
20
Mant
+2-27
8
Exp
+28
19
Mant
+2-28
7
Exp
+271
18
Mant
+2-29
6
Exp
+26
17
Mant
+2-30
5
Exp
+25
16
Mant
+2-31
4
Exp
+24
15
Mant
+2-32
3
Exp
+23
14
Mant
+2-33
2
Exp
+22
13
Mant
+2-34
1
Exp
+21
12
Mant
+2-35
0
Exp
+20
Y-word:
Bit:
Part:
Val:
Bit:
Part:
Val:
23
Mant
+2-24
11
Exp
+211
Mant:
Mantissa – The mantissa of a floating-point number in standard format. The
mantissa should have the range 0.5 <= Mant < 1.0, or –1.0 <= Mant < -0.5. Mant
= 0.0 when number is exactly 0.
Exp:
Exponent – The exponent of a floating-point number in modified format. The 12bit value here (0 <= Exp <= 4095) should have a value of 2047 subtracted from it
(n = Exp -2047; -2048 <= n <= +2047); then 2 is raised to the power n, and the
resulting value is multiplied by the mantissa. Exp = 0 when number is exactly 0.
Receiving Values
Constant values sent from the host as part of command lines are sent as ASCII text, either as decimal
values or hexadecimal values. Hexadecimal values must be preceded by a $ character; they must be
unsigned, and they cannot include fractional values. Decimal values can be positive or negative, and can
include fractional values. The Turbo PMAC value interpreter does not support exponential notation, and
it is limited to passing through values in the range +2+35, or +3.43 x 10+10. Values outside this range are
truncated to the maximum or minimum values of the range.
Examples:
1234
3
03
-27.656
0.001
.001
$ff00
(leading zeros OK)
(leading zero not required)
(interpreted as hexadecimal)
Reporting Values
Turbo PMAC reports numerical values to the host computer as part of response lines in decimal ASCII
text form (although address values can be reported in hexadecimal ASCII form if I9 = 2 or 3 – see below).
+47
+14
The value reporter is limited to passing values in the range of +2 , or +1.41 x 10 . Values outside of
this range are truncated to the maximum or minimum values of the range.
Turbo PMAC Computational Features
241
Turbo PMAC User Manual
Addresses
Turbo PMAC uses a Motorola DSP563xx as its processor. The DSP563xx has dual 24-bit address spaces
(of which 19 bits are used by the Turbo PMAC) for memory and I/O. (Note that the I/O in Turbo PMAC
is memory-mapped; it does not have a separate I/O space as your PC does.) When specifying an address
in Turbo PMAC, you must state which half of memory (X or Y) – or both halves (L) for a long 48-bit
word – followed by an optional colon, followed by the numerical address itself. The numerical address is
a constant in the range of $000000 - $07FFFF. Usually, the address values are specified as a hexadecimal
value (with the $ prefix).
Do not confuse the memory and I/O addresses of Turbo PMAC itself with those of the host computer.
Examples of legal address specifications are:
Y:$078A02
X:136
X$078003
(word containing machine I/O)
(Motor 1 commanded position – also X:$88)
(captured encoder 1 position)
This form of address specification is used particularly in M- variable definitions and direct read (R) and
write (W) commands. There are I-variables that specify addresses, but usually these are pre-defined to the
X or Y space, so all that is needed is the numerical value. The data-gathering-address I- variables (I5001I5048) use an extra hex digit in front of the numerical value to specify the memory half (see I5001
description).
Variables
Turbo PMAC has several types of variables. In Turbo PMAC, a variable is specified by a single letter (I,
P, Q, or M) followed by a number from 0 to 8191. Each letter denotes a different type of variable, each
type with its own properties. The different types share the characteristics that when their name is cited in
an expression, the current value of the variable is used (reading from them); and values can be assigned to
them in an equation (writing to them).
Custom variable names on Turbo PMAC may not be specified; however the Editor in the PMAC
Executive Program has a substitution (macro) scheme that allows programs to be written using userdefined variable names, but changes these names into Turbo PMAC-legal variable names during the
download process. This substitution scheme is strongly recommended for managing large applications.
I-Variables
I-Variables (initialization or setup variables) determine the personality of the card for a given application.
They are at fixed locations in memory and have pre-defined meanings. Most are integer values, and their
range varies depending on the particular variable. There are 1024 I-variables, from I0 to I8191, and they
are organized as follows:
I0 – I99
I100 – I199
I200 – I299
…
I3200 – I3299
I3300 – I4799
I4900 – I4999
I5000 – I5099
I5100 – I5199
I5200 – I5299
…
I6600 – I6699
I6800 – I6999
I7000 – I7999
I8000 – I8191
242
Global card setup
Motor 1 setup
Motor 2 setup
Motor 32 setup
Supplemental Motor setup
Configuration status
Data gathering/ADC demux setup
Coordinate System 1 setup
Coordinate System 2 setup
Coordinate System 16 setup
MACRO IC setup
Servo IC setup
Encoder conversion table setup
Turbo PMAC Computational Features
Turbo PMAC User Manual
Value Assignment
Values assigned to an I-variable may be either a constant or an expression. The commands to do this are
on-line (immediate) if no buffer is open when sent, or buffered program commands if a buffer is open.
Examples:
I120=45
I120=I120+5
I(P1*100+20)=10
Limited Range
For I-variables with limited range, an attempt to assign an out-of-range value does not cause an error.
The value is automatically rolled over to within the range by modulo arithmetic (truncation). For
example, I3 has a range of 0 to 3 (4 possible values). The command I3=5 would assign a value of (5
modulo 4) = 1 to the variable.
Non-Volatile Storage
When a value is assigned to an I-variable, the value in an active memory register that is volatile (does not
retain its value when power is removed) is being removed. Copy all of these active values to non-volatile
flash memory with the SAVE command. On a power-up or reset, the last values saved into flash memory
are copied back into active memory automatically for use.
Default Values
Default values for all I-variables are contained in the manufacturer-supplied firmware. They can be used
individually with the I{constant}=* command, or in a range with the
I{constant}..{constant}=* command. Upon board re-initialization by the $$$*** command
or by a reset with the re-initialization jumper in the non-default setting, all default settings are copied
from the firmware into active memory. The last saved values are not lost; they are just not used.
See the I-variable description section for the functions of individual variables.
P-Variables
P-variables are general-purpose user variables. They are 48-bit floating-point variables at fixed locations
in Turbo PMAC’s memory, but with no pre-defined use. There are 8192 P-variables, from P0 to P8191.
A given P-variable means the same thing from any context within the card; all coordinate systems have
access to all P-variables (contrast Q-variables, which are coupled to a given coordinate system, below).
This allows for useful information passing between different coordinate systems. P-variables can be used
in programs for any purpose desired: positions, distances, velocities, times, modes, angles, intermediate
calculations, etc.
Type of Memory Used
P-variables can be located in either the main memory, or in the supplemental battery-backed parameter
memory (if Option 16 is ordered). If I46 is set to 0 (default) or 2, the P-variables are located in the main
memory, which has fast access (1 wait state) but whose values are not retained without a SAVE command
copying the values to flash memory. On power-up/reset, the last saved values are copied from flash
memory into the active variable registers in RAM.
If I46 is set to 1 or 3, the P-variables are located in the Option 16 battery-backed RAM, which has slow
access (9 wait states) but whose values are retained automatically by the battery when power is removed.
Special-Use P-Variables
Generally, Turbo PMAC firmware has no automatic use of P-variables. However, it can make special use
of variables P0 – P32 and P101 – P132, as described below.
Turbo PMAC Computational Features
243
Turbo PMAC User Manual
If a command consisting simply of a constant value is sent to Turbo PMAC, Turbo PMAC assigns that
value to variable P0 (unless a special table buffer such as a compensation table or stimulus table has been
defined but not yet filled – in that case the constant value will be entered into the table. For example, if
the command 342<CR> is sent to Turbo PMAC, it will interpret it as P0=342<CR>.)
This capability is intended to facilitate simple operator terminal interfaces. It does mean, however, that it
is not a good idea to use P0 for other purposes, because it is easy to change this variable’s value
accidentally.
If your application uses kinematic subroutines to convert between tool-tip (axis) positions and joint
(motor) positions, variables P1 – P32 and P101 – P132 are used for the motor positions in these
subroutines (Pn is Motor n position; if PVT moves are converted, P10n is Motor n velocity). If you are
using the kinematic subroutines, make sure not to use the P-variables employed in the subroutines for any
other purpose.
Q-Variables
Q-variables, like P-variables, are general-purpose user variables – 48-bit floating-point variables at fixed
locations in memory, with no pre-defined use. However, the register using a given Q-variable (and hence
the value contained in it) is dependent on which coordinate system is utilizing it. This allows several
coordinate systems to use the same program (for instance, containing the line X(Q1+25) Y(Q2), but to
have different values in their own Q variables (which in this case, means different destination points).
Allotting Q-Variables
There is a total of 8192 Q-variables. If using only a single coordinate system (Coord.Sys. 1 – specified as
&1), use all of these: Q0 to Q8191. The Q-variables of Coordinate System 2 (&2) overlap these: Q0 of
&2 is the same thing as Q4096 of &1, and Q4095 of &2 is the same thing as Q8191 of &1. (The Q buffer
is actually rotary, so Q4096 of &2 is the same thing as Q0 of &1, and Q8191 of &2 is Q4095 of &1.)
Thus, both coordinate systems have 4096 unique Q-variables: Q0 to Q4095.
Note:
There is no protection against overwriting another coordinate system’s Qvariables. It is your responsibility to keep Q-numbers within the proper range.
If using 3 or 4 coordinate systems, there are 2048 unique Q-variables for each coordinate system (Q0 –
Q2047). If using 5 to 8 coordinate systems, there are 1024 unique Q-variables for each coordinate system
(Q0 – Q1023). If using 9 to 16 coordinate systems, there are 512 unique Q-variables for each coordinate
system (Q0 – Q511).
The following table shows the addresses used to store Q0 for each coordinate system when the Qvariables are stored in main memory. It shows how they potentially overlap.
CS #
Q0
Address
CS #
Q0
Address
CS #
Q0
Address
CS #
Q0
Address
1
2
3
4
$8000
$9000
$8800
$9800
5
6
7
8
$8400
$8C00
$9400
$9C00
9
10
11
12
$8200
$8600
$8A00
$8E00
13
14
15
16
$9200
$9600
$9A00
$9E00
Addressing a Q-Variable Set
The type of command determines which set of Q-variables being used. When accessing a Q-variable
from an on-line (immediate) command from the host, the Q-variable for the currently host-addressed
coordinate system (with the &n command) is being used.
244
Turbo PMAC Computational Features
Turbo PMAC User Manual
When accessing a Q-variable from a motion program statement (including kinematic subroutines), the Qvariable belonging to the coordinate system running the program is being used. If a different coordinate
system runs the same motion program, it will use different Q-variables.
When accessing a Q-variable from a PLC program statement, the Q-variable for the coordinate system
that has been addressed by that PLC program with the ADDRESS command is being used. Each PLC
program can address a particular coordinate system independent of other PLC programs and independent
of the host addressing. If no ADDRESS command is used in the PLC program, the program uses the Qvariables for C.S. 1.
Type of Memory Used
Q-variables can be located in either the main memory or in the supplemental battery-backed parameter
memory (if Option 16 is ordered). If I46 is set to 0 (default) or 1, the Q-variables are located in the main
memory, which has fast access (1 wait state) but whose values are not retained without a SAVE command
copying the values to flash memory. On power-up/reset, the last saved values are copied from flash
memory into the active variable registers in RAM.
If I46 is set to 2 or 3, the Q-variables are located in the Option 16 battery-backed RAM, which has slow
access (9 wait states) but whose values are retained by the battery automatically when power is removed.
Special-Use Q-Variables
Several Q-variables have special uses to watch for. The ATAN2 (two-argument arctangent) function
automatically uses Q0 as its second argument (the cosine argument). The READ command places the
values it reads following letters A through Z in Q101 to Q126, respectively, and a mask word denoting
which variables have been read in Q100. The S (spindle) statement in a motion program places the value
following it into Q127.
If the application uses kinematic subroutines to convert between tool-tip (axis) positions and joint (motor)
positions, variables Q1 – Q10 and possibly Q11 – Q19 for the coordinate system are used for the axis data
in these subroutines (Q1 – Q9 are for axis positions; Q10 tells whether PVT moves are being converted; if
PVT moves are converted, Q11 – Q19 are for axis velocities). If using the kinematic subroutines, make
sure not to use the Q-variables employed in the subroutines for any other purpose.
M-Variables
M-variables are provided to permit easy user access to Turbo PMAC’s memory and I/O space. Generally,
a definition only needs to be made once, with an on-line command. The SAVE command must be used to
retain the definition through a power-down or reset. Define an M-variable by assigning it to a location
and defining the size and format of the value in this location. An M-variable can be a bit, a nibble (4
bits), a byte (8 bits), 1-1/2 bytes (12 bits), a double-byte (16 bits), 2-1/2 bytes (20 bits), a 24-bit word, a
48-bit fixed-point double word, a 48-bit floating-point double word, or special formats for dual-ported
RAM and for the JTHW multiplexer port.
There are 8192 M-variables (M0 to M8191), and as with other variable types, the number of the Mvariable may be specified with either a constant or an expression: M576 or M(P1+20).
M-Variable Definitions
The definition of an M-variable is done using the defines-arrow (->) composed of the minus sign and
greater than symbol. Generally, a definition needs to be made only once, with in an on-line command,
because it is stored in battery-backed RAM or saved to flash memory. The M-variable thus defined may
be used repeatedly.
An M-variable may take one of the following types, as specified by the address prefix in the definition:
X:
Y:
D:
1 to 24 bits fixed-point in X-memory
1 to 24 bits fixed-point in Y-memory
48 bits fixed-point across both X- and Y-memory
Turbo PMAC Computational Features
245
Turbo PMAC User Manual
L:
DP:
F:
TWD:
TWB:
TWS:
TWR:
*:
48 bits floating-point across both X- and Y-memory
32 bits fixed-point (low 16 bits of X and Y) (for use in dual-ported RAM)
32 bits floating-point (low 16 bits of X and Y) (for use in dual-ported RAM)
Multiplexed BCD decoding from Thumbwheel port
Multiplexed binary decoding from Thumbwheel port
Multiplexed serial I/O decoding from Thumbwheel port
Multiplexed serial resolver decoding from Thumbwheel port
No address definition; uses part of the definition word as generalpurpose variable
If an X or Y type of M-variable is defined, you must also define the starting bit to use, the number of bits,
and the format (decoding method).
Typical M-variable definition statements are:
M1->Y:$078C02,8,1
M102->Y:$78003,8,16,S
M103->X:$078003,0,24,S
M161->D:$8B
M5141->L:$2041
M50->DP:$060401
M51->F:$0607FF
M100->TWD:4,0.8.3,U
See the instructions for each type of M-variable definition in the On-Line Commands reference section of
the manual. This can be found in the Talking to Turbo PMAC section of this manual. Many suggested
M-variable definitions are given in the software reference manual as well.
It is a good idea to prepare a single file with all of your M-variable definitions and to put at the top of this
file the command M0..8191->*. This will remove all existing definitions, and help to prevent
mysterious problems caused by “stray” M-variable definitions.
The M-variable definitions are stored as 48-bit codes at Turbo PMAC memory addresses $004000 (for
M0) to $005FFF (for M8191). The Y-register contains the address of the register pointed to by the
definition; the X-register contains a code that determines what part of the register is used and how it is
interpreted.
If another M-variable points to the Y-register, it can be used to change the subject register. The main use
of this technique is to create arrays of registers, which can be used to walk through tables in memory.
Limited Range
Many M-variables have a more limited range than Turbo PMAC’s full computational range. If a value
outside of the range of an M-variable is placed to that M-variable, Turbo PMAC rolls over the value
automatically to within that range and does not report any errors.
For example, with a single bit M-variable, any odd number written to the variable ends up as 1, any even
number ends up as 0. If an attempt is made to place a non-integer value in an integer M-variable, Turbo
PMAC automatically rounds to the nearest integer.
Using M-Variables
Once defined, an M-variable may be used in programs just as any other variable – through expressions.
When the expression is evaluated, Turbo PMAC reads the defined memory location, calculates a value
based on the defined size and format, and utilizes it in the expression.
Care should be exercised in using M-variables in expressions. If an M-variable is something that can be
changed by a servo routine (such as instantaneous commanded position), which operates at a higher
priority than the background expression evaluation, there is no guarantee that the value will not change in
the middle of the evaluation. For instance, if in the expression
(M16-M17)*(M16+M17)
246
Turbo PMAC Computational Features
Turbo PMAC User Manual
the M-variables are instantaneous servo variables, there is no guarantee that M16 or M17 will have the
same value in both places in the expression or that the values for M16 and M17 will come from the same
servo cycle. The first problem can be overcome by setting P1=M16 and P2=M17 right above this, but
there is no general solution to the second problem.
Use for Indirect Addressing
As pointers, M-variables can be used for a technique known as indirect addressing to access a range of
registers without having to define a separate variable for each register. This technique uses two Mvariables. The first is assigned to a register in the address area of interest, with a format of the desired
type; the second is assigned to the register that contains the address of the first definition.
M-variable address definitions are in fixed locations in Turbo PMAC memory, starting at $004000 (for
M0) and ending at $005FFF (for M8191). The X-register at each of these addresses holds the code that
determines the format of the M-variable; the Y-register holds the address of the register being pointed to.
By changing the contents of this Y-register, you can change the address of the register that this Mvariable points to.
This technique is best illustrated by an example. Suppose that a 2048-word UBUFFER had been created
in a Turbo PMAC with standard memory. This UBUFFER would occupy addresses $010000 to
$0107FF. In this buffer, we wanted to create a 2048-entry floating-point sine table. We would start off
with two M-variable definitions:
M64->L:$010000
M65->Y:$004040,0,12
; Floating-point M-var def to start of UBUFFER
; M64 definition address, low 12 bits
Now, by changing the value of M65, we change the address to which M64 points. Note that by assigning
M65 to only the low 12 bits (last 3 hex digits) of the M64 definition address, we can in this case just
assign values to M65 representing offsets from the beginning of the register set. To create the sine table,
use the following code:
M65=0
WHILE (M65<2048)
M64=SIN(360*M65/2048)
M65=M65+1
ENDHWILE
; Point M64 to L:$010000
; Write sine value
; Index M64 to next register
Operators
Turbo PMAC operators work like those in any computer language: they combine values to produce new
values. Detailed descriptions of the operators are given in the Software Reference manual; overviews are
given here.
Arithmetic Operators
Turbo PMAC uses the four standard arithmetic operators: +, -, *, and /. The standard algebraic
precedence rules are used: multiply and divide are executed before add and subtract, operations of equal
precedence are executed left to right, and operations inside parentheses are executed first.
Modulo Operator
Turbo PMAC also has the ‘%’ modulo operator, which produces the resulting remainder when the value
in front of the operator is divided by the value after the operator. Values may be integer or floating point.
This operator is particularly useful for dealing with counters and timers that roll over.
When the modulo operation is done by a positive value x, the results can range from 0 to x (not including
x itself). When the modulo operation is done by a negative value x, the results can range from -x to x (not
including x itself). This negative modulo operation is useful when a register can roll over in either
direction.
Turbo PMAC Computational Features
247
Turbo PMAC User Manual
Logical Operators
Turbo PMAC has three logical operators that do bit-by-bit operations: & (bit-by-bit AND), | (bit-by-bit
OR), and ^ (bit-by- bit EXCLUSIVE OR). If floating-point numbers are used, the operation works on the
fractional as well as the integer bits. & has the same precedence as * and /; | and ^ have the same
precedence as + and -. Use of parentheses can override these default precedence levels.
Note:
These bit-by-bit logical operators are different from the simple Boolean operators
AND and OR used in compound conditions (q.v.).
Functions
As in any programming language, Turbo PMAC mathematical functions perform mathematical operations
on constants or expressions to yield new values. The general format is:
{function name} ({expression})
The available functions are:
SIN:
COS:
TAN:
ASIN:
ACOS:
ATAN:
ATAN2:
LN:
EXP:
SQRT:
ABS:
INT:
Trigonometric sine
Trigonometric cosine
Trigonometric tangent
Trigonometric inverse sine (arcsine)
Trigonometric inverse cosine (arccosine)
Trigonometric inverse tangent (arctangent)
Special two-argument, four-quadrant trigonometric inverse tangent
Natural logarithm (log base e)
Exponentiation (ex)
Positive square root
Absolute value
Truncation to integer (towards minus infinity)
Global I-variable I15 controls whether the trigonometric functions use degrees (I15 = 0) or radians (I15 = 1).
Detailed descriptions of each function, including domain and range, are given in the Software Reference
Manual.
Expressions
A Turbo PMAC expression is a mathematical construct consisting of constants, variables, and functions,
connected by operators. Expressions can be used to assign a value to a variable, to determine a motion
program parameter, or as part of a condition. A constant can be an expression, so if the syntax calls for
{expression}, a constant may be used as well as a more complicated expression. No extra
parentheses are required for non-constant expressions, unlike when {data} is specified (see below).
Examples of expressions are:
512
P1
P1-Q18
1000*COS(Q25*3.14159/180)
I100*ABS(M347)/ATAN(P(Q3+1)/6.28)+5
The {DATA} Syntax
For Turbo PMAC purposes, if command syntax requires {data}, utilize either a constant that is not
surrounded by parentheses or an expression that is surrounded by parentheses. (Since a constant can be
an expression, it is legal to put a constant in parentheses, but this takes more storage and more calculation
time.)
For example, if the listed command syntax is T{data}, it is legal to use T100, T(P1+250*P2), or
T(100) (which is legal but wasteful).
248
Turbo PMAC Computational Features
Turbo PMAC User Manual
Variable Value Assignment Statement
This type of statement calculates and assigns a value to a variable. When a value assignment statement is
sent to Turbo PMAC, if a program buffer is open, the statement is added to the buffer. If not, it is
executed immediately. The standard assignment syntax is:
{variable name}={expression}
where {variable name} specifies which variable is to be used, and {expression} represents the
value to be assigned to the variable.
I-Variable Default Value Assignment
A statement with the syntax:
{I-variable}=*
will assign to the specified I-variable the manufacturer’s default value for that variable (not the user’s last
saved value).
Synchronous M-Variable Value Assignment
In a motion program, when Turbo PMAC is blending or splining moves together, it must be calculating in
the program ahead of the actual point of movement. This is necessary in order to be able to blend moves
together at all, and also to be able to do reasonable velocity and acceleration limiting. Depending on the
mode of movement, calculations while blending may occur one, two, or three moves ahead of the actual
movement.
Why Needed
When assigning values to variables is part of the calculation, the variables will get their new values ahead
of their place in the program when looking at actual move execution. Generally, for P and Q-variables,
this is not a problem, because they exist only to aid further motion calculations. However, for Mvariables, particularly outputs, this can be a problem, because with a normal variable value assignment
statement, the action will take place sooner than is expected, looking at the statement’s place in the
program.
For example, in the program segment
X10
M1=1
X20
; Move X-axis to 10
; Turn on Output 1
; Move X-axis to 20
it might be expected that Output 1 would be turned on at the time the X-axis reached position 10. But
since Turbo PMAC is calculating ahead, at the beginning of the move to X10, it will have already
calculated through the program to the next move, working through all program statements in between,
including M1=1, which turns on the output. Therefore, using this technique, the output will be turned on
sooner than desired.
How They Work
Synchronous M-variable assignment statements were implemented as a solution to this problem. When
one of these statements is encountered in the program, it is not executed immediately; rather, the action is
put on a stack for execution at the start of the actual execution of the next move in the program. This
makes the output action properly synchronous with the motion action.
In the modified program segment
X10
M1==1
X20
; Move X-axis to 10
; Turn on Output 1 synchronously
; Move X-axis to 20
the statement M1==1 (the double-equals indicates synchronous assignment) is encountered at the
beginning of the move to X10, but the action is not actually performed until the start of blending into the
next move (X20).
Turbo PMAC Computational Features
249
Turbo PMAC User Manual
Note:
With synchronous assignment, the actual assignment is performed where the
blending to the new move begins, which is generally ahead of the programmed
point. In LINEAR and CIRCLE mode moves, this blending occurs V*TA/2
distance ahead of the specified intermediate point, where V is the commanded
velocity of the axis, and TA is the acceleration (blending) time.
Also, note that the assignment is synchronous with the commanded position, not necessarily the actual
position. It is the responsibility of the servo loop to make the commanded and actual positions match
closely.
In applications in which Turbo PMAC is executing segmented moves (Isx13 > 0), the synchronous Mvariables are executed at the start of the first Isx13 spline segment after the start of blending into the
programmed move.
Turbo PMAC checks to see whether it is time to pull synchronous assignments out of the queue and
execute them every real-time interrupt (every I8+1 servo cycles). The smaller I8 is, and the smaller the
servo cycle time is, the tighter the timing control of the synchronous outputs is.
Note:
Synchronous M-variables after the last move or DWELL in the program do not
execute when the program ends or temporarily stops. Use a DWELL as the last
statement of the program to execute these statements.
If synchronous assignments are left in the queue because the program ended or was stopped before it was
time for their execution, they can be removed from the queue with the on-line MFLUSH command.
Syntax. There are four forms of synchronous M-variable assignment statements:
M{constant}=={expression}
M{constant}&={expression}
M{constant}|={expression}
M{constant}^={expression}
;
;
;
;
Straight equals assignment
AND-equals assignment
OR-equals assignment
XOR-equals assignment
In all of these forms, the expression on the right side of the statement is evaluated when the line is
encountered in the program, ahead of the execution of the move. The value of the expression, the
variable number, and the operator are placed on a stack for execution at the proper time.
Execution
When actual execution of the appropriate move starts, these items are pulled off the stack, and the actual
action is performed. In the case of the == syntax, the value is simply assigned to the variable at this time.
In the case of the other forms (&=, |=, and ^=), the variable is read at this time, the bit-by-bit Boolean
operation (AND, OR, XOR, respectively) is performed between the variable value and the expression
value, and the result is written back to the variable.
Special Boolean Feature
These Boolean assignment operators are subtly different from what would seem to be equivalent ==
statements. Consider the two statements acting on an 8-bit M-variable, which attempt to make all of the
odd bits 1, while leaving the even bits where they are:
M50==M50 & $AA
M50&=$AA
The difference between the two statements is apparent when M50 is read for the operation. In the first
case, it is read when the statement is first evaluated in the program. In the second case, it is read when the
operation is pulled off the stack, immediately before the variable is written to. In this second technique
there is no chance that the value of the M-variable can be changed by some other task in the mean time.
250
Turbo PMAC Computational Features
Turbo PMAC User Manual
Synchronous Assignment of Other Variables
Only M-variables can be used in these synchronous assignments, but M-variables can be assigned to the
registers of any other variables (I, P, or Q), so these synchronous assignments can be used effectively on
other variable types as well. Refer to the detailed memory map for addresses of other variables.
Limitations
There are a few limitations to these synchronous assignments that the user must be aware of:
Valid Forms: First, these statements may not be used with any of the thumbwheel-multiplexer-port Mvariable forms (TWB, TWD, TWR, or TWS). The Boolean assignments (&=, |=, ^=) cannot be used
with any double-width M-variable forms (D, L, or F).
Queue Limits: The pending synchronous outputs must be stored in a queue of finite size. Without the
special lookahead buffer enabled (see below), the queue is fixed at a size of 256 words for all coordinate
systems combined, with two words required to store each assignment. Global variable I68, which
controls the number of active coordinate systems, determines how much of this queue is assigned to each
coordinate system.
The following table shows how many words of the queue are assigned to each active coordinate system
for the different values of I68, and how many assignments can be made per move with and without cutter
radius compensation active. (Cutter radius compensation requires the motion program to work ahead an
extra move, so reduces the number per move.) Note that two words are required per assignment, and one
additional word per move.
I68
Value
Highest Numbered
Coordinate System
Activated
Sync. M-Var.
Stack per C.S.
Max. Sync M-Var.
Assignments per
move, no cutter comp
Max. Sync M-Var.
Assignments per
move, cutter comp on
0
1
2-3
4-7
8 - 15
C.S. 1
C.S. 2
C.S. 3 - 4
C.S. 5 - 8
C.S. 9 - 16
256 words
128 words
64 words
32 words
16 words
63
31
15
7
3
42
20
10
4
2
If you are using the special multi-block lookahead feature, the storage requirements for pending
synchronous assignments can be significant. For this reason, the special lookahead buffer permits you to
explicitly reserve space for pending synchronous outputs as well as pending move segments. If the
following command were given to define a lookahead buffer:
&1 DEFINE LOOKAHEAD 500,100
then a lookahead buffer for Coordinate System 1 would be defined large enough to store 500 move
segments (each of Isx13 programmed time) for each motor in the coordinate system, plus 100
synchronous M-variable assignments.
Comparators
A comparator evaluates the relationship between two values (constants or expressions). It is used to
determine the truth of a condition in a motion or PLC program. The valid comparators for Turbo PMAC are:
=
!=
>
!>
<
!<
~
!~
(equal to)
(not equal to)
(greater than)
(not greater than; less than or equal to)
(less than)
(not less than; greater than or equal to)
(approximately equal to -- within one)
(not approximately equal to -- at least one apart)
These are described in detail in the Software Reference manual under Mathematical Features.
Turbo PMAC Computational Features
251
Turbo PMAC User Manual
Note:
<= and >= are not valid Turbo PMAC comparators. The comparators !> and !<,
respectively, should be used in their place.
Conditions
A condition can be used to control program flow in motion or PLC programs. It is evaluated as either
true or false. It can be used in an IF branching statement or WHILE looping statement. Turbo PMAC
supports both simple and compound conditions.
Note:
A condition in a command line – IF or WHILE – must be surrounded by
parentheses.
Simple Conditions
A simple condition consists of three parts:
{expression} {comparator} {expression}
If the relationship between the two expressions defined by the comparator is valid, then the condition is
true; otherwise, the condition is false. Examples of simple conditions in commands are:
WHILE (1<2)
(always true)
IF (P1>5000)
WHILE (SIN(P2-P1)!>P300/1000)
Note that parentheses are required around the condition itself.
Note:
Unlike in some programming languages, a Turbo PMAC condition may not be
simply a value, evaluated for zero or non-zero (e.g. IF (P1) is not valid). It
must explicitly be a condition with two expressions and a comparator.
Compound Conditions
A compound condition is a series of simple conditions connected by the logical operators AND and OR.
The compound condition is evaluated from the values of the simple conditions by the rules of Boolean
algebra. In the Turbo PMAC, AND has execution precedence over OR (that is, ORs operate on blocks of
ANDed simple conditions). Turbo PMAC will stop evaluating compound AND conditions after one false
simple condition has been found. Examples of compound conditions in command lines are:
IF (P1>-20 AND P1<20)
WHILE (P80=0 OR I120>300 AND I120<400)
IF (Q16!<Q17 AND Q16!>Q18 OR M136<256 AND M137<256)
Note:
The simple conditions contained within a compound condition on a single line
must not be separated by parentheses. For example, IF((P1>-20) AND
(P1<20)) is an illegal condition and will be rejected for illegal syntax.
Single-Line Condition Actions
In Turbo PMAC motion programs (but not in PLC programs) the action(s) to be executed on a true
condition can be put on the same line as the condition itself. In this case, no ENDIF or ENDWHILE is
required to mark the end of the conditional action, and none may be used; the end of the line is
automatically the marker for the end of the conditional action. Examples of this form are:
IF (P1<0) P1=0
WHILE (M11=0) DWELL 10
252
Turbo PMAC Computational Features
Turbo PMAC User Manual
In Turbo PMAC rotary program buffers single-line condition actions are the only types of conditional
statements permitted. Multiple-line conditions are not permitted because it cannot be guaranteed that the
line that must be jumped to will be in the rotary buffer at that time.
Multiple-Line Conditions
In Turbo PMAC PLC programs (but not in motion programs) compound conditions over several program
lines are allowed. The first line of the condition must start with IF or WHILE; following lines of the
condition must start with AND or OR. Simple and compound conditions within a program line are always
evaluated before the conditions on separate lines are combined. Between the conditions on multiple lines,
AND takes precedence over OR. Turbo PMAC will stop evaluating a multi-line AND condition after one
single-line condition has been found false.
An example is:
IF (M11=1 OR M12=1)
AND (M13=1 OR M14=1)
...
Timers
Each active coordinate system (those numbered from 1 to I68+1) has two timer variables running: Isx11 and
Isx12. These two 24-bit registers are general-purpose timers for user program use. Turbo PMAC decrements
them once per servo cycle. Usually a value equal to the time to wait is written scaled in servo cycles. Then the
program waits for the register to become less than 0. The registers will continue to count down until they
23
reach –2 (-8,388,608). They will not roll over back to positive values.
Since these timers have units of servo cycles, and most users will prefer to work in milliseconds, a
23
conversion must be done. To convert from milliseconds to servo cycles, multiply by 2 (8,388,608) and
divide by the value of I10.
The timer variables that belong to a coordinate system can be used by any task on Turbo PMAC,
including motion programs running in other coordinate systems.
Example:
In a PLC program, to turn on an output for a fixed number of milliseconds:
M1=1
I5111=125*8388608/I10
WHILE (I5111>0)
ENDWHILE
M1=0
; Turn on Machine Output 1
; Set timer to 125 msec, in servo cycles
; Wait for counter to count down to zero
; Turn off Machine Output 1
Computational Considerations
When Turbo PMAC is doing calculations in a PLC program, motion program, or on-line, it uses its 48-bit
floating-point format for the intermediate form of the calculation. This gives Turbo PMAC the ability to
convert between its different numerical formats automatically, and enables it to do bit-wise operations on
its P and Q-variables even though they are floating-point values.
The process of converting a number to 48-bit format is very fast and will not be noticeable in most Turbo
PMAC applications. However, skipping the conversion step can help increase Turbo PMAC’s speed and
efficiency for computationally demanding applications. In such applications, using P, Q, and L- (long-)
format M-variables skip the conversion step (they are already in 48-bit floating-point format) and are
computed faster than other variable types.
Turbo PMAC Computational Features
253
Turbo PMAC User Manual
PMAC Program Mathematical Calculations
For All Motion and PLC Programs
Except Fixed-Point Compiled PLC Lines
1
1
Intermediate Working Form:
48-Bit Floating Point
4
4
(PMAC Handles all type conversions automatically)
8
X
12
S
G
N
16
1
2
1
4
1
8
8
Y
... MANTISSA
12
16
EXPONENT
20
20
36
12
24
24
48
I & M-Variables
I & M-Variables
48
36
12
P & Q-Variables, Program Constants
36
P & Q-Variables
Inputs to Calculations
12
Outputs from Calculations
However, when Turbo PMAC is doing calculations with L-variables in a compiled PLC program (PLCC)
it uses a 24-bit fixed-point format for the intermediate form of the calculation. This gives Turbo PMAC
the ability to perform the calculations extremely quickly. L-variable calculations are about 10 times faster
than equivalent calculations using intermediate floating-point data formats.
PMAC Program Mathematical Calculations
Fixed Point Compiled PLC
Program Lines Only
1
4
1
4
Intermediate Working Form:
Signed 24-Bit Integer
8
8
24
12
16
12
S
G
N
16
20
20
S
G
N
24
L-Variables
Inputs to Calculations
254
Program line may only contain L-Variables
and integer constants in range
All values - in,out, and intermediate must stay in range - 8,388,608 to 8,388,607
Calculations are about 10 times faster than
equivalent calculations in floation point
S
G
N
24
L-Variables
Outputs from Calculations
Turbo PMAC Computational Features
Turbo PMAC User Manual
SETTING UP A COORDINATE SYSTEM
Once you have set up your motors, gotten them well tuned, and are doing controlled jogging and homing
search moves, you will want to assemble one or more coordinate systems from the motors so that you can
run motion programs.
Turbo PMAC has several methods of coordinating multiple motions, whether they are all under Turbo
PMAC’s direct control or not. Depending on the user’s situation and needs, one of the coordination
strategies below can be implemented.
What is a Coordinate System?
A coordinate system in Turbo PMAC is a grouping of one or more motors for the purpose of
synchronizing movements. A coordinate system (even with only one motor) can run a motion program; a
motor by itself cannot. Turbo PMAC can have up to 16 coordinate systems, addressed as &1 to &16, in a
flexible fashion (e.g. 16 coordinate systems of 1 or 2 motors each, 1 coordinate system of 9 motors, 4
coordinate systems of 8 motors each, etc.).
In general, if you want certain motors to move in a coordinated fashion, put them in the same coordinate
system. If you want them to move independently of each other, put them in separate coordinate systems.
Different coordinate systems can run separate programs at different times (including overlapping times), or
even run the same program at different (or overlapping) times.
A coordinate system must first be established by assigning motors to axes in that coordinate system. For
simple relationships between motors (actuators) and axes (tool coordinates), this is done with on-line
commands called “axis-definition statements” (see below). For more complex relationships, it is done by
writing special kinematic subroutines that describe the relationship (covered in a following section).
A coordinate system must have at least one motor assigned to an axis within that system, or it cannot run a
motion program, even non-motion parts of it. When a program is written for a coordinate system, if
simultaneous motions are desired of multiple motors, their move commands are simply put on the same
line, and the moves will be coordinated.
What is an Axis?
An axis is an element of a coordinate system. It can be thought of as one of the coordinates of the tool, or
of the mechanics relative to the tool. An axis is often similar to a motor, but not the same thing. An axis is
referred to by letter. There can be up to nine independent axes in a coordinate system, selected from X, Y,
Z, A, B, C, U, V, and W (it is possible to assign multiple motors to an axis). Normally, an axis is defined
by assigning it to a motor with a scaling factor and an offset (X, Y, and Z may be defined as linear
combinations of three motors, as may U, V, and W). The variables associated with an axis defined in this
manner are scaled floating-point values.
Single-Motor Axes
In the vast majority of cases, there will be a one-to-one correspondence between motors and axes. That is,
a single motor is assigned to a single axis in a coordinate system. Even when this is the case, however, the
matching motor and axis are not completely synonymous. The axis is scaled into engineering units, and
deals only with commanded positions. Except for the PMATCH function, calculations go only from axis
commanded positions to motor commanded positions, not the other way around.
Setting Up a Coordinate System
255
Turbo PMAC User Manual
Multiple-Motor Axes
More than one motor may be assigned to the same axis in a coordinate system. This is common in gantry
systems, where motors on opposite ends of the cross-piece are always trying to do the same movement.
By assigning multiple motors to the same axis, a single programmed axis move in a program causes
identical commanded moves in multiple motors. This is commonly done with two motors, but up to eight
motors have been used in this manner with Turbo PMAC. Remember that the motors still have
independent servo loops, and that the actual motor positions will not necessarily be exactly the same.
Coordinating parallel gantry motors in this fashion is in general superior to using a master/slave technique
(which can be done on Turbo PMAC with the position following feature described in the Synchronizing
Turbo PMAC to External Events section of this manual). In the master/slave technique, the actual
trajectory of the master as measured at the encoder, with all of the disturbances and quantization errors,
becomes the commanded trajectory for the slave, whose actual trajectory will have even more errors. The
roughness in the slave motor’s commanded trajectory makes it difficult or impossible to use feedforward
properly, which introduces a lag. True, if the master gets a disturbance, the slave will see it and attempt to
match it, but if the slave gets a disturbance, the master will not see it.
Care must be taken in the startup and homing of gantry motors that have a tight mechanical linkage. In
general, the motors will power up not quite in ideal alignment with each other. The usual procedure is to
do a homing search move on one motor with the second motor slaved to it, followed by an offset back out
far enough that the second motor knows which way it has to go to its home trigger. Next the second motor
is made the master and is told to do a homing search move with the first motor slaved to it. This will leave
the first motor slightly off from its home position; it can now be told to go there with just a J=0 command.
The slaving is then turned off, and the motors are commanded identically through joint axis commands.
Phantom Axes
An axis in a coordinate system can have no motors attached to it (a phantom axis), in which case
programmed moves for that axis cause no movement, although the fact that a move was programmed for
that axis can affect the moves of other axes and motors. For instance, if sinusoidal profiles are desired on a
single axis, the easiest way to do this is to have a second, phantom axis and program circularly interpolated
moves.
Axis Definition
A coordinate system is established by using axis definition statements. An axis is defined by matching a
motor (which is numbered) to one or more axes (which are specified by letter).
Matching Motor to Axis
The simplest axis definition statement is something like #1->X. This simply assigns motor #1 to the Xaxis of the currently addressed coordinate system. When an X-axis move is executed in this coordinate
system, motor #1 will make the move.
Scaling and Offset
The axis definition statement also defines the scaling of the axis’ user units. For instance,
#1->10000X also matches motor #1 to the X axis, but this statement sets 10,000 encoder counts to one
X-axis user unit (e.g. inches or centimeters). This scaling feature is used almost universally. Once the
scaling has been defined in this statement, you can program the axis in engineering units without ever
needing to deal with scaling again.
The statement #1->10000X+20000 also sets the axis zero at 20,000-count (2-user-unit) distance from
the motor zero (home position). This offset is rarely used. Further, an axis definition statement can match
a motor to a linear combination of Cartesian axes (see below), which allows for rotation of a coordinate
system, or orthogonality correction.
256
Setting Up a Coordinate System
Turbo PMAC User Manual
Axis Types
An axis can have several attributes, as specified below. Note that for most axis functions, it does not
matter what type of axis is used, or what letter is given it. However, for some features, only particular axis
names may be used.
Cartesian Axes
A Cartesian axis is one that may be put into a grouping of two or three axes so that movement along an
axis is a linear combination of motion on two or three motors. X, Y, and Z form one set of Cartesian axes;
U, V, and W form the other. In addition, there are several commands (NORMAL, circular move) that can
reference the X, Y, and Z-axes through the use of I, J, and K-vectors, respectively.
To make a Cartesian axis a linear combination of several motors, do so with an extended form of the axis
definition statement. For instance, a 30o rotation of the axes can be received from the motors with the
following axis definition statements:
#1->8660.25X-5000Y
#2->5000X+8660.25Y
In this case, a request for a Y-axis (or an X-axis) move would cause both motors #1 and #2 to move.
Only the X, Y, and Z Cartesian axes may be used for PMAC’s circular interpolation routines, cutter radius
compensation routines, and matrix axis transformation routines. If you want to do circular interpolation on
other axes, you can do it through blended short moves and trigonometry in subroutines.
1) Scaling and Translation
Y
(in)
Motor #2 (10,000
#1 -> 10000X + 40000
#2 -> 10000Y + 20000
cts
)
in
4"
X (in)
2"
cts
Motor #1 (10,000 in )
2) Scaling and Rotation
#1 -> 7071.07X - 7071.07Y
#2 -> 7071.07X + 7071.07Y
cts
Motor #2 (10,000 in )
45
Motor #1 (10,000
cts
)
in
3) Orthogonality Correction
Y (in)
cts
Motor #2 (10,000 in )
#1 -> 10000.00X - 2.91Y
#2 -> 10000.00Y
X (in)
1 arc min
(exaggerated)
Setting Up a Coordinate System
cts
Motor #1 (10,000 in )
257
Turbo PMAC User Manual
Rotary Axes
A rotary axis is one that permits rollover, but cannot be assigned to combinations of motors. A rotary axis
must be named A, B, or C. The rollover is technically a motor function, specified by Ixx27 for Motor xx,
but it can only operate when the motor is assigned to a rotary axis. Rollover permits the motor to take the
shortest path around the rotary range, or the specified direction to the destination, when an absolute axis
move is specified in a program.
Vector Feedrate Axes
A vector feedrate axis is an axis in a coordinate system that figures into the calculations of a feedratespecified move. The time for a feedrate-specified move is calculated as the vector distance for the feedrate
axes (calculated by Pythagorean theorem as the square root of the sum of the squares of the axis distances)
divided by the feedrate itself. If other axes are commanded to move in the same statement, they will be
linearly interpolated over this same computed time.
The default feedrate axes are the Cartesian axes X, Y, and Z. This setting can be changed with the FRAX
(feedrate axis) command. Any axis involved in the arc of a CIRCLE-mode move is automatically a vector
federate axis for that move, even if not explicitly declared as such.
Conversion from Axis to Motor Position
Technically, the conversion from axis (tool-tip) positions to motor (joint, or actuator) positions is known as
the inverse-kinematic conversion. Turbo PMAC automatically converts from the programmed axis
positions (in user engineering units) to motor positions (in counts) every programmed move, or in the case
of segmented moves, every segment of the move.
In the case of axes defined with axis-definition statements, Turbo PMAC simply plugs the axis values into
the equation of the definition statement, and computes the resulting motor position. (The axis-definition
statement is therefore an inverse-kinematic equation.) In the case of axes defined using kinematic
subroutines, Turbo PMAC executes the user-written inverse-kinematic subroutine to compute these (see
the Coordinate-System Kinematic Calculation section in this manual).
Conversion from Motor to Axis Positions
Technically, the conversion from motor (joint, or actuator) positions to axis (tool-tip) positions is known as
the forward-kinematic conversion. There is only one type of calculation in which Turbo PMAC
automatically performs these forward-kinematic calculations. This is in the PMATCH (position-match)
function, which converts from commanded motor positions to commanded axis positions. This is needed
in only a few cases.
First, when a motion program is started with an R (run) or S (step) command, Turbo PMAC automatically
executes a PMATCH command internally to compute the starting axis positions for the first move
calculations. Within a motion program, it normally assumes that the endpoint of the previous move is the
starting point for the subsequent move, and so does not do these calculations each move. However, when
a program is started, it does these calculations because there is a good chance that motors may have been
moved independently (e.g. jog moves, open-loop moves, stopping on an error condition); in other words,
the axes does not know where the motors have gone and motor and axis positions may not match properly.
Second, if the relationship between motor and axis positions is changed inside a motion program (e.g.
changing position-following offset mode, directly writing to the position-bias or axis scaling registers),
issue a PMATCH command before the next programmed move (using the CMD"PMATCH" structure).
Otherwise, the next move will not execute properly.
For axes defined with a simple definition statement, the PMATCH function inverts the equations contained
in the axis-definition statements for the coordinate system, using motor commanded positions, and solves
for axis commanded positions. If more than one motor is assigned to the same axis (e.g. #1->10000X,
#2->10000X), the commanded position of the lower-numbered motor is used in the PMATCH
calculations. For axes in a coordinate system with a user-written forward-kinematic subroutine (see
below), this subroutine is executed automatically for the PMATCH function.
258
Setting Up a Coordinate System
Turbo PMAC User Manual
The PMATCH function assumes that the position referencing – either a homing search move or an absolute
position-sensor read – has been done for each motor in the coordinate system. Each motor has a home
complete status bit that is set true if either has been done, and to check for this, do it at the application
level.
Some users will want to read instantaneous motor position registers (particularly motor actual position) –
Turbo PMAC does not compute instantaneous axes positions automatically – and convert them to axis
positions for position-reporting purposes. These forward-kinematic calculations must be done at the
application level, either in Turbo PMAC (usually in a PLC program) or in the host computer.
Coordinate-System Kinematic Calculations
Turbo PMAC provides structures that enable easy implementation and execution of complex kinematic
calculations. Kinematic calculations are required when there is a non-linear mathematical relationship
between the tool-tip coordinates and the matching positions of the actuators (joints) of the mechanism,
typical in non-Cartesian geometries. They are used most commonly in robotic applications, but can be
used with other types of actuators that are not considered robotic. For example, in 4-axis or 5-axis
machine tools with one or two rotary axes, it is desirable to program the cutter-tip path and let the
controller compute the necessary motor positions.
This capability permits the motion for the machine to be programmed in the natural coordinates of the
tool-tip – usually Cartesian coordinates, whatever the underlying geometry of the machine. The kinematic
routines are embedded in the controller by the integrator, and operate invisibly to the people programming
paths and the machine operators. These routines can be unchanging for the machines, but with
parameterization and/or logic, they can adapt to normal changes such as tool lengths and different endeffectors.
In Turbo PMAC terminology, the tool-tip coordinates are for axes, which are specified by letter, and have
user-specified engineering units. The joint coordinates are for motors, which are specified by numbers,
and have the raw units of “counts”.
Note:
PMAC’s standard axis-definition statements handle linear mathematical
relationships between joint motors and tool-tip axes. This section pertains to the
more difficult case of the non-linear relationships.
The forward-kinematic calculations use the joint positions as input, and convert them to tool-tip
coordinates. These calculations are required at the beginning of a sequence of moves programmed in tooltip coordinates to establish the starting coordinates for the first programmed move. The same type of
calculations can also be used to report the actual position of the actuator in tool-tip coordinates, converting
from the sensor positions on the joints. (The Turbo PMAC forward-kinematic program buffer does not
support this position-reporting functionality, but functionally identical calculations can be used in a PLC
program for this purpose.)
The inverse-kinematic calculations use the tool-tip positions as input, and convert them to joint
coordinates. These calculations are required for the end-point of every move that is programmed in tooltip coordinates, and if the path to the end-point is important, they must be done at periodic intervals during
the move as well.
Setting Up a Coordinate System
259
Turbo PMAC User Manual
Note:
Formal robotic analysis makes a distinction between joint position, and the actuator
positions required for that joint position. Usually, while the two positions are the
same, there are cases, such as when two motors drive a joint differentially, where
there is an important difference. If your system has a distinction between joint and
actuator positions, your kinematic calculations must include this distinction, to go
all the way between actuator positions and tool-tip positions, with joint positions as
an intermediate step. This documentation will refer to only joint positions, although
this could technically refer to actuator positions in some applications.
Creating the Kinematic Program Buffers
Turbo PMAC implements the execution of kinematic calculations through special forward-kinematic and
inverse-kinematic program buffers. Each coordinate system can have one of each of these program
buffers, and the algorithms in them can be executed automatically at the required times, called as
subroutines from the motion program.
Creating the Forward-Kinematic Program
The on-line OPEN FORWARD command opens the forward-kinematic buffer for the addressed coordinate
system for entry. The on-line CLEAR command erases any existing contents of that buffer. Subsequently,
any program command sent to Turbo PMAC that is legal for a PLC program (except ADDRESS, CMDx,
and SENDx) will be entered into the open buffer. The on-line CLOSE command stops entry into the
buffer.
Before any execution of the forward-kinematic program, Turbo PMAC will automatically place the present
commanded motor positions for each Motor xx in the coordinate system into global variable Pxx. These
are floating-point values, with units of counts. The program can then use these variables as the “inputs” to
the calculations.
After any execution of the forward-kinematic program, Turbo PMAC will take the values in Q1 – Q9 for
the coordinate system in the user’s engineering units, and copy these into the 9 axis target position
registers for the coordinate system. There they are used as the starting positions for the first programmed
move that follows. The following table shows the axis whose position each variable affects, and the
suggested M-variable number for each of these registers (listed for debugging purposes).
AxisPosition QVariable
Axis
Letter
Q1
Q2
Q3
A
B
C
Target
Register
Suggested
M-Variable
Msx41
Msx42
Msx43
AxisPosition QVariable
Axis
Letter
Q4
Q5
Q6
U
V
W
Target
Register
Suggested
M-Variable
Msx44
Msx45
Msx46
AxisPosition QVariable
Axis
Letter
Q7
Q8
Q9
X
Y
Z
Target
Register
Suggested
M-Variable
Msx47
Msx48
Msx49
The basic purpose of the forward-kinematic program, then, is to take the joint-position values found in P1
– P32 for the motors used in the coordinate system, compute the matching tip-coordinate values, and place
them in variables in the Q1 – Q9 range.
It is a good idea to check in your forward-kinematics program to make sure that a position reference has
been properly established for each motor, either through a homing search move or an absolute position
read. This can be done by evaluating the home complete status bit for each motor; the run-time error bit
can be set if the referencing has not been done (see example).
260
Setting Up a Coordinate System
Turbo PMAC User Manual
Reserved Variables
If kinematic calculations are used in a system, the global variables P1 – P32 and the coordinate-system
variables Q1 – Q10 should not be used for any other purposes, because Turbo PMAC will write to these
variables automatically in executing the kinematic routines. (Q10 is used to distinguish between inversekinematic calculations that involve velocity calculations and those that do not, as explained below.) If
inverse-kinematic calculations involving PVT-mode moves are used, additionally the global variables
P101-P132 and the coordinate-system variables Q11 – Q19 should not be used for any other purposes,
because Turbo PMAC will write to these variables automatically in executing the velocity portions of the
inverse-kinematic routines.
Example
Take the example of a 2-axis shoulder-elbow robot, with an upper-arm length (L1) of 400mm, and a lowerarm length (L2) of 300mm. Both the shoulder joint (A) and the elbow joint (B) have resolutions of 1000
counts per degree. When both joints are at their zero-degree positions, the two links are both extended
along the X-axis. The forward-kinematic equations are:
X = L1 cos( A ) + L2 cos( A + B )
Y = L1 sin( A ) + L2 sin( A + B )
To implement these equations in a Turbo PMAC forward-kinematic program for Coordinate System 1 that
converts the shoulder angle in Motor 1 and the elbow angle in Motor 2 to the X and Y tip coordinates in
millimeters, the following setup and program could be used:
; Setup for program
I15=1
&1
M145->Y:$0000C0,10,1
M245->Y:$000140,10,1
M5182->Y:$00203F,22,1
Q91=400
Q92=300
Q93=1000
;
;
;
;
;
;
;
;
Trig calculations in degrees
Address CS 1
Motor 1 home complete bit
Motor 2 home complete bit
CS 1 run-time error bit
L1
L2
Counts per degree for A and B
; Forward-kinematic program buffer for repeated execution
&1 OPEN FORWARD
; Forward kinematics for CS 1
CLEAR
; Erase existing contents
IF (M145=1 AND M245=1) ; Properly position referenced?
Q7=Q91*COS(P1/Q93)+Q92*COS((P1+P2)/Q93) ; X position
Q8=Q91*SIN(P1/Q93)+Q92*SIN((P1+P2)/Q93) ; Y position
ELSE
; Not valid; halt operation
M5182=1
; Set run-time error bit
CLOSE
Setting Up a Coordinate System
261
Turbo PMAC User Manual
The forward-kinematic program must calculate the axis positions for all of the axes in the coordinate
system, whether or not all of the motor positions are calculated in the inverse-kinematic program (see
below). For instance, if this arm had a vertical axis at the tip with a normal axis definition statement in
C.S. 1 of #3->100Z (100 counts per millimeter – a linear relationship between motor and axis), the above
program would still need to perform the forward-kinematic calculation for this motor/axis with a line such
as Q9=P3/100.
Note:
If the forward-kinematic algorithm is not correct, and does not yield a true
mathematical inverse of the inverse-kinematic algorithm, there will be a sudden and
potentially dangerous jump at the beginning of the first move executed after the
forward kinematic algorithm is executed. Make sure early in development that the
Ixx11 fatal following error limits are set as tight as possible to ensure that any large
errors will cause a trip and not result in violent motion.
Iterative Solutions
Some systems, particularly parallel-link mechanisms such as Stewart platforms (hexapods), do not have
reasonable closed-form solutions for the forward-kinematic equations, and require iterative numerical
solutions. Typically, these cases are handled by a looping WHILE … ENDWHILE construct in the
forward-kinematic program. Do not permit indefinite looping – if the solution does not converge in the
expected number of cycles, the program should be stopped (see the inverse-kinematic equations, below,
for examples of how to stop the program).
In this case, it is best to leave the I11 program-calculation delay variable at its default value of 0, so the
calculations can take as long as needed. If I11 is greater than 0, and the forward-kinematic calculations
plus the first move calculations do not finish within I11 msec, Turbo PMAC will stop the program with a
run-time error. In any case, if the forward-kinematic calculations take more than about 25 msec, it is
possible to trip the watchdog timer.
Position-Reporting Forward Kinematics
Another use of forward-kinematic calculations is for the position reporting function, reading actual joint
positions at any time, and converting them to tip positions for reporting. The forward-kinematic program
buffer on Turbo PMAC does not support this function. (Using the program for both initial-position
calculations and position reporting could lead to potential overlapping use and register conflicts.)
If the application requires the Turbo PMAC to do forward-kinematic calculations for position reporting as
well as for establishing initial tip position, the position-reporting calculations should be put into a PLC
program. The following PLC program could be used for the position-reporting function of the example
“shoulder-elbow” robot:
; M-variable definitions for actual position registers
M162->D:$8B
M262->D:$10B
; Motor 1 actual position
; Motor 2 actual position
; Forward-kinematic PLC program buffer for position reporting
OPEN PLC 10
CLEAR
P51=M162/(I108*32*Q93)
P52=M262/(I208*32*Q93)
Q27=Q91*COS(P51)+Q92*COS(P51+P52)
Q28=Q91*SIN(P51)+Q92*SIN(P51+P52)
CLOSE
262
;
;
;
;
;
;
Forward kinematics for CS 1
Erase existing contents
Actual A position (deg)
Actual B position (deg)
Actual X position
Actual Y position
Setting Up a Coordinate System
Turbo PMAC User Manual
Creating the Inverse-Kinematic Program
The on-line OPEN INVERSE command opens the inverse-kinematic buffer for the addressed coordinate
system for entry. The on-line CLEAR command erases any existing contents of that buffer. Subsequently,
any math or logic program command sent to Turbo PMAC that is legal for a PLC program (this does not
include ADDRESS, DISPLAY, CMDx, or SENDx) will be entered into the open buffer. The on-line CLOSE
command stops entry into the buffer.
Before any execution of the inverse-kinematic program, Turbo PMAC will place the present axis target
positions for each axis in the coordinate system into variables in the range Q1 – Q9 for the coordinate
system. These are floating-point values, in engineering units. The program can then use these variables as
the “inputs” to the calculations. The following table shows the variable for each axis:
AxisPosition QVariable
Axis
Letter
AxisPosition QVariable
Axis
Letter
AxisPosition QVariable
Axis
Letter
Q1
Q2
Q3
A
B
C
Q4
Q5
Q6
U
V
W
Q7
Q8
Q9
X
Y
Z
After any execution of the inverse-kinematic program, Turbo PMAC will read the values in those variables
Pxx (P1 – P32) that correspond to Motors xx in the coordinate system with axis-definition statements of
#xx->I. These are floating-point values, and Turbo PMAC expects to find them in the raw units of
“counts.” Turbo PMAC will automatically copy these values into the target position registers for these
motors (suggested M-variable Mxx63), where they are used for the fine interpolation of these motors.
There can be other motors in the coordinate system that are not defined as inverse-kinematic axes; these
motors get their position values directly from the axis-definition statement and are not affected by the
inverse-kinematic program.
The basic purpose of the inverse-kinematic program, then, is to take the tip-position values found in Q1 –
Q9 for the axes used in the coordinate system, compute the matching joint-coordinate values, and place
them in variables in the P1 – P32 range.
Example
Continuing with our example of the two-axis shoulder-elbow robot, and for simplicity’s sake limiting
ourselves to positive values of B (the right-armed case), we can write our inverse-kinematic equations as
follows:
 X 2 + Y 2 − L2 − L2 
1
2
B = + cos −1 

2 L1 L2



A + C = a tan 2( Y , X )
 X 2 + Y 2 + L2 − L2
1
2
C = + cos −1 
 2L X 2 + Y 2
1

A = ( A+C )−C




(X, Y)
Y
2
2
L2
√(X +Y )
B
C
L1
A
X
Setting Up a Coordinate System
263
Turbo PMAC User Manual
To implement these equations in a Turbo PMAC inverse-kinematic program for Coordinate System 1 that
converts the X and Y tip coordinates in millimeters to the shoulder angle in Motor 1 and the elbow angle in
Motor 2, the following program could be used. System constants Q91, Q92, and Q93 are the same as for
the above forward kinematic program.
; Setup for program
&1
#1->I
#2->I
M5182->Y:$00203F,22,1
; Motor 1 assigned to inverse kinematic axis in CS 1
; Motor 2 assigned to inverse kinematic axis in CS 1
; CS 1 run-time error bit
; Pre-compute additional system constants
Q94=Q91*Q91+Q92*Q92
Q95=2*Q91*Q92
Q96=Q91*Q91-Q92*Q92
; L1^2 + L2^2
; 2*L1*L2
; L1^2 – L2^2
; Inverse-kinematic algorithm to be executed repeatedly
&1 OPEN INVERSE
; Inverse kinematics for CS 1
CLEAR
; Erase existing contents
Q20=Q7*Q7+Q8*Q8
; X^2+Y^2
Q21=(Q20-Q94)/Q95
; cos(B)
IF (ABS(Q21)<0.9998)
; Valid solution w/ 1 deg margin?
Q22=ACOS(Q21)
; B (deg)
Q0=Q7
; X into cos argument for ATAN2
Q23=ATAN2(Q8)
; A+C = ATAN2(Y,X)
Q24=ACOS((Q20+Q96)/(2*Q91*SQRT(Q20))) ; C (deg)
Q25=Q23-Q24
; A (deg)
P1=Q25*Q93
; Motor 1 = 1000A
P2=Q22*Q93
; Motor 2 = 1000B
ELSE
; Not valid, halt operation
M5182=1
; Set run-time error bit
ENDIF
CLOSE
Notes on the example:
•
•
•
•
By choosing the positive arc-cosine solutions, we are automatically selecting the right-armed case. In
a more general solution, we would have to choose whether the positive or negative is used, based on
some criterion.
Increased computational efficiency could be obtained by combining more operations into single
assignment statements. Calculations were split out here for clarity’s sake.
This example does not use the substitution macros permitted by the Executive program to substitute
meaningful names for variables. Use of these substitution macros in complex applications is strongly
encouraged.
This example stops the program for cases in which no inverse kinematic solution is possible. It does
this by setting the “run-time error” status bit for the coordinate system, which causes Turbo PMAC to
halt motion program execution and issue the Abort command. Other strategies may be used to cope
with this problem.
If this robot had a vertical axis at the tip, the relationship between motor and axis could be defined with a
normal linear axis-definition statement (e.g. #3->100Z for 100 counts per millimeter), and the motor
position would be calculated without the special inverse-kinematic program. Alternately, the motor could
be defined as an inverse-kinematic axis (#3->I) and the motor position could be calculated in the inversekinematic program (e.g. Q3=Q49*100 to set Motor 3 position from the Z-axis with 100 counts per unit).
264
Setting Up a Coordinate System
Turbo PMAC User Manual
Rotary Axis Rollover
If a rotary inverse-kinematic axis in the system has the capability to “roll over,” the inverse-kinematic
program must handle the rollover calculations explicitly. The automatic rollover capability of the A, B,
and C axes with Ixx27 is not available for inverse-kinematic axes. The key to handling rollover properly is
to take the difference between the new and the old values and make sure that this difference is in the +180o
range. This can be done in Turbo PMAC with the ‘%’ modulo (remainder) operator. This difference is
then added to the old value. Mathematically, the equations are:
(
)
∆θ = θ new−temp − θ old %(− 180 )
θ new = θ old + ∆θ
When the modulo operation is done in Turbo PMAC with a negative operand ‘-n’ (such as –180), the
result is always in the +n range.
For example, if the A-axis in the above example had the capability of rolling over, the line Q25=Q23Q24 could be replaced with:
Q25=P1/Q93+(Q23-Q24-P1/Q93)%-180
; Handle rollover cases
The value (P1/Q93) is θold, from the previous cycle of the inverse kinematics or initially from the forward
kinematics; and value (Q23-Q24) is θnew-temp, both in degrees.
Velocity Calculation Flag
In every move mode other than PVT mode, Turbo PMAC sets the variable Q10 for the coordinate system
to 0 automatically as a flag to the inverse-kinematic program not to compute velocity values. If planning
to use both PVT mode and other modes, evaluate Q10 explicitly in the inverse-kinematic program (see
below).
Iterative Solutions
Some robot geometries do not have closed-form inverse-kinematic solutions, and require iterative
numerical solutions. Typically these cases are handled by a looping WHILE … ENDWHILE construct in
the inverse-kinematic program. Multiple executions of the WHILE loop inside the inverse-kinematic
program do not disable blending as they would inside the main motion program (due to the double jumpback rule), but excessive iterations can cause the calculations not to be done within the required time.
This will cause a run-time error, aborting the program automatically.
Inverse-Kinematic Program for PVT Mode
The Turbo PMAC can also support the conversion of velocities from tip space to joint space in the inversekinematic program to enable the use of PVT mode with kinematic calculations. With PVT-mode moves,
the position calculations are done just as for any other move mode. An additional set of velocityconversion calculations must also be done.
When executing PVT-mode moves with kinematics active (Isx50 = 1), Turbo PMAC will automatically
place the commanded axis velocity values from the PVT statements into variables Q11 – Q19 for the
coordinate system before each execution of the inverse-kinematic program. These are signed floatingpoint values in the engineering velocity units defined by the engineering length/angle units and the
coordinate system’s Isx90 time units (e.g. mm/min or deg/sec). The following table shows the variable
used for each axis:
Axis-Velocity
Q-Variable
Axis
Letter
Axis-Velocity
Q-Variable
Axis
Letter
Axis-Velocity QVariable
Axis Letter
Q11
Q12
Q13
A
B
C
Q14
Q15
Q16
U
V
W
Q17
Q18
Q19
X
Y
Z
Setting Up a Coordinate System
265
Turbo PMAC User Manual
Turbo PMAC will also set Q10 to 1 in this mode as a flag to the inverse-kinematic program that it should
use these axis (tip) velocity values to compute motor (joint) velocity values.
In this mode, after any execution of the inverse-kinematic program, Turbo PMAC will read the values in
those variables P1xx (P101 – P132) for each Motor xx in the coordinate system defined as inversekinematic axes (#xx->I). These are floating-point values, and Turbo PMAC expects to find them in
units of counts per Isx90 milliseconds. Turbo PMAC will use them as motor (joint) velocity values along
with the position values in Pxx to create a PVT move for the motor.
For PVT moves, then, the inverse-kinematic program not only must take the axis (tip) position values in
Q1 – Q9 and convert them to motor (joint) position values in P1 – P32; but also it must take the axis (tip)
velocity values in Q11 – Q19 and convert them to motor (joint) velocity values in P101 – P132.
Technically, the velocity conversion consists of the solution of the “inverse Jacobian matrix” for the
mechanism.
Example
Continuing with the “shoulder-elbow” robot of the above examples, the equations for joint velocities as a
function of tip velocities are:
L cos( A + B ) X& + L2 sin( A + B )Y&
A& = 2
L1 L2 sin B
− L1 cos A − L2 cos( A + B ) X& + − L1 sin A − L2 sin( A + B ) Y& − XX& − YY&
B& =
=
L1 L 2 sin B
L1 L2 sin B
[
]
[
]
The angles A and B have been computed in the position portion of the inverse-kinematic program. Note
that the velocities become infinite as the angle B approaches 0 degrees or 180 degrees. Since in our
example we are limiting ourselves to positive values for B, we will trap any solution with a value of B less
than 1o or greater than 179o (sin B < 0.0175) as an error.
&1
OPEN INVERSE
CLEAR
{Position calculations from above}
IF (Q10=1)
Q26=SIN(Q25)
IF (Q26>0.0175)
Q27=Q91*Q92*Q26
Q28=COS(Q25+Q22)
Q29=SIN(Q25+Q22)
Q30=(Q92*Q28*Q17+Q92*Q29*Q18)/Q27
Q31=(-Q7*Q17-Q8*Q18)/Q27
P101=Q30*Q93
P102=Q31*Q93
ELSE
M5182=1
ENDIF
ENDIF
CLOSE
;
;
;
;
;
;
;
;
;
;
;
;
PVT mode?
sin(B)
Not near singularity?
L1*L2*sinB
cos(A+B)
sin(A+B)
dA/dt
dB/dt
#1 speed in cts/(Isx90 msec)
#2 speed in cts/(Isx90 msec)
Near singularity
Set run-time error bit
Note that in this case the check to see if B is near 0o or 180o is redundant because we have already done
this check in the position portion of the inverse-kinematic algorithm. This check is shown here to illustrate
the principle of the method. In this example, a run-time error is created if too near a singularity; other
strategies are possible.
266
Setting Up a Coordinate System
Turbo PMAC User Manual
Coordinate System Transformations with Kinematics
A coordinate system established with kinematic algorithms can still use the on-line {axis}= and
buffered PSET translations of the axis programming origins, just as for a coordinate system with standard
axis definition statements. When one of these commands is executed, Turbo PMAC executes the inverse
kinematic algorithm to calculate a new position bias register (suggested M-variable Mxx64) for each
affected motor in the coordinate system. This is done invisibly to the user; the effect is to offset the
programming origin for the axis. Axis transformations for the X, Y, and Z axes may also be used with the
kinematic algorithms.
Executing the Kinematic Programs
Once the forward-kinematic and inverse-kinematic program buffers have been created for a coordinate
system, Turbo PMAC will execute them automatically at the proper times once the kinematic calculations
have been enabled by setting coordinate system I-variable Isx50 to 1. No modification to a motion
program is required for access to the kinematic programs at the proper time.
The forward-kinematic program is executed automatically each time an R (run) or S (step) command is
given to the coordinate system if Isx50 is 1. This is done to ensure that the starting tip (axis) position is
correct for the calculation of the initial move, even if joint (motor) moves, such as jogs, have been done
since the last programmed move. The forward-kinematic program is also executed automatically each
time a PMATCH command is given to the coordinate system if Isx50 is 1.
(With Isx50 = 0 and normal axis definition statements, Turbo PMAC executes this same function by
mathematically inverting the equations of the axis-definition statements to derive the starting axis positions
from present commanded motor positions. The axis-definition statements are technically inversekinematic equations, so their mathematical inverse forms the forward-kinematic equations. Because the
standard axis-definition statements are limited to mathematically linear equations, in general their inverse
can be derived automatically.)
The inverse-kinematic program is executed automatically each time Turbo PMAC computes new axis
positions during the execution of a motion program. This occurs at the end-point of each programmed
move block for non-segmented moves, such as those in RAPID mode. It occurs at the end of each
intermediate segment – every Isx13 milliseconds – for segmented moves (LINEAR and CIRCLE-mode
moves with Isx13 > 0).
(With normal axis definition statements, Turbo PMAC executes this same function by using the equations
of the axis definition statements to derive motor positions from axis positions.)
Setting Up a Coordinate System
267
Turbo PMAC User Manual
When the inverse-kinematic program is executed only at programmed end-points, as in RAPID mode, all
interpolation occurs in joint space. In this case, the path of the tip from point to point is not well defined if
the programmed end-points are far apart, and in general it will not be a straight line.
When the inverse-kinematic program is executed at each intermediate segment boundary, the coarse
interpolation (segmentation) is done in tip space, so the path is well defined. After the conversion of the
segment coordinates to joint positions, the fine interpolation between segment boundaries is done in joint
space as a cubic spline, but with the segments close together (typically 5 to 20 msec each) any deviations
from the ideal tip path are negligible.
If the special lookahead buffer for the coordinate system is active (LINEAR or CIRCLE-mode moves with
the lookahead buffer defined for the coordinate system, Isx13 > 0, and Isx20 > 0), the internal spline
segments computed for the joints (motors) are entered into the lookahead buffer automatically. Here they
are continually checked against position, velocity, and acceleration limits for each motor. This permits
Turbo PMAC to check and correct automatically for the motion anomalies that occur near singularities, so
you do not need to do so.
Coordinate System Time-Base
Each coordinate system has its own time base that helps control the speed of interpolated moves in that
coordinate system. Turbo PMAC’s interpolation routines increment an “elapsed-time” register every servo
cycle. While the true time for the servo cycle is set in hardware for the entire system (by jumpers E98,
E29-E33, and E3-E6 on a Turbo PMAC1, by hardware-control variables I7m00, I7m01, and I7m02 for a
Turbo PMAC2 without MACRO, or by I6800, I6801, and I6802 for a Turbo PMAC2 with MACRO) and
does not change for a given application, the value of time added to the “elapsed-time” register for a
coordinate system each servo cycle is just a number in a memory register. It does not have to match the
true physical time for the cycle.
The units for the time base register are such that 223 (8,388,608) equals 1 millisecond. The default value
for the time-base register is equal to the value of I10. The factory default value for I10 of 3,713,707
represents the default physical servo cycle time of 442 microseconds.
If the value of the time base register is changed from I10, interpolated moves will move at a different
speed from that programmed. Many people call this capability feedrate override. Note that the physical
time does not change, so servo loop dynamics remain unchanged.
268
Setting Up a Coordinate System
Turbo PMAC User Manual
Each coordinate system has a variable Isx93 that contains the address of the register that the coordinate
system uses for its time base. With the default value of Isx93, the coordinate system gets its time base
information from a register that is set by % commands from the host computer. A %100 command puts a
value equal to I10 in this register; a %50 command puts a value equal to I10/2 in this register.
Note:
A motor that is not assigned to any coordinate system uses Coordinate System 1’s
time base value for executing its jogging and homing moves. To use another base
value, assign the motor to a different coordinate system, even if there is no need to
write a motion program to run this motor.
Regardless of the source of the time base information, a % query command cause PMAC to report back the
value of the present time base expressed as a percentage of I10.
Time base information can come from other sources. The most common alternative to command-sourced
time base is external frequency-sourced time base, in which the time base value is proportional to the
frequency of a master encoder. This provides a powerful position-synchronized slaving mechanism that is
commonly called electronic cam.
See instructions for using an external time base, in the Synchronizing Turbo PMAC to External Events
section of this manual.
Setting Up a Coordinate System
269
Turbo PMAC User Manual
270
Setting Up a Coordinate System
Turbo PMAC User Manual
WRITING AND EXECUTING MOTION PROGRAMS
Motion programs are Turbo PMAC’s chief mechanism for describing the desired motion with the
associated math, logic, and I/O operations. They provide a simple, yet powerful and flexible means for
describing the motion and operations synchronous to that motion.
Turbo PMAC can hold up to 224 motion programs at one time. Any coordinate system can run any of
these programs at any time, even if another coordinate system is already executing the same program.
Turbo PMAC can run as many motion programs simultaneously as there are coordinate systems defined
on the card (up to 16). A motion program can call any other motion program as a subprogram, with or
without arguments.
Turbo PMAC’s motion program language is perhaps best described as a cross between a high-level
computer language like BASIC or Pascal, and G-Code (RS-274) machine tool language. In fact, it can
accept straight “G-Code” programs directly, provided it has been set up properly. It has the
computational and logical constructs of a computer language, and move specification constructs very
much like machine tool languages. Numerical values in the program can be specified as constants or
expressions.
Sequenced Motion Program Execution
A powerful feature of Turbo PMAC motion programs is their automatic sequencing of calculations in
synchronization with the programmed moves. Unlike many motion-programming languages, it is not
necessary to include in your program explicit structures to wait for the end of a programmed move.
Instead, the Turbo PMAC’s operating system automatically monitors the progress of the programmed
move execution and triggers pending calculations (motion, I/O, and/or logic) at the end of a programmed
move. This greatly simplifies the writing of motion-program sequences.
A key implication of this scheme is that calculations in motion programs occur only at the boundaries of
programmed moves. If you have calculations that you want to occur at other times, these calculations
should be executed in Turbo PMAC PLC programs instead. See the Writing and Executing PLC
Programs section of this manual for details.
Flow Control
In a motion program, Turbo PMAC has WHILE loops and IF...ELSE branches that control program
flow. These constructs can be nested indefinitely. In addition, there are GOTO statements, with either
constant or variable arguments (the variable GOTO can perform the same function as a Case statement).
GOSUB statements (constant or variable destination) allow subroutines to be executed within a program.
CALL statements permit other programs to be entered as subprograms. Entry to the subprogram does not
have to be at the beginning – the statement CALL 20.15000 causes entry into Program 20 at line
N15000. GOSUBs and CALLs can be nested only 15 deep.
G-Codes
To handle machine-tool-style G-codes, which provide direct access to part programs created by
CAD/CAM programs, Turbo PMAC treats a Gnn statement as CALL 1000.nn000. The following
values on the line (e.g. X1000) can be treated as parameters to be passed, as for a canned cycle, or the
subprogram can execute without arguments, return, and execute the rest of the line (as for a modal Gcode). The machine tool designer writes Program 1000 to implement the G-codes as he wishes, allowing
customization and enhancements. Delta Tau provides a sample file implementing all of the standard Gcodes. M, S, T, and D codes are similarly implemented.
Writing and Executing Motion Programs
271
Turbo PMAC User Manual
Modal Commands
Many of the statements in Turbo PMAC motion programs are modal in nature. These include move
modes, which specify what type of trajectory a move command will generate; this category includes
LINEAR, RAPID, CIRCLE, PVT, and SPLINE. Moves can be specified either incrementally (distance)
or absolutely (location) – individually selectable by axis – with the INC and ABS commands. Move
times (TA, TS, and TM) and/or speeds (F), are implemented in modal commands. Modal commands can
precede the move commands they are to affect, or they can be on the same line as the first of these move
commands.
Move Commands
The move commands themselves consist of a one-letter axis-specifier followed by one or two values
(constant or expression). All axes specified on the same line will move simultaneously in a coordinated
fashion on execution of the line; consecutive lines execute sequentially (with or without stops in between,
as determined by the mode). Depending on the modes in effect, the specified values can mean,
destination, distance, and/or velocity (see Trajectory Features section).
Motion Program Trajectories
Among Turbo PMAC’s outstanding characteristics are the power and flexibility of its trajectory
generation algorithms. These algorithms allow a great variety of difficult maneuvers to be performed,
and permit you to make your own tradeoffs between ease of use and degree of control. It is important to
remember that these trajectories are a series of commanded positions only. It is up to the servo loops for
each axis to try to make the actual positions match the commanded positions. All the times, speeds,
distances, and profiles discussed in this section are commanded ones, unless otherwise noted.
Linear Blended Moves
The easiest class of moves to make is the linear blended move category. In this type of move, an axis
moves toward the target position at a designated speed, accelerating to and decelerating from this speed in
a controlled fashion. If more than one move is specified in succession with no pause in between, the first
move will blend into the second with the same type of controlled acceleration as is done to and from a
stop.
Linear blended move mode is the default mode for motion programs. If in another move mode, the
program can be put into this mode with the LINEAR statement. The program can be taken out of
LINEAR mode with another move mode statement (e.g. CIRCLE1, CIRCLE2, RAPID, PVT, SPLINE).
It is good programming practice to declare the LINEAR mode in each program, and not rely on the
default. The LINEAR statement is equivalent to the RS-274 G-Code G01.
Position or Distance Specification
The destination point of a linear-mode move is specified in the move command itself (e.g. X10Y20). The
commanded destination for each axis can be specified as a position relative to the programming origin (if
the axis is in ABS absolute mode) or with a distance from the last commanded position (if the axis is in
INC incremental mode). The program specifies one of these; Turbo PMAC automatically calculates the
other.
Feedrate or Move-Time Specification
You can specify either the target velocity (feedrate) for the move, with an F command, or the move time
with the TM command. If F is specified, the move time is calculated, and if TM is specified, the feedrate is
calculated. The relationship between the two values is reciprocal for a given move distance. Move-time
and feedrate values are modal; they affect all subsequent linear (and circle) mode moves until another F
or TM value is specified in the program.
272
Writing and Executing Motion Programs
Turbo PMAC User Manual
The units of the TM time are milliseconds; the units of the F velocity are the user length (or angle) units of
the feedrate axes divided by the time units as defined by coordinate system variable Isx90 in milliseconds.
If Isx90 is at the default value of 1000, the F units are length units per second; if Isx90 is set to 60,000,
the F units are length units per minute.
If no F or TM value is specified after power-up/reset, the value of Isx89 is used for the moves as a feedrate
value. Any F value specified in a program is compared to maximum feedrate parameter Isx98; if greater
than this parameter, Isx98 is used instead.
Note:
Feedrate is a magnitude and should therefore always be a positive number. A
negative feedrate will cause the motion to be opposite of what is defined as
positive in the Coordinate System definition.
Vector Feedrate Axes
If a multi-axis move is specified by feedrate (and not time), theres is the further flexibility of specifying
which axes control the vector feedrate using the FRAX command (on-line or buffered) and velocity is
apportioned among these axes so that their vector combination (root of sum of squares) is the specified
velocity. Turbo PMAC calculates the move time as the vector distance of the feedrate axes divided by the
programmed feedrate. Therefore, each axis’ velocity does not have to be computed individually for each
different angle of movement. If a simultaneous move is requested of a non-feedrate axis, that move is
completed in the same time as that computed for the feedrate axes. The default feedrate axes for a
coordinate system are the X, Y, and Z-axes.
If there are other axes (non-feedrate axes) commanded on the same line, Turbo PMAC compares the
move time computed for the vector feedrate axes to the move time derived by taking the greatest distance
of a non-feedrate axis divided by the coordinate system’s alternate feedrate parameter Isx86. Whichever
of these move times is the longest is used for all axes.
Example Vector Feedrate Calculations (Isx86=40)
INC
FRAX(X,Y)
X3 Y4 F10
INC
FRAX(X,Y)
X3 Y4 Z12 F10
INC
FRAX(X,Y,Z)
X3 Y4 Z12 F10
INC
FRAX(X,Y,Z)
C10 F10
Vect Dist = SQRT(32 + 42) = 5
Move Time = 5/10 = 0.5
Vx = 3/0.5 = 6
Vy = 4/0.5 = 8
Vect Dist = SQRT(32 + 42) = 5
Vect Move Time = 5/10 = 0.5
Non-Vect Dist = 12
Non-Vect Move Time = 12/40 = 0.3
Vx = 3/0.5 = 6
Vy = 4/0.5 = 8
Vz = 12/0.5 = 24
Vect Dist = SQRT(32+42+122)=13
Move Time = 13/10 = 1.3
Vx = 3/1.3 = 2.31
Vy = 4/1.3 = 3.08
Vz = 12/1.3 = 9.23
Vect Dist = 0, Non-Vect Dist = 10
Move Time = 10/40 = 0.25
Vc = 40
Motor Velocity Limits
Turbo PMAC provides a velocity limit parameter Ixx16 for each Motor xx that can be used to
automatically limit the commanded velocity in linear-mode moves even if the motion program requests a
higher rate. The details of how this limiting function operates are dependent on the mode of operation.
Writing and Executing Motion Programs
273
Turbo PMAC User Manual
This velocity limiting is active either if segmentation mode is not active (Isx13 = 0), in which case
circular interpolation and cutter-radius compensation are not permitted, or if segmentation mode is active
(Isx13 > 0) and the special lookahead buffer is active (Isx20 > 0, defined lookahead buffer).
If the velocity limits are active, Turbo PMAC compares the motor velocity magnitudes requested by the
motion program to the Ixx16 limit for each motor. If the request for any motor exceeds the limit, the
move time is extended so that motor will not exceed its limit; this automatically slows the other motors in
the coordinate system in proportion so that the relationship between motors (path in space) is maintained.
Acceleration Parameters
The acceleration to and from velocity can be constant, providing trapezoidal velocity profiles; it can be
linearly varying, yielding S-curve velocity profiles; or it can be a combination of the two. Specify the
time for the full acceleration (TA – default parameter is coordinate system I-variable Isx87) and the time
in each half of the S (TS – default parameter Isx88). If the specified TA time is less than twice the
specified TS time, the TA time used will be twice TS (users who want pure S-curve acceleration can just
set TA to 0). Turbo PMAC can use only integer values for TA and TS. If a non-integer value is
specified, PMAC will round it to the nearest integer before using it in trajectory calculations.
274
Writing and Executing Motion Programs
Turbo PMAC User Manual
AUTOMATIC
"S" CURVE ACCELERATION
SPECIFY t
ACCEL
AND t
S
AND ACCELERATION LIMIT
ACCELERATION
Parabolic
Sinusoidal For Comparison
tL =0
t ACCEL = 2tS
2a
2
a
tL = tS
3
2
tS = 0
t ACCEL = 2tS
a
0
4
1
TIME
VELOCITY
Parabolic (sinusoidal)
for comparison
t L = t ACCEL - 2tS
tL
t
S
4
t
S
MAX.
tL = 0
t ACCEL = 2t S
3
t L = tS
2
tS = 0
t ACCEL = t L
1
TIME
0
t
ACCEL
t ACCEL = 2tS +t L
Acceleration Limits
Turbo PMAC provides an acceleration limit parameter Ixx17 for each Motor xx that can be used to
automatically limit the commanded rate of acceleration in linear-mode moves even if the motion program
requests a higher rate. The details of how this limiting function operates are dependent on the mode of
operation.
• Segmentation mode not active (Isx13 = 0), special lookahead buffer not active (Isx20 = 0). This is the
simplest mode of operation. Circular interpolation is not permitted in this mode. Ixx17 acceleration
limits are active in this mode, but if the acceleration would have to be extended over more than a full
programmed move in order to observe the limit, the limit will be violated.
• Segmentation mode active (Isx13 > 0), special lookahead buffer not active (Isx20 = 0). This mode
permits circular interpolation and cutter-radius compensation, and to have circle-mode and linearmode moves to blend together, but the Ixx17 acceleration limits are not active in this mode.
• Segmentation mode active (Isx13 > 0), special lookahead buffer active (Isx20 > 0, defined lookahead
buffer). This is the most sophisticated mode of operation. This mode permits circular interpolation
and cutter-radius compensation, to have circle-mode and linear-mode moves blend together, to have
the Ixx17 acceleration limits active for both linear and circle-mode moves, and to be able to extend
the acceleration over multiple programmed moves if necessary.
Writing and Executing Motion Programs
275
Turbo PMAC User Manual
This section covers the acceleration limiting algorithms of Mode 1 only. The acceleration limiting
function of the special lookahead buffer (Mode 3) is discussed in a manual section devoted to that
function, below.
If the acceleration calculated from the motion exceeds the maximum programmed acceleration (set by
2
Ixx17 in counts/msec ) for any motor involved in the move, the acceleration for all motors in the move is
decreased in proportion so that no motor exceeds this limit. The path through space is not changed, nor is
the shape of the velocity profile for any motor. If you want to specify acceleration rate directly, set TA
and TS to very small so as always to violate the limit, in which case the acceleration is controlled by the
motors’ I-variables Ixx17.
Note:
Unless the coordinate system is in segmentation mode (Isx13 > 0), do not set both
the TA and TS (Isx87 and Isx88) times to zero, even if you are planning to rely on
the acceleration. This would cause a divide-by-zero error, yielding possible erratic
performance.
When Too Effective
When blending linear moves together, the Ixx17 limit is enforced even for the intermediate decelerations
to a stop that are removed to blend into the next move. As Turbo PMAC calculates each move in the
blended sequence, it has to assume that it could be the last move in the sequence, and it will try to make
sure that the deceleration to a stop at the programmed position obeys this limit. This can result in a
deceleration time longer than the programmed move time (specified either directly with TM, or indirectly
by distance over F feedrate), which will cause the move to execute at lower than the programmed speed.
This can be especially limiting when moves are broken into very small pieces to be blended together. The
Ixx17 limit must be set higher than the top speed divided by the smallest segment time in order not to
limit the speed. This can make it too high for effective acceleration control.
When Not Effective Enough
Without the special lookahead buffer enabled, Turbo PMAC looks two moves ahead of actual move
execution to perform its acceleration limit, and can recalculate these two moves to keep the accelerations
under the Ixx17 limit. However, there are cases in which more than two moves, some much more than
two, would have to be recalculated in order to keep the accelerations under the limit. In these cases,
Turbo PMAC will limit the accelerations as much as it can, but because the earlier moves have already
been executed, they cannot be undone, and therefore the acceleration limit will be exceeded.
If you desire robust acceleration control in cases where the acceleration limiting may require modifying
the speed of several moves, the special lookahead buffer should be used. See the Turbo PMAC
Lookahead Function section in this manual for details.
Minimum Move Time
If a feedrate-specified move segment is so short in distance that it cannot reach its target velocity, it will
spend its entire time in acceleration (yielding a triangular rather than trapezoidal profile). The minimum
time for such a move is thus the specified acceleration time (the larger of TA or 2*TS). For a single move
remember to add on the extra acceleration time to decelerate to a stop.
In a time-specified move segment, if TM is less than the acceleration time, the segment will be done in
acceleration time, not TM time.
In other words, the acceleration time is the minimum time for an individual blended move or blended
move segment. This is in part a protection against move times getting so short that Turbo PMAC cannot
calculate them in real time. If you are working with very short move segments and your move sequence
is going more slowly than you want, this acceleration-time limit may well be causing the problem.
276
Writing and Executing Motion Programs
Turbo PMAC User Manual
Maximum Move Time
23
The maximum time for one programmed move is 2 -1 (8,388,607) msec, approximately 2 hours and 20
minutes. This is the maximum value that Turbo PMAC will accept with a TM command. It is also the
maximum value Turbo PMAC will compute for a feedrate-specified move when it divides the vector
distance for the move by the feedrate. If the vector distance for the move divided by the feedrate yields a
time greater than 8,388,607 msec, Turbo PMAC will use 8,388,607 msec as the move time, and the speed
will be higher than what was programmed.
The Blending Function
If more than one move is specified in succession without any intervening dwell commands, each motor
blends smoothly from its velocity for the first move to the velocity for the second move according to the
acceleration and S-curve values in force at the time. This change in speed (which can be a zero change)
starts at the point where the first move would start to decelerate to a stop at its specified end position, not
at the first move’s endpoint itself. (However, if Isx92 is set to one, “blended” moves in that coordinate
system always come to a stop before the next move.)
The acceleration parameters TA and TS can change between each move. To have the final deceleration to
a stop use a different TA or TS from the previous blending acceleration time in a sequence, declare the
new TA or TS after the final move command in the sequence, but before the DWELL or other feature that
stops the continuous sequence.
Writing and Executing Motion Programs
277
Turbo PMAC User Manual
Linear Mode Trajectories
Small acceleration time
V
TA
TM or
∆P/F
TM or
∆P/F
TA
time
V
TA
TA
TM or
∆P/F
time
TA
V
TM or
∆P/F
TM or
∆P/F
TA
TM or
∆P/F
TA
TA
time
V
TA
TA
278
TM or
∆P/F
TA
TM or
∆P/F
time
TA
Writing and Executing Motion Programs
Turbo PMAC User Manual
Linear Mode Trajectories
Acceleration time matches move time
V
TM or
∆ P/F
TM or
∆ P/F
time
V
TA
TM or
∆ P/F
TA
time
TA
V
TM or
∆ P/F
TA
TM or
∆ P/F
TA
time
TA
V
TM or
TA
∆ P/F
TM or
∆ P/F
time
TA
TA
Writing and Executing Motion Programs
TA
279
Turbo PMAC User Manual
Linear Mode Trajectories
Large (velocity limiting) acceleration time
V
TM or
∆P/F
TA
time
TA
V
TM or
∆P/F
TM or
∆P/F
TA
TA
time
TA
V
TM or
TA
280
∆P/F
TM or
TA
∆P/F
time
TA
Writing and Executing Motion Programs
Turbo PMAC User Manual
Linear Mode Trajectories
Changing acceleration times
V
TM or
∆P/F
TA1
TM or
∆P/F
TA2
time
TA2
V
TM or
∆P/F
TA1
TM or
∆P/F
TA2
time
TA2
V
TA1
TA2 actual
TA2 specified
TA2 specified
time
V
TM1
TA1
Writing and Executing Motion Programs
TM2
TA2
time
TA2
281
Turbo PMAC User Manual
Circular Blended Moves
Note:
In order for Turbo PMAC to do circular moves, coordinate-system parameter Isx13
must be greater than zero. See below for details.
Turbo PMAC allows circular interpolation on the X, Y, and Z-axes in a coordinate system. As with linear
blended moves, TA and TS control the acceleration to and from a stop, and between moves. Circular
blended moves can be either feedrate-specified (F) or time-specified (TM), just as with linear moves. It is
possible to change back and forth between linear and circular moves without stopping.
Specifying the Interpolation Plane
The first thing that should be done in preparing for a circular move is to specify the orientation of the
plane that will contain the circle. This is done by specifying the normal vector to that plane with the
NORMAL command. The arguments of this command are the component magnitudes of the vector: I (Xaxis direction), J (Y-axis direction), and K (Z-axis direction). A typical command might be NORMAL
I0.866 J0.5 K0.0. The length of the normal vector specified here is not important; only the ratio
between the component magnitudes (which determines the direction) is.
Standard Planes
To specify the circles in the XY plane, command NORMAL K-1 (equivalent to G17 in machine-tool
code). Similarly, for circles in the ZX plane, command NORMAL J-1 (G18 equivalent); for circles in the
YZ plane, command NORMAL I-1 (G19 equivalent).
Clockwise Direction Sense
The directional sense of the normal vector is right handed. The standard clockwise sense is obtained by
using normal vectors that point in the negative direction along their axes.
Circle Modes
To put the program in circular mode, use the program command CIRCLE1 for clockwise arcs (G02
equivalent) or CIRCLE2 for counterclockwise arcs (G03 equivalent). Any other move mode command –
LINEAR, RAPID, PVT, or SPLINEn – will take the program out of circular move mode. LINEAR will
restore you to linear blended moves. Once in circular mode, a circular move is specified with a move
command specifying the move endpoint and either the vector to the arc center or the distance (radius) to
the center. The endpoint may be specified either as a position or as a distance from the starting point,
depending on whether the axes are in absolute (ABS) or incremental (INC) mode (individually
specifiable).
282
Writing and Executing Motion Programs
Turbo PMAC User Manual
CIRCULAR INTERPOLATION
+Z
+Z
CW
CW
+X
+Y
+X
+Y
NORMAL K-1
NORMAL K1
+Z
+Z
CW
+X
+Y
+X
CW
+Y
NORMAL J-1
NORMAL J1
+Z
+Z
CW
+X
CW
+X
+Y
NORMAL I-1
+Y
NORMAL I1
NORMAL VECTORS FOR CIRCULAR MOVES:
THE PLANES AND CLOCK WISE ARCS THEY DEFINE
Center Vector
If the vector method of locating the arc center is used, the vector is specified by its I, J, and Kcomponents (I specifies the component parallel to the X axis, J to the Y axis, and K to the Z-axis). This
vector can be specified as a distance from the starting point (i.e. incrementally), or from the XYZ origin
(i.e. absolutely). The choice is made by specifying ‘R’ in an ABS or INC statement (e.g. ABS(R) or
INC(R)). This affects I, J, and K-specifiers together. (ABS and INC without arguments affect all axes,
but leave the vectors unchanged). The default is for incremental vector specification.
Note:
The standard machine-tool usage is for incremental vector specification even when
move endpoint specification is absolute.
A typical circular move command with a vector specification is:
X1000 Y2000 I500 J-500
Example:
Starting from the point X0 Y0, make a quarter circle clockwise in the XY plane to X20 Y20, then a linear
move to X40 Y20, then a three-quarters circle clockwise to X20 Y0. With the default modes of absolute
move and incremental vector specification, the program would be:
NORMAL K-1
F10
CIRCLE1
; XY plane
; Clockwise circle
Writing and Executing Motion Programs
283
Turbo PMAC User Manual
; Arc move; I=20-0=20; J=0-0=0
X20 Y20 I20 J0
LINEAR
X40 Y20
CIRCLE1
X20 Y0 I0 J-20
; Arc move; I=40-40=0; J=0-20=-20
PMAC
END
(25,30)
Y
NORMAL K-1
DefaultsABS (X,Y)
INC (R)
CIRCLE1
F10
X25Y30I20J5
Y
Y
I
CENTER
(30,10)
START
(10,5)
J
END
(15,10)
I
Y
Y
X
X
X
CIRCLE2
TM1000
X15Y10I-10
START
CENTER (25,20)
(15,20)
X
Y
J
CENTER
(20,20)
I
Y
CIRCLE1
F25
X30Y10I-10J10
or
I-10J10
1 CIRCLE2
TM2000
X0Y10R10
END
(0,10)
2 CIRCLE2
TM2000
X0Y10R-10
2
START,END
(30,10)
1
X
X
START (10,0)
X
Radius Size Specification
If the radius method of locating the arc center is used, the radius is the number after the letter R in the
move command. This value always represents the distance from the move starting point. With radius
specification, it is also necessary to specify whether the arc to the move endpoint is the long route (>=
180 degrees) or the short route (<= 180 degrees). Turbo PMAC’s convention is to take the short arc path
if the R value is positive and the long arc path if R is negative. R values are not modal – a value must be
specified on each move command line. It is not possible to do a full circle in a single move command
with a radius specification; the circle must be broken into at least two parts.
A typical circular move command with a radius specification is:
X1000 Y2000 R750
Example:
To do the same moves as in the above example, except with radius center specification, the program
would be:
NORMAL K-1
F10
CIRCLE1
X20 Y20 R20
X40 Y20
X20 Y0 R-20
; XY plane
; Arc move < 180 deg
; Automatically linear
; Arc move > 180 deg
Note:
Do not use the R radius specification if using the axis transformation matrices for
scaling purposes with the AROT or IROT statements. The radius value will not
scale with the axes.
284
Writing and Executing Motion Programs
Turbo PMAC User Manual
No Center Specification
If there is neither a vector specification nor a radius specification on a given move command line, the
move will be linearly interpolated between start and end points, even if the program is in circular move
mode. However, cutter compensation will not work properly if this is done. LINEAR move mode must
be explicitly declared if cutter compensation is on.
Vector Feedrate Axes
Any axes used in the circular interpolation are automatically feedrate axes for circular moves, even if they
were not so specified in an FRAX command. Other axes may or may not be feedrate axes. Any nonfeedrate axes commanded to move in the same move command will be linearly interpolated so as to finish
in the same time. This permits easy helical interpolation. See the Feedrate Axes section in this manual.
Circle-Radius Errors
If the endpoint is not the same distance from the center as the starting point, the change in radius is taken
up smoothly over the course of the move. Technically, Turbo PMAC generates an exponential spiral
curve with a constant vector feedrate and a constant change in radius with respect to time (constant
dR/dθ).
When using an IJK center vector, this spiral curve can be created over any angle between the start point,
the “center” and the end point. When using the R radius-magnitude specification, a spiral can be created
only if the distance between the start point and the end point is more than twice the radius magnitude. In
this case, the start point, the center, and the end point will all be collinear (like a semi-circle). For moves
specified this way, this technique is really used just to permit the execution of semi-circles with some
tolerance for round-off errors.
Each coordinate system has an I-variable (Isx96) that determines the limit in distance difference for which
this compensation will be done. Above this limit, a run-time error will be generated and the program will
stop. This limit allows distinguishing between round-off errors and major mistakes. Regardless of this
limit, if the distance from starting point to center or from ending point to center is zero, an error will be
generated and the program will stop. If the specified vector does not lie in the plane of interpolation, the
projection of that vector into the plane is used.
Move Segmentation Mode
Turbo PMAC computes circular trajectories through a rapid and continuous cubic spline technique called
“segmentation mode.” The spline segments are of a time specified by coordinate system variable Isx13
(in units of milliseconds). Typically a value of 5 to 10 milliseconds will be used, depending on the
number of axes being controlled by the card. When Isx13 is greater than zero, all blended moves – linear
and circular – are computed through this ongoing cubic spline technique. The exact move calculations
are performed every Isx13 milliseconds, and fine interpolation between these intermediate points is done
at the servo update rate using an efficient cubic B-spline interpolation (identical to Turbo PMAC’s
SPLINE1 mode).
If Isx13 is zero, linear moves are not computed using this spline technique and circular moves are not
permitted (if a circular move is requested, it will be done as a linear move). The difference in the actual
performance of linear-mode moves between Isx13=0 mode and Isx13>0 mode is virtually imperceptible,
unless the feature sizes of the moves are in the same range as the Isx13 time.
Segmentation mode (Isx13 > 0) also is required to use cutter-radius compensation (even if circle moves
are not explicitly programmed), kinematic calculations, and the special lookahead buffer.
Writing and Executing Motion Programs
285
Turbo PMAC User Manual
Rapid-Mode Moves
Rapid-mode moves provide for minimum-time point-to-point moves, subject to pre-defined motor
constraints. These moves are essentially jog moves for each motor assigned to an axis specified in the
move. The acceleration for each motor is controlled by the Ixx20 and Ixx21 acceleration times; however,
if the rate of acceleration determined by the motor’s move speed and these variables exceeds the Ixx19
acceleration limit, the times will be extended so that the rate will not exceed Ixx19. Many users will set
Ixx20 and Ixx21 to very small values (e.g. 1 and 0; do not set both to 0) so that the Ixx19 rate controls all
of these acceleration profiles.
The speed of the move is controlled either by the maximum-speed parameter Ixx16 (default) or the jogspeed parameter Ixx22, depending on the setting of motor variable Ixx90. A short move may not reach
the programmed speed, just accelerating and decelerating through the move.
A motion program is put into this mode using the RAPID statement. It is taken out of this mode by
another move mode command (e.g. LINEAR, CIRCLEn, PVT, SPLINEn). RAPID is equivalent to the
RS-274 G-Code G00.
Move Time
On a multi-axis rapid mode move, only the motor calculated to take the longest time at its specified Ixx16
or Ixx22 speed will actually be commanded to move at that speed. The commanded speeds for other
motors are lessened so that they have the same ratio of distance to speed, yielding the same move time for
all motors (before acceleration and deceleration are added). This makes the move path in a Cartesian
system approximately linear. However, if the acceleration times are not the same for all motors, as will
happen if the Ixx19 acceleration limits are hit, the commanded move path will not be perfectly linear.
Maximum Move Time Limit
The maximum move time for a RAPID-mode move is 8,388,607 msec, approximately 2 hours and 20
minutes. If the distance and speed for any motor request a move time greater than this time, Turbo
PMAC will use this maximum time instead, and the speed will be higher than what was programmed.
Minimum Move Time Limit
The minimum move time for a RAPID-mode move is the longest acceleration time (Ixx20 or 2*Ixx21,
whichever is greater) for any motor in the coordinate system, including motors not explicitly commanded
on the program line. Users who may command very short RAPID moves and want them to conclude very
quickly should set Ixx20 and Ixx21 very small for all motors, and use the Ixx19 acceleration limit to
control the acceleration profile.
Move Path
Because the move times (before accel/decel) for all motors are made the same, the move path in a
Cartesian system will be at least approximately linear. However, the acceleration times (not rates) for all
motors must be the same for a truly linear path in a Cartesian system. To obtain this fully linear path,
Ixx20 and Ixx21 must be the same for all motors in the coordinate system, and Ixx19 for all motors must
be set high enough not to be used for the move.
If kinematic subroutines are used to program a non-Cartesian system in Cartesian coordinates, RAPIDmode moves will not in general have a linear path in the system. The kinematic calculations for RAPIDmode moves are done only at the programmed end points for RAPID-mode moves, not at many points
along the way.
286
Writing and Executing Motion Programs
Turbo PMAC User Manual
No Blending
A rapid-mode move is never blended with another move at the programmed end-point; all motors will be
commanded to at least a momentary stop before the next move is commanded to start. However, unlike
in other modes, rapid-mode moves may be broken into at arbitrary points along the trajectory to change to
a different destination. This can be implemented with the move-until-trigger and altered-destination
functions described below.
Motion Program Move-Until-Trigger
The move-until-trigger function permits a programmed rapid-mode move to be interrupted by a trigger
and terminated by a move relative to the position at the time of the trigger. It is similar to a homing
search move, except that the motor zero position is not altered, and there is a specific destination in the
absence of a trigger.
The move-until-trigger is a variant of the RAPID move mode on Turbo PMAC. Speeds and accelerations
are governed by the same variables as for regular rapid moves. The move-until-trigger function for an
axis, and therefore for any motors defined to that axis, is specified by adding a ^{data} specifier to the
move command for the axis, where {data} is the distance to be traveled relative to the trigger before
stopping.
This makes the axis command for a move-until trigger {axis}{data}^{data}, something like
X50^-5. The first value is the destination position or distance (depending on whether the axis is in
absolute or incremental mode) to be traveled in the absence of a trigger. The second value is the distance
to be traveled relative to the position at the time of the trigger. This value is always expressed as a
relative distance, regardless of whether the axis is in absolute or incremental mode. Both values are
expressed in the axis user units.
Other axes can be specified on the same line without a caret and second value. These axes will do a
simultaneous normal RAPID move. However, if an axis matrix transformation is active for the X, Y, and
Z-axes in this coordinate system (TSELECT has been used to select a matrix), if there is a move-untiltrigger on any of the X, Y, and Z-axes, the other axes in the XYZ triplet will also execute a move until
trigger. If no post-trigger move is specified for the axis, the post-trigger distance is assumed to be zero.
If no move at all is specified for the axis, a zero-distance pre-trigger move is assumed.
The trigger condition for each motor is set up just as for homing-search moves:
Ixx97 bit 1 specifies whether input flags are used to create the trigger, or the warning-following-error
limit status bit is the trigger (the magnitude of the following error exceeds Ixx12): 0=flags, 1=error
status. Triggering on following error is often known as “torque-limited triggering.”
• If input flags are to create the trigger, Ixx25 specifies the flag register.
• If input flags are to create the trigger, Encoder/Flag I-variables I7mn2 and I7mn3 for this channel
specify which edges of which signals will cause the trigger.
• Ixx97 bit 0 specifies whether the hardware-captured counter value is used as the trigger position –
suitable for incremental encoder signals, real or simulated – or the software-read position is used
instead – suitable for other types of feedback (0=hardware, 1=software). The software-read position
must be used if the warning-following-error status is used for the trigger.
•
Note that each motor has an independent triggering function and move relative to the trigger, even if the
motors are assigned to the same axis. If a common trigger signal is desired for multiple motors, the same
trigger signal must be wired into the flag inputs for all of those motors.
Turbo PMAC will blend each motor smoothly from the pre-trigger move to the post-trigger move
according to the jog/home acceleration parameters Ixx19, Ixx20, and Ixx21.
Writing and Executing Motion Programs
287
Turbo PMAC User Manual
All motors must come to a stop, either at the originally specified position, or at the post-trigger position,
before Turbo PMAC will calculate any further in the motion program. This means that there is no
blending of the post-trigger move into any subsequent moves.
The captured value of the sensor position at the trigger is stored in a dedicated register if later access is
needed. The units are in counts; for incremental encoders, they are relative to the power-up/reset
position.
Turbo PMAC sets the motor home-search-in-progress status bit (bit 10 of the first motor status word
returned on a ? command) true (1) at the beginning of a programmed move-until-trigger move. The bit is
set false (0) either when the trigger is found, or at the end of the move.
Turbo PMAC also sets the motor trigger move status bit (bit 7 of the second motor status word returned
on a ? command) true at the beginning of a programmed move-until-trigger move, and keeps it true at
least until the end of the move. If a trigger is found during the move, this bit is set false at the end of the
post-trigger move; however, if the pre-trigger move finishes without finding a trigger, the bit is left true at
the end of the move. Therefore, this bit can be used at the end of the move to tell whether the trigger was
found or not.
Altered-Destination Moves
Turbo PMAC gives you the capability for altering the destination of certain moves in the middle of the
execution of those moves by issuing an on-line command. This allows you to start a move with a
tentative destination and then change the destination during the move, with a smooth transition to the
altered destination. If no move is currently executing, this feature also gives the capability of
commanding a simple programmed move without using a program buffer.
This technique works with RAPID-mode moves only. The only motion mode whose destination can be
altered on the fly is RAPID mode, and the only motion mode that can be used to approach the new
destination is RAPID mode.
Altered-Destination Command
This feature is implemented by the on-line coordinate-system-specific command
!{axis}{constant}[{axis}{constant}…] or its variant
!{axis}Q{constant}[{axis}Q{constant}…] . The exclamation point identifies this
command as the on-line altered-destination command. The axis letters and their associated values specify
the new destination.
In the first case (e.g. !X3.0Y2.7), the constant value associated with each axis letter directly specifies
the new destination of the axis. Typically, this first case is used when the command is issued from a host
computer.
In the second case (e.g. !XQ21YQ22), the constant value associated with each axis letter after the ‘Q’
character specifies the number of the Q-variable for the coordinate system whose value represents the
new destination for the axis. For example, if Q21=3.0 and Q22=2.7, then !XQ21YQ22 is equivalent to
!X3.0Y2.7. Usually, this second case is used when the command is issued from a Turbo PMAC PLC
program.
The values specified in this command are always positions of the new destinations (relative to “program
zero”), not distances from previous commanded positions. That is, this command is always effectively in
absolute mode, regardless of whether the axes are in absolute or incremental mode. If the axes are in
incremental mode, they will stay in incremental mode for subsequent buffered program commands.
If there is no commanded move in progress when this command is issued, Turbo PMAC will execute a
RAPID-mode point-to-point move to the specified coordinates. This makes it equivalent to a jog-toposition command in action, although the destination is specified in user units, not counts.
288
Writing and Executing Motion Programs
Turbo PMAC User Manual
If a RAPID-mode move is in progress when the command is issued, Turbo PMAC will extend the current
trajectory of each motor for Ixx92 milliseconds. At that point, it will break into the trajectory of each
motor, compute a smooth blending for each motor to the RAPID-mode trajectory toward the new
destination, and execute the modified trajectory. Because the altered-destination move is itself a RAPIDmode move, its destination can be modified with a subsequent altered-destination command.
If a move of some other mode is in progress when this command is issued, Turbo PMAC will reject the
command with an error.
Use of Altered Destination
The altered-destination command is most often used to modify the destination of a RAPID-mode move
executing from the coordinate system’s rotary motion-program buffer as the last move in that buffer. In
typical use, the RAPID move will be started with an approximate idea of the final destination, while some
sensor, such as a vision system, determines the exact location. The altered-destination command is then
sent to the coordinate system with the exact coordinates of the final destination.
If the altered-destination command is not received before the end of the move, there will be a momentary
pause before the move to the final end position is started, but all axes end up in the same location as if the
command were received before the end of the move. Note, however, that in this case, certain status bits
such as desired-velocity-zero, and in-position may get set at the end of the initial move, and so cannot be
counted on by themselves to show that the modified end-point has been reached.
The altered-destination command can also be used to modify a RAPID-mode move that is not at the end
of the rotary buffer, or one that is in a “fixed” motion-program buffer. In this case, there are a couple of
things to watch. First, if axes are in incremental mode, the subsequent moves in the program are modified
by the altered destination. Second, if the altered-destination command is received after the RAPID-mode
move is finished; it may be rejected with an error, depending on what the program is executing
subsequently.
Spline-Mode Moves
Turbo PMAC can perform two types of cubic splines (cubic in terms of the position-vs.-time equations)
to blend together a series of points on an axis. Its SPLINE1 mode is a uniform non-rational cubic Bspline and its SPLINE2 mode is a non-uniform non-rational cubic B-spline. It can, of course, do either
spline for all of the axes simultaneously. Splining is particularly suited to odd (non-Cartesian)
geometries, such as radial tables and rotary-axis robots, where there are odd axis profile shapes even for
regular tip movements.
How They Work
In SPLINE1 mode, a long move is split into equal-time segments, each of TM or TA time (depending on
the setting of global variable I42). Each axis is given a destination position in the motion program for
each segment with a normal move command line like X1000Y2000. Looking at the move command
before this and the move command after this, Turbo PMAC creates a cubic position-vs-time curve for
each axis so that there is no sudden change of either velocity or acceleration at the segment boundaries.
The commanded position at the segment boundary may be “relaxed” slightly to meet the velocity and
acceleration constraints (see figures below).
The spline move time as used in the actual spline calculations is a 24-bit fixed-point value with 12 bits of
integer and twelve bits of fraction. This provides a range of up to 4096 milliseconds (just over 4 seconds)
with a resolution of about ¼-microsecond. If I42 is set to the default value of 0, this time is specified in a
TM command, which supports the fractional resolution. If I42 is set to 1, this time is specified in a TA
command, which does not support the fractional resolution. This mode is mainly for compatibility with
older non-Turbo PMAC applications.
Writing and Executing Motion Programs
289
Turbo PMAC User Manual
Turbo PMAC computes intermediate “way-points” WPi for each axis for each point along the spline by
taking a weighted average of the specified point Pi and the specified points on either side. For the
uniform spline, this is done according to the equation:
P + 4 Pi + Pi +1
WPi = i −1
6
Turbo PMAC also computes the velocity Vi for each axis at each way-point along the spline. In the
uniform spline, it does this by taking the velocity halfway between the average velocities of the segments
on either side of the way point:
Vi =
(Pi+1 − Pi ) + (Pi − Pi−1 )
2T
− Pi −1
P
= i +1
2T
Similar calculations are done for the non-uniform spline.
Having computed exact positions and velocities at segment boundaries, Turbo PMAC calculates the
unique cubic position equation (parabolic velocity profile) that meets these constraints, and uses this
equation for interpolation.
Added Pieces
At the beginning and end of a series of splined moves, Turbo PMAC automatically adds a zero-distance
segment of the same segment time for each axis, and performs the spline between this segment and the
adjacent one. This results in S-curve acceleration to and from a stop.
Quantifying the Position Adjustment
The difference between the splined commanded position and the pre-splined (program-line) commanded
position for an axis at the end of segment i in the uniform spline can be calculated according to the simple
equation:
Diff i =
Dist i +1 − Dist i
6
where Disti is the programmed distance for segment i of the spline (whether in absolute or incremental
mode), and Disti+1 is the programmed distance for segment i+1.
5-Point Spline Correction
In contouring applications, it is often desired to pass through the series of points as closely as possible. In
these applications, the error introduced by the standard spline algorithm may be too large to tolerate.
However, in the uniform spline, a simple pre-compensation can reduce the splining errors dramatically.
For each point Pi in the spline, replace with a point P’i with the following formula before sending to
Turbo PMAC:
P' i =
− Pi −1 + 8 Pi − Pi +1
6
Non-Uniform Spline
Turbo PMAC’s SPLINE2 mode is similar to the SPLINE1 mode, except that the requirement that the
TA spline segment time remain constant is removed. The removal of this constraint makes the SPLINE2
mode a non-uniform non-rational cubic B-spline, whereas the SPLINE1 mode is a uniform non-rational
cubic B-spline. The “non-rational” specification indicates that there are no independent weightings
(ratios) of the different points in the spline.
The added segment at the beginning of a spline has the same time as the first programmed segment; the
added segment at the end of a spline has the same time as the last programmed segment.
The combined time of any three consecutive segments in a SPLINE2 continuous spline must be less than
8,388,608 msec, or about 2 hours and 20 minutes.
290
Writing and Executing Motion Programs
Turbo PMAC User Manual
Cubic Spline Trajectories
V
V
TA (added)
TA
TA (added) time
TA (added)
TA
TA (added)time
TA
Two Programmed Segments
One Programmed Segment
V
TA (added)
TA
TA
TA
TA (added)
time
Three Programmed Segments
V
TA (added)
TA
TA
TA
TA (added) time
TA
Four Programmed Segments
SPLINED MOVES
If segment were
done at constant
velocity:
All segments same time
etc.
VEL
Vc =
No velocity or acceleration
discontinuities at segment
boundaries
P
TA
etc.
INCREMENTAL
SPLINE
TA500
X10000
X9000
X10500
X12000
TA
TA
TA
TA
TA
(added)
TA
(added)
TIME
PMAC Transition Point Moves
Position, Velocity, and Time (PVT mode) [Parabolic Velocity]
VEL
A0
(calculated)
P0 ,V 0
(from
before)
dA
= constant (calculated)
dt
P1 (specified)
V1 (specified)
A 1 (calculated)
PVT200
...
X9000:150
...
P P1 P0
TA
(specified)
Writing and Executing Motion Programs
TIME
291
Turbo PMAC User Manual
PVT-Mode Moves
For the user who desires more direct control over the trajectory profile, Turbo PMAC offers PositionVelocity-Time (PVT) mode moves. In these moves, you specify the axis states directly at the transitions
between moves (unlike in blended moves). This requires more calculation by the host, but allows tighter
control of the profile shape. For each piece of a move, you specify the end position or distance, the end
velocity, and the piece time.
Mode Statement
Turbo PMAC is put in this mode with the program statement PVT{data}, where {data} is a floatingpoint constant, variable, or expression, representing the piece time in milliseconds. If I42 is set to the
default value of 0, Turbo PMAC converts this value to a 24-bit value with 12 bits of integer and 12 bits of
fraction. This provides a range of up to 4096 milliseconds (just over 4 seconds) with a resolution of about
¼-microsecond. If I42 is set to 1, Turbo PMAC converts this to a 12-bit integer value with no fractional
component (rounding to the nearest integer). This mode is mainly for compatibility with older non-Turbo
PMAC applications.
The move time may be changed between moves, either with another PVT{data} statement, or with a
TM{data} statement if I42 = 0, or a TA{data} statement if I42 = 1. The program is taken out of this
mode with another move mode statement (e.g. LINEAR, RAPID, CIRCLE, SPLINE).
Move Statements
A PVT mode move is specified for each axis to be moved with a statement of the form
{axis}{data}:{data}, where {axis} is a letter specifying the axis, the first {data} is a value
specifying the end position or the piece distance (depending on whether the axis is in absolute or
incremental mode), and the second {data} is a value representing the ending velocity.
The units for position or distance are the user length or angle units for the axis, as set in the Axis
Definition statement. The units for velocity are defined as length units divided by time units, where the
length units are the same as those for position or distance, and the time units are defined by variable Isx90
for the coordinate system (feedrate time units). The velocity specified for an axis is a signed quantity.
Turbo PMAC Calculations
From the specified parameters for the move piece, and the beginning position and velocity (from the end
of the previous piece), Turbo PMAC computes the only third-order position trajectory path to meet the
constraints. This results in linearly changing acceleration, a parabolic velocity profile, and a cubic
position profile for the piece.
Problems in Stepping
Since you can specify (directly or indirectly) a non-zero end velocity for the move, it is not a good idea to
step through a program of transition-point moves, and great care must be exercised in downloading these
moves in real time. With the use of the BLOCKSTART and BLOCKSTOP statements surrounding a series
of PVT moves, the last of which has a zero end velocity, it is possible to use a Step command to execute
only part of a program.
Use of PVT to Create Arbitrary Profiles
The PVT mode is the most useful for creating arbitrary trajectory profiles. It provides a "building block"
approach to putting together parabolic velocity segments to create whatever overall profile is desired.
The diagram PVT Segment Shapes, below, shows common velocity segment profiles. PVT mode can
create any profile that any other move mode can.
292
Writing and Executing Motion Programs
Turbo PMAC User Manual
PVT
Segme
nt
Shapes
Vel
V
Vel
V
∆P=1/2 Vt
∆P=1/3 Vt
t
Time
Vel
t
Time
t
Time
2t
Time
Vel
V
V
∆P=2/3 Vt
∆P=Vt
t
Time
Vel
Vel
V2
V
V
V/2
∆P =1/6 Vt
1
∆ P =5/6 Vt
∆P=1/2(V +V ) t
1
2
2
t
Time
t
Use of PVT in Contouring
PVT mode provides excellent contouring capability, because it takes the interpolated commanded path
exactly through the programmed points. It creates a path known as a Hermite Spline. LINEAR and
SPLINE modes are 2nd and 3rd-order B-splines, respectively, which pass to the inside of programmed
points.
Compared to Turbo PMAC’s SPLINE mode, PVT produces a more accurate profile. Its worst-case error
can be estimated as:
E=
V 4T 4
Rθ 4
=
3
384
384 R
where V is the vector velocity, T is the segment time, R is the local radius of curvature, and θ is the
subtended angle.
Writing and Executing Motion Programs
293
Turbo PMAC User Manual
Cutter Radius Compensation
Turbo PMAC provides the capability for performing cutter (tool) radius compensation on the moves it
performs. This compensation can be performed among the X, Y, and Z axes, which should be physically
perpendicular to each other. The compensation automatically offsets the described path of motion
perpendicular to the path by a programmed amount, compensating for the size of the tool. This allows
you to program the path along the edge of the tool, letting Turbo PMAC calculate the tool-center path
based on a radius magnitude that can be specified independently of the program.
Turbo PMAC supports both two-dimensional (2D) and three-dimensional (3D) cutter radius
compensation. In the older and more common 2D compensation, described immediately below, first
specify the plane of compensation, then the direction of compensation relative to the path and the radius
magnitude. In the newer 3D compensation, specify the surface-normal vector and the tool-orientation
vector, as well as major and minor radii for the tool.
Cutter radius compensation is valid only in LINEAR and CIRCLE move modes. The moves must be
specified by F (feedrate), not TM (move time). Turbo PMAC must be in move segmentation mode (Isx13
> 0) to do this compensation (Isx13 > 0 is required for CIRCLE mode anyway.)
Note:
In CIRCLE mode, a move specification without any center specification results in
a linear move. This move is executed correctly without cutter radius compensation
active, but if the compensation is active, it will not be applied properly in this case.
A linear move must be executed in LINEAR mode for proper cutter-radius
compensation.
Defining the Plane of Compensation
Several parameters must be specified for the compensation. First, the plane in which the compensation is
to be performed must be set using the buffered motion-program NORMAL command. Any plane in XYZspace may be specified. This is done by specifying a vector normal to that plane, with I, J, and Kcomponents parallel to the X, Y, and Z-axes, respectively.
For example, NORMAL K-1, by describing a vector parallel to the Z-axis in the negative direction,
specifies the XY-plane with the normal right/left sense of the compensation (NORMAL K1 would also use
the XY-plane, but invert the right/left sense).
294
Writing and Executing Motion Programs
Turbo PMAC User Manual
This same command also specifies the plane for circular interpolation. NORMAL K-1 is the default. The
compensation plane should not be changed while compensation is active.
Other common settings are NORMAL J-1, which specifies the ZX-plane for compensation, and NORMAL
I-1, which specifies the YZ-plane. These three settings of the normal vector correspond to RS-274 Gcodes G17, G18, and G19, respectively. If implementing G-codes in Turbo PMAC program 1000,
incorporate them in PROG 1000:
N17000 NORMAL K-1 RETURN
N18000 NORMAL J-1 RETURN
N19000 NORMAL I-1 RETURN
Defining the Magnitude of Compensation
The magnitude of the compensation – the cutter radius – must be set using the buffered motion program
command CCR{data} (Cutter Compensation Radius). This command can take either a constant
argument (e.g. CCR0.125) or an expression in parentheses (e.g. CCR(P10+0.0625)). The units of
the argument are the user units of the X, Y, and Z-axes. In RS-274 style programs, these commands are
often incorporated into “tool data” D-codes using Turbo PMAC motion program 1003.
Negative and zero values for cutter radius are possible. Note that the behavior in changing between a
positive and negative magnitude is different from changing the direction of compensation. See the
Changes in Compensation section of this manual. Also, the behavior in changing between a non-zero
magnitude and a zero magnitude is different from turning the compensation on and off. See the
appropriate sections below.
Turning On Compensation
The compensation is turned on by buffered motion program command CC1 (offset left) or CC2 (offset
right). These are equivalent to the RS-274 G-Codes G41 and G42, respectively. If implementing GCode subroutines in Turbo PMAC motion program 1000, incorporate them in PROG 1000:
N41000 CC1 RETURN
N42000 CC2 RETURN
Turning Off Compensation
The compensation is turned off by buffered motion program command CC0, which is equivalent to the
RS-274 G- Code G40. If implementing G-Code subroutines in Turbo PMAC motion program 1000,
incorporate them in PROG 1000:
N40000 CC0 RETURN
How Turbo PMAC Introduces Compensation
Turbo PMAC gradually introduces compensation over the next LINEAR or CIRCLE-mode move
following the CC1 or CC2 command that turns on compensation. This lead-in move ends at a point one
cutter radius away from the intersection of the lead-in move and the first fully compensated move with
the line from the programmed point to this compensated endpoint being perpendicular to the path of the
first fully compensated move at the intersection.
Note:
Few controllers can make their lead-in move a CIRCLE-mode move. This
capability permits establishing contact with the cutting surface very gently,
important for fine finishing cuts.
Inside Corner Introduction
If the lead-in move and the first fully compensated move form an inside corner, the lead-in move goes
directly to this point. When the lead-in move is a LINEAR-mode move, the compensated tool path will
be at a diagonal to the programmed move path. When the lead-in move is a CIRCLE-mode move, the
compensated tool path will be a spiral.
Writing and Executing Motion Programs
295
Turbo PMAC User Manual
Introducing Compensation – Inside Corner
Line
Programmed
Path
r
Line
Programmed
Path
r
Tool Center
Path
Line
Arc
Line
Arc
Line
Tool Center
Path
Line
CC2
CC2
Line to Line
Line to Arc
Line
Programmed
Path
r
Arc
Tool Center
Path
Line
Programmed
Path
r
Arc
Arc
Arc
Spiral
Spiral
Tool Center
Path
CC2
CC2
Arc to Line
Arc to Arc
Outside Corner Introduction
If the lead-in move and the first fully compensated move form an outside corner, the lead-in move first
moves to a point one cutter radius away from the intersection of the lead-in move and the first fully
compensated move, with the line from the programmed point to this compensated endpoint being
perpendicular to the path of the lead-in move at the intersection. When the lead-in move is a LINEARmode move, this compensated tool path will be at a diagonal to the programmed move path.
When the lead-in move is a CIRCLE-mode move, this compensated tool path will be a spiral. Then a
circular arc move with radius equal to the cutter radius is added, ending at a point one cutter radius away
from the intersection of the lead-in move and the first fully compensated move, with the line from the
programmed point to this compensated endpoint being perpendicular to the path of the first fully
compensated move at the intersection.
Introducing Compensation – Outside Corner
CC2
CC2
Line
Line
Line
Programmed
Path
Line
r
r
Arc
Arc
Line
Tool Center
Path
r
Line
CC2
Arc
r
Tool Center
Path
Line to Line
Line to Arc
CC2
Arc Programmed
Path
Arc
Line
Programmed
Path
Spiral
r
Arc
Spiral
Arc
Arc
r
r
Tool Center
Path
r
Tool Center
Path
Line
Arc to Line
296
Programmed
Path
Arc
Arc to Arc
Writing and Executing Motion Programs
Turbo PMAC User Manual
Note that the behavior for lead-in moves is different from changing the compensation radius from zero to
a non-zero value while compensation is active. An arc move is always added at the corner, regardless of
the setting of Isx99. This ensures that the lead-in move never cuts into the first fully compensated move.
Treatment of Compensated Inside Corners
Inside corners are still subject to the blending due to the TA and TS times in force (default values set by
coordinate system I-variables Isx87 and Isx88, respectively). The longer the acceleration time the larger
the rounding of the corner. (The corner rounding starts and ends a distance F*TA/2 from the
compensated, but unblended corner.) The greater the portion of the blending is S-curve, the squarer the
corner will be.
When coming to a full stop (e.g. Step, Quit, or DWELL at the corner) at an inside corner, Turbo PMAC
will stop at the compensated, but unblended, corner point.
Inside Corner Cutter Compensation
Line
Programmed
Path
r
r
Line
Tool Center
Path
Line
Programmed
Path
r
Line
Arc
r
Arc
Line
Line
Tool Center
Path
Line to Line
Line to Arc
Line
Programmed
Path
Programmed
Path
r
r
Arc
Tool Center
Path
Line
Arc
r
Arc
Arc
r
Arc
Arc
Tool Center
Path
Arc to Line
Arc to Arc
Treatment of Outside Corners
For outside corners, Turbo PMAC will either blend the incoming and outgoing moves directly together,
or it will add an arc move to cover the additional distance around the corner. Which option it chooses is
dependent on the relative angle of the two moves and the value of I-variable Isx99.
The relative angle between the two moves is expressed as the change in directed angle of the motion
vector in the plane of compensation. If the two moves are in exactly the same direction, the change in
o
o
directed angle is 0 ; if there is a right angle corner, the change is +/-90 ; if there is a complete reversal, the
o
change in directed angle is 180 .
Isx99 specifies the boundary angle between directly blended outside corners and added-arc outside
o
corners. It is expressed as the cosine of the change in the directed angle of motion (cos0 =1.0,
o
o
cos90 =0.0, cos180 =-1.0) at the boundary of the programmed moves. The change in directed angle is
o
equal to 180 minus the “included angle” at the corner.
Writing and Executing Motion Programs
297
Turbo PMAC User Manual
Sharp Outside Corner
If the cosine of the change in directed angle is less than Isx99, which means the corner is sharper than the
specified angle, then an arc move will be added around the outside of the corner.
Outside Corner Cutter Compensation, Sharp Angle (cos ∆Θ < Isx99)
Line
Arc
Line
Arc
Line
r
Line
r
Programmed
Path
r
Arc
Tool Center
Path
Programmed
Path
r
Arc
Tool Center
Path
Line
Line
Line to Line
Line
Line to Arc
Arc
Line
Arc
r
r
Programmed
Path
Arc
r
Arc
Programmed
Path
Arc
r
Arc
Arc
Arc
Tool Center
Path
Tool Center
Path
Arc to Line
Arc to Arc
Shallow Outside Corner
However, if the cosine of the change in directed angle is greater than Isx99, which means that the corner
is flatter than the specified angle, the moves will be directly blended together without an added arc.
Outside Corner Cutter Compensation, Shallow Angle ( cos ∆Θ > Isx99)
Line
Programmed
Path
Line
r
Line
Line
Programmed
Path
r
Line
r
r
Arc
Tool Center
Path
Line
Tool Center
Path
Line to Line
Line to Arc
Arc
Arc
Arc
Arc
Line
Programmed
Path
r
r
Tool Center
Path
Line
Arc to Line
Arc
Programmed
Path
r
Arc
r
Arc
Tool Center
Path
Arc to Arc
The added arc prevents the compensated corner from extending too far out on the outside of a sharp
corner. However, as an added move, it has the minimum time of the acceleration time, which can cause a
slowdown on a very shallow angle. While the default value for Isx99 of 0.9998 (cos1o) causes an arc to
be added on any change in angle greater than 1o, many users will set Isx99 to 0.707 (cos45o) or 0.0
(cos90o) so arcs are only added on sharp corners.
298
Writing and Executing Motion Programs
Turbo PMAC User Manual
When coming to a full stop (e.g. Step, Quit, /, or DWELL) at an outside corner with an added arc, Turbo
PMAC will include the added arc move before stopping. When coming to a full stop at an outside corner
without an added arc, Turbo PMAC will stop at the compensated, but unblended, corner point.
Treatment of Full Reversal
If the change in directed angle at the boundary between two successive compensated moves is 180o + 1o
(the included angle is less than 1o), this is considered a “full reversal” and special rules apply. If both the
incoming and outgoing moves are lines, the corner is always considered an outside corner, and an arc
move of approximately 180o is added. If one or both of the moves is an arc, Turbo PMAC will check for
possible inside intersection of the compensated moves. If such an intersection is found, the corner will be
treated as an inside corner. Otherwise, it will be treated as an outside corner with an added 180o arc
move.
Reversal In Cutter Compensation
Line
Programmed
Path
Line
<1o
Line
Line
Tool Cen
ter Path
r
r
Arc
Arc
Arc
Line
r
Arc
Tool Center
Arc
Path Right
Line to Line
Tool Center
Path
Programmed Path
Line
r
Tool Center
Path Left
Line to Arc,
Inside and Outside
Arc
r
Programmed
Path
r
Arc
Arc
Arc
Arc
r
r
Arc
Arc
Programmed
Path
Arc to Arc Outside
Arc
Tool Center
Path
Arc
Arc to Arc Inside
Note on Full Circles
If a full-circle move is executed while in cutter compensation, and one or both of the ends produces a
shallow outside corner that is directly blended (no added arc – see Treatment of Outside Corners, above),
the compensated arc move will be extended beyond 360o, and Turbo PMAC may produce just a very
short arc, 360o shorter than what is desired (making it appear that the circle has been “skipped.”
Although typically this is the result of sloppy programming – an outside corner with a full circle causes
an overcut into the circle – many machine designers may want to permit slight cases of this. Coordinate
system parameter Isx97 defines the shortest arc angle that may be executed; the longest arc angle is 360o
plus this angle.
The default value of Isx97 sets a minimum arc angle of one-millionth of a semi-circle, enough to account
for numerical round-off, but sometimes not enough for compensated full circles. To handle these cases,
Isx97 should be set to a somewhat larger value.
Writing and Executing Motion Programs
299
Turbo PMAC User Manual
Failure When Compensation Extends Full Circle
Tool Center
Path
1
1
r
r 2
Programmed
Path
r
Short
Arc
Executed
Compensated
Circle
“Skipped”
3
2
Programmed
Full Circle
Speed of Compensated Moves
Tool center speed for the compensated path remains the same as that programmed by the F parameter.
On an arc move, this means that the tool edge speed (the part of the tool in contact with the part) will be
different from that programmed by the fraction Rtool/Rarc.
Changes in Compensation
Turbo PMAC permits changes both to the radius of compensation and the direction of compensation
while the compensation is active. It is important to understand exactly how Turbo PMAC changes the
compensated path in these cases.
Radius Magnitude Changes
Changes in the magnitude of compensation (new CCR values) made while compensation is active are
introduced linearly over the next move. When this change is introduced over the course of a LINEAR-mode
move, the compensated tool path will be at a diagonal to the programmed move path. When this change is
introduced over the course of a CIRCLE-mode move, the compensated tool path will be a spiral.
Compensation Direction Changes
Changes in the direction of compensation (between CC1 and CC2) made during compensation are
generally introduced at the boundary between the two moves.
Cutter Compensation Change of Direction
CC1
Arc
Line
Line
Arc
Line
CC2
Line
Programmed
Path
CC2
Line
Programmed
Path
Tool Center
Path
Line
Line to Line
Line to Arc
Arc
CC1
CC2
CC2
Programmed
Path
Line
Tool Center
Path
Line
Arc to Line
300
Programmed
Path
Arc
Arc
CC1
Tool Center
Path
Arc
Arc
Arc
Tool Center
Path
CC1
Arc to Arc
Writing and Executing Motion Programs
Turbo PMAC User Manual
However, if there is no intersection between the two compensated move paths, the change is introduced
linearly over the next move.
Cutter Compensation Change of Direction – No Intersection
Programmed
CC1
Path
Line
Line
CC2
Line
Line
Line
Programmed
Path
Line
Line
CC1
Line
Spiral
CC2
Tool Center
Path
Arc
Line
Tool Center
Path
Line
Change Through a Line
Change Through an Arc
Tool Center
Path
Arc
Arc
CC1
Line
Line
CC2
Arc
CC2
Programmed Path
CC1
Line
r
Programmed
Path
Arc
Tool Center Path
Line
Change Through a Line
Change in Reversal
How Turbo PMAC Removes Compensation
Turbo PMAC gradually removes compensation over the next LINEAR or CIRCLE-mode move following
the CC0 command that turns off compensation. This lead-out move starts at a point one cutter radius
away from the intersection of the lead-in move and the first fully compensated move with the line from
the programmed point to this compensated endpoint being perpendicular to the path of the first fully
compensated move at the intersection. Note that few controllers can make their lead-out move a
CIRCLE-mode move. This capability permits releasing contact with the cutting surface very gently,
important for fine finishing cuts.
Inside Corner
If the last fully compensated move and the lead-out move form an inside corner, the lead-out move starts
directly from this point to the programmed endpoint. When the lead-out move is a LINEAR-mode move,
the compensated tool path will be at a diagonal to the programmed move path. When the lead-in move is
a CIRCLE-mode move, the compensated tool path will be a spiral.
Removing Compensation – Inside Corner
Line
Programmed
Path
Tool Center
Path
Line
CC0
CC0
Programmed
Path
r
r
Arc
Line
Line
Arc
Line
Tool Center
Path
Line to Line
Line
Programmed
Path
Tool Center
Path
Line
Line to Arc
CC0
CC0
r
Arc
Programmed
Path
r
Arc
Arc
Spiral
Arc to Line
Writing and Executing Motion Programs
Line
Arc
Tool Center
Path
Spiral
Arc to Arc
301
Turbo PMAC User Manual
Outside Corner
If the last fully compensated move and the lead-out move form an outside corner, the last fully
compensated move ends at a point one cutter radius away from the intersection of the last fully
compensated move and the lead-out move, with the line from the programmed point to this compensated
point being perpendicular to the path of the fully compensated move at the intersection.
Turbo PMAC then adds a circular arc move with radius equal to the cutter radius, ending at a point one
cutter radius away from the same, with the line from the programmed point to this compensated endpoint
being perpendicular to the path of the lead-out move at the intersection.
Finally, Turbo PMAC gradually removes compensation over the lead-out move itself, ending at the
programmed endpoint of the lead-out move. When the lead-out move is a LINEAR-mode move, this
compensated tool path will be at a diagonal to the programmed move path. When the lead-in move is a
CIRCLE-mode move, this compensated tool path will be a spiral.
Removing Compensation – Outside Corner
Line
Line
Programmed
Path
Tool Center
Path
CC0
Line
r
Arc
Line
r
Tool Center
Path
r
Line
Programmed
Path
CC0
Arc
r
Arc
Line
Line to Line
Line to Arc
Arc
Line
Programmed
Path
CC0
Spiral
r
r
Tool Center
Path
Line
Arc
Tool Center
Path
Spiral
r
r
Arc
Arc to Line
Arc
Programmed
Path
CC0
Arc
Arc
Arc to Arc
Note that this behavior is different from changing the magnitude of the compensation radius to zero while
leaving compensation active. An arc move is always added at the corner, regardless of the setting of
Isx99. This ensures that the lead-out move will never cut into the last fully compensated move.
Failures in Cutter Compensation
It is possible to give Turbo PMAC a program sequence in which the cutter compensation algorithm will
fail, not producing desired results. There are three types of reasons the compensation can fail:
1. Inability to calculate through corner
2. Inside corner smaller than radius
3. Inside arc radius smaller than cutter radius
Inability to Calculate through Corner
First, if Turbo PMAC cannot see ahead far enough in the program to find the next move with a
component in the plane of compensation before the present move is calculated, then it will not be able to
compute the intersection point between the two moves. This can happen for several reasons:
• There is a move with no component in the plane of compensation (i.e. perpendicular to the plane of
compensation, as in a Z-axis-only move during XY compensation) before the next move in the plane
of compensation, and no CCBUFFER compensation block buffer declared (see below).
• There are more moves with no component in the plane of compensation before the next move in the
plane of compensation than the CCBUFFER compensation block buffer can hold (see below).
• There are more than 10 DWELLs before the next move in the plane of compensation.
302
Writing and Executing Motion Programs
Turbo PMAC User Manual
•
Program logic causes a break in blending moves (e.g. looping twice through a WHILE loop).
Failures in Cutter Compensation
Overcut
Programmed
Path
Line
r
r
r
Tool
Center Point
at Failure
Line
Tool
Center Point
at Failure
(No Overcut)
Line
Tool Center
Path
Arc
r
r
Line
Line
Tool Center
Path
Programmed
Path
Line
Line
Line
Failure to See Through
Inside Corner
Tool Center
Path
Line
Arc
Line
r
Line
Line
r
Programmed
Path
Overcut
Inside Corner Smaller Than
Cutter Radius
Failure to See Through
Outside Corner
Line
Programmed
Path
Line
Arc
r
r
Tool Center
Path
Line
Stopping
Point
(Not
Executed)
Arc Radius Smaller Than
Cutter Radius
If Turbo PMAC cannot find the next move in time, it will end the current move as if the intersection with
the next move would form an outside corner. If the next move, when found, does create an outside
corner, or continues straight on, compensation will be correct. On an outside corner, an arc move is
always added at the corner, regardless of the setting of Isx99. However, if the next move creates an inside
corner, the path will have overcut into the corner. In this case, Turbo PMAC then moves to the correct
intersection position and continues with the next move, leaving the overcutting localized to the corner.
Inside Corner Smaller Than Radius
Second, if the compensated path produces an inside corner with one of the moves shorter than the cutter
radius, the cutter compensation will not work properly. This situation results in a compensated move that
is in the opposite direction from that of the uncompensated move, and there will be overcutting at the
corner.
Inside Arc Radius Smaller Than Cutter Radius
Third, if the program requests an arc move with compensation to the inside, and the programmed arc
radius is smaller than the cutter radius, then no proper path can be calculated. In this case, Turbo PMAC
ends the program at the end of the previous move with a run-time error, setting the internal run-time error
code in register Y:$002x14 to 7.
Block Buffering for Cutter Compensation
If the application requires the execution of moves perpendicular to the plane of compensation while cutter
compensation is active, it will require that a special buffer be defined to hold these moves while Turbo
PMAC scans ahead to find the next move in the plane of compensation so it can compute the proper
intersection between the incoming move to this point in the plane and the outgoing move.
This buffer is created with the on-line coordinate-system-specific command DEFINE
CCBUF{constant}, where {constant} is a positive integer representing the number of moves
perpendicular to the compensation plane that can be stored in the buffer. This number should be at least
as large as the largest number of consecutive perpendicular moves between any two moves in the plane.
Writing and Executing Motion Programs
303
Turbo PMAC User Manual
With this buffer defined for the coordinate system, if Turbo PMAC encounters one or more moves
perpendicular to the plane of compensation while compensation is active, these moves will be stored in
the CCBUF temporarily while the next move in the plane is found, so the intersections can be computed
correctly. However, if there is not enough room in the buffer to store all of the perpendicular moves
found, Turbo PMAC will assume an outside-corner intersection; if the next move in the plane actually
forms an inside corner, overcut will have occurred.
When programmed moves are actually stored in the CCBUF, commands that change the current position
value – HOME, HOMEZ, and PSET – are not permitted. Turbo PMAC will report an ERR019 if I6 is set to
1 or 3.
The CCBUF, which stores motion program blocks for the purpose of computing proper cutter
compensation intersection points, should not be confused with the LOOKAHEAD buffer, which stores
small motion “segments” generated from these programmed blocks for the purpose of guaranteeing
observance of position, velocity, and acceleration limits. Both of these buffers may be defined and active
for a coordinate system at the same time.
The CCBUF is a temporary buffer. Its contents are never retained through a power-down or card reset;
the buffer itself is only retained through a power-down or reset if it was defined, and I14 was set to 1, at
the time of the last SAVE command.
Single-Stepping While in Compensation
It is possible to execute moves in “single-step” mode while cutter compensation is active, but the user
should be aware of several special considerations for this mode of operation. Because of the need for the
program to see ahead far enough to find the next move in the plane of compensation before the current
move can be executed, the execution of an S single-step command may not produce the intuitively
expected results. The single-step command on a move in compensation causes the preliminary
calculations for that move to be done, not for the move actually to be executed. This has the following
ramifications:
•
•
•
•
A single-step command on the lead-in move for compensation will produce no motion, because the
next move has not yet been found.
Single-step commands on compensated moves in the plane of compensation will cause the previous
move to execute.
Single-step commands on compensated moves perpendicular to the plane of compensation will
produce no motion, as these will just be held in the CCBUFFER. A single-step command on the next
move in the plane of compensation will cause the previous move in the plane, plus all buffered moves
perpendicular to the plane to execute.
A single-step command on the lead-out move will cause both the last fully compensated move and the
lead-out move to execute.
Unlike many controllers, Turbo PMAC can execute non-motion program blocks with single-step
commands with cutter compensation active. However, you should be aware that the execution of these
blocks may appear out of sequence, because the motion from the previous programmed move block will
not yet have been executed.
Synchronous M-variable assignments in this mode are still buffered and not executed until the actual start
of motion execution of the next programmed move.
304
Writing and Executing Motion Programs
Turbo PMAC User Manual
Three-Dimensional Cutter Radius Compensation
Turbo PMAC provides the capability for performing three-dimensional (3D) cutter (tool) radius
compensation on the moves it performs. This compensation can be performed among the X, Y, and Z
axes which should be physically perpendicular to each other (even if the motors assigned to the axes are
not). Unlike the more common two-dimensional (2D) compensation, you can specify independently the
offset vector normal to the cutting surface, and the tool orientation vector.
The 3D compensation algorithm automatically uses this data to offset the described path of motion,
compensating for the size and shape of the tool. This permits you to program the path along the surface
of the part, letting Turbo PMAC calculate the path of the center of the end of the tool.
3D compensation is valid only in LINEAR and CIRCLE move modes, and is really intended only for
LINEAR moves.
A note on terminology: Much of the documentation on the older two-dimensional cutter-radius
compensation refers to just “cutter-radius compensation”, since there was no 3D compensation at the
time. Documentation specific to 3D compensation will always specify “3D” compensation.
Defining the Magnitude of 3D Compensation
The magnitude of 3D compensation is determined by two user-declared radius values. The first of these
is the radius of the rounded end of the cutter, set by the buffered motion program command CCR{data}
(Cutter Compensation Radius). This command can take either a constant argument (e.g. CCR2.35) or an
expression in parentheses (e.g. CCR(Q20-0.001)). The units of the argument are the user units of the
X, Y, and Z-axes. In operation, the compensation first offsets the path by the cutter’s end radius along the
surface-normal vector (see below).
3D Compensation: Cutting Tool Cross Sections
TR
TR
TR
C
C
C
C
CCR=0
0<CCR<TR
R
R
CCR=TR
The second value is the tool radius itself, the radius of the shaft of the tool. This is set by the buffered
motion program command TR{data} (Tool Radius). This command can take either a constant
argument (e.g. TR7.50) or an expression in parentheses (e.g. TR(7.50-Q99)). The units of the
argument are the user units of the X, Y, and Z-axes. In operation, the compensation next offsets the path
by an amount equal to the tool radius minus the cutter’s end radius, perpendicular to the “tool-orientation”
vector (see below).
A flat-end cutter will have a cutter-end radius of zero. A ball-end cutter (hemispherical tip) will have a
cutter-end radius equal to the tool (shaft) radius. Other cutters will have a cutter-end radius in between
zero and the tool radius.
Turning on 3D Compensation
3D cutter compensation is turned on by the buffered motion program command CC3. Since the offset
vector is specified explicitly, there is no left or right compensation here. When 3D compensation is
turned on, the surface-normal vector is set to the null (zero-magnitude) vector automatically, and the toolorientation vector is also set to the null vector automatically.
Writing and Executing Motion Programs
305
Turbo PMAC User Manual
Until a surface-normal vector is declared explicitly with 3D compensation active, no actual compensation
will occur. A tool-orientation vector must also be declared for compensation to work on anything other
than a ball-nose cutter.
Turning Off 3D Compensation
3D cutter compensation is turned off by the buffered motion program command CC0, just as for 2D
compensation. Compensation will be removed over the next LINEAR or CIRCLE mode move after
compensation has been turned off.
Declaring the Surface-Normal Vector
The direction of the surface-normal vector is determined by the NX{data}, NY{data}, and
NZ{data} components declared in a motion program line. The absolute magnitude of these components
does not matter, but the relative magnitudes define the direction. The direction must be from the surface
into the tool.
Generally, all three components should be declared together. If only one or two components are declared
on a program line, the remaining component(s) are left at their old value(s), which could lead to
unpredictable results. If it is desired that a component value be changed to zero, it should be declared
explicitly as zero.
Note that the coordinates of the surface-normal vector must be expressed in the machine coordinates. If
the part is on a rotating table, these coordinates will not in general be the same as the original part
coordinates from the part design – the vector must be rotated into machine coordinates before sending to
Turbo PMAC.
The surface-normal vector affects the compensation for the move on the same line of the motion program,
and all subsequent moves until another surface-normal vector is declared. In usual practice, a surfacenormal vector is declared for each move, affecting that move alone.
Declaring the Tool-Orientation Vector
If the orientation of the cutting tool can change during the compensation, as in five-axis machining, the
orientation for purposes of compensation is declared by means of a tool-orientation vector. (If the
orientation is constant, as in three-axis machining, the orientation is usually declared by the normal vector
to the plane of compensation, although the tool-orientation vector may be used.)
The direction of the tool-orientation vector is determined by the TX{data}, TY{data}, and
TZ{data} components declared in a motion program line. The absolute magnitude of these components
does not matter, but the relative magnitudes define the direction. The direction sense of the toolorientation vector is not important; it can be from base to tip, or from tip to base
Generally, all three components should be declared together. If only one or two components are declared
on a program line, the remaining components are left at their old values, which could lead to
unpredictable results. If it is desired that a component value be changed to zero, it should be declared
explicitly as zero.
Note that the coordinates of the surface-normal vector must be expressed in the machine coordinates. If
the part is on a rotating table, these coordinates in general will not be the same as the original part
coordinates from the part design.
The tool-orientation vector affects the compensation for the move on the same line of the motion
program, and all subsequent moves until another tool-orientation vector is declared. In usual practice, a
tool-orientation vector is declared for each move, affecting that move alone.
Note that the tool-orientation vector declared here does not command motion; it merely tells the
compensation algorithm the angular orientation that has been commanded of the tool. Typically the
motion for the tool angle has been commanded with A, B, and/or C-axis commands, often processed
through an inverse-kinematic subroutine on Turbo PMAC.
306
Writing and Executing Motion Programs
Turbo PMAC User Manual
How 3D Compensation is Performed
In operation, Turbo PMAC starts from the uncompensated X, Y, and Z-axis positions for each end-point
programmed while 3D compensation is active. Then two offsets are applied to the X, Y, and Z-axis
positions. The first offset is taken along the surface-normal vector, of a magnitude equal to the tip radius.
The second offset is then taken toward the center of the tool, in the plane containing both the surfacenormal vector and the tool-orientation vector, perpendicular to the tool-orientation vector, of a magnitude
equal to the cutter radius minus the tip radius.
Once the modified end-point is calculated, the move to that end-point is calculated just as it would be
without compensation. If the program is in LINEAR mode, it will be linearly interpolated. If the
program is in CIRCLE mode (not advised), arc interpolation will be applied.
Because the offset to the end-point is directly specified for each move, there are no intersection points for
Turbo PMAC to compute using the equations for the next move. This means there are no special
lookahead or single-step execution considerations, as there are in 2D compensation.
All moves in 3D compensation are directly blended together. There are no special considerations for
outside corners, as there are in 2D compensation. Also, there are no special considerations for the lead-in
and lead-out moves. The lead-in move is an interpolated move from the last uncompensated position to
the first compensated position. The lead-out move is an interpolated move from the last compensated
position to the first uncompensated position.
3D Cutter Radius Compensation
T
TR
N
CP
R
CC
TR-C
CR
PP
PP: Programmed Position
CP: Compensated Position
N: Surface-Normal Vector
T: Tool-Orientation Vector
CCR: Cutter’s End Radius
TR: Tool Shaft Radius
Turbo PMAC Lookahead Function
Turbo PMAC can perform highly sophisticated lookahead calculations on programmed trajectories to
ensure that the trajectories do not violate specified maximum quantities for the axes involved in the
moves. This permits writing the motion program simply to describe the commanded path. Vector
feedrate becomes a constraint instead of a command; programmed acceleration times are used only to
define corner sizes and minimum move block times. Turbo PMAC will control the speed along the path
automatically (but without changing the path) to ensure that axis limits are not violated.
Lookahead calculations are appropriate for any execution of a programmed path in which throughput has
been limited by the need to keep execution slow throughout the path because of the inability to anticipate
the few sections where slow execution is required. The lookahead function’s ability to anticipate these
problem areas permits much faster execution through most of the path, dramatically increasing
throughput.
Writing and Executing Motion Programs
307
Turbo PMAC User Manual
Because of the nature of the lookahead calculations – trajectory calculations are done well in advance of
the actual move execution, and moves are kept within machine limits by the automatic adjustment of
move speeds and times – they are not appropriate for some applications. Any application requiring quick
reaction to external conditions should not use lookahead. Also, also any application requiring precise
synchronization to external motion, such as those using PMAC’s external time base feature, should not
use lookahead.
When the lookahead function is enabled, Turbo PMAC will scan ahead in the programmed trajectories,
looking for potential violations of its position, velocity, and acceleration limits. If it sees a violation, it
will then work backward through the pre-computed buffered trajectories, slowing down the parts of these
trajectories necessary to keep the moves within limits. These calculations are completed before these
sections of the trajectory are actually executed.
Turbo PMAC can perform these lookahead calculations on LINEAR and CIRCLE mode moves. The
coordinate system must be put in segmentation mode (Isx13 > 0) to enable lookahead calculations, even if
only LINEAR mode moves are used. (The coordinate system must be in segmentation mode anyway to
execute CIRCLE mode moves or cutter radius compensation.) In segmentation mode, Turbo PMAC
automatically splits the moves into small segments, which are executed as a series of smooth splines to recreate the programmed moves.
Turbo PMAC stores data on these segments in a specially defined lookahead buffer for the coordinate
system. Each segment takes Isx13 milliseconds when it is put into the buffer, but this time can be
extended if it or some other segment in the buffer violates a velocity or acceleration limit.
This technique permits Turbo PMAC to create deceleration slopes in the middle of programmed moves, at
the boundaries of programmed move, or over multiple programmed moves, whichever is required to
create the fastest possible move that does not violate constraints. All of this is done automatically and
invisibly inside the Turbo PMAC; the part programmer and operator do not need to understand the
workings of the algorithm.
The following diagram shows the principle of how the lookahead function can create acceleration and
deceleration profiles automatically over multiple programmed moves. In this case, the programmed
moves are too short to permit the full acceleration to and from programmed speed in a single programmed
move. Without any change to the motion program, the lookahead function will create a profile that does
not violate acceleration constraints.
Lookahead for Multi-Block Accel / Decel
V
ACCEL & DECEL TIME MUST BE EQUAL TO
OR SHORTER THAN BLOCK MOVE TIME
Before Lookahead
F100
X1
X2
X3
X4
X5
X6
X7
X8
X9
X10
DWELL50
1 2 3 4 5 6 7 8 9 10
t
V
After Lookahead
ACCEL & DECEL ARE CONTROLLED BY
SPECIFIED ACCEL LIMIT (G’s)
1
2
3 4 5 6 7 8
9
10
t
308
Writing and Executing Motion Programs
Turbo PMAC User Manual
The next diagram shows how the lookahead function can create a deceleration into a tight corner
automatically, permitting the corner to be taken slowly to keep it within acceleration constraints, and then
to accelerate back up to the programmed speed coming out of the corner. This permits the user to
command high speeds, and to have Turbo PMAC slow down the path only where needed. Note that the
post-lookahead profile in this diagram is not time-extended as it would really be; this was done to show
the correspondence of points on the profiles.
Lookahead & Small, Tight Corners
Y
5
F50
X10 Y10
X0 Y20
4
VX
Time
3
1
2
3
4
5
VY
2
1
F50
X10 Y10
X0 Y20
Y
7
Time
Before Lookahead
X
VX
Time
6
5
3
ACCEL & DECEL
CONTROLLED BY
4 SPECIFIED G’s
AT CORNER
1
2
3
4
5
6
7
VY
2
1
X
After Lookahead
Time
Note how ‘Y” decelerates
to make the corner
If Turbo PMAC’s inverse kinematic calculations are used, the conversion from tip coordinates to joint
coordinates takes place before lookahead calculations, segment by segment for LINEAR and CIRCLE
mode moves. Therefore, Turbo PMAC can execute the lookahead calculations in joint space, motor by
motor, even if the system has been programmed in tip coordinates.
Once the lookahead function has been set up, the lookahead function operates transparently to the
programmer and the operator. No changes need to be made to a motion program to use the lookahead
function, although the programmer may choose to make some changes to take advantage of the increased
performance capabilities that lookahead provides.
Quick Instructions: Setting Up Lookahead
The following list quickly explains the steps required for setting up and using the lookahead function on
the Turbo PMAC. Greater detail and context are given in the subsequent section.
1. Assign all desired motors to the coordinate system with axis definition statements.
2. Set Ixx13 and Ixx14 positive and negative position limits, plus Ixx41 desired position-limit band, in
counts for each motor in coordinate system. Set bit 15 of Ixx24 to 1 to enable desired position limits.
3. Set Ixx16 maximum velocity in counts/msec for each motor in coordinate system.
4. Set Ixx17 maximum acceleration in counts/msec2 for each motor in coordinate system.
5. Set Isx13 segmentation time in msec for the coordinate system to minimum programmed move block
time or 10 msec, whichever is less.
6. Compute maximum stopping time for each motor as Ixx16/Ixx17.
7. Select motor with longest stopping time.
8. Compute number of segments needed to look ahead as this stopping time divided by (2 * Isx13).
9. Multiply the segments needed by 4/3 (round up if necessary) and set the Isx20 lookahead length
parameter to this value.
Writing and Executing Motion Programs
309
Turbo PMAC User Manual
10. If the application involves high block rates, set the Isx87 default acceleration time to the minimum
block time in msec; the Isx88 default S-curve time to 0.
11. If the application does not involve high block rates, set the Isx87 default acceleration time and the
Isx88 default S-curve time parameters to values that give the desired blending corner size and shape
at the programmed speeds.
12. Store these parameters to non-volatile memory with the SAVE command if you want them to be an
automatic part of the machine state.
13. After each power-up/reset, send the card a DEFINE LOOKAHEAD {# of segments},{# of
outputs} command for the coordinate system, where {# of segments} is equal to Isx20 plus
any segments for which backup capability is desired, and {# of outputs} is at least equal to the
number of synchronous M-variable assignments that may need to be buffered over the lookahead
length.
14. Load your motion program into the Turbo PMAC. Nothing special needs to be done to the motion
program. The motion program defines the path to be followed; the lookahead algorithm may reduce
the speed along the path, but it will not change the path.
15. Run the motion program, and let the lookahead algorithm do its work.
Detailed Instructions: Setting up to use Lookahead
A few steps are required to calculate and set up the lookahead function. Typically, the calculations have
to be done only once in the initial configuration of the machine. Once configured, the lookahead function
operates automatically and invisibly.
Defining the Coordinate System
The lookahead function checks the programmed moves against all motors in the coordinate system.
Therefore, the first step is to define the coordinate system by assigning motors to axes in the coordinate
system with axis definition statements. This action is covered in the Setting up the Coordinate System
section of this manual.
Lookahead Constraints
Turbo PMAC’s lookahead algorithm forces the coordinate system to observe four constraints for each
motor. These constraints are defined in I-variables for each motor representing maximum position
extents, velocities, and accelerations. These I-variables must be set up properly in order for the lookahead
algorithm to work properly.
Position Limits
Variables Ixx13 and Ixx14 for each Motor xx define the maximum positive and negative position values,
respectively, that are permitted for the motor (software overtravel limits). These variables are defined in
counts, and are referenced to the motor zero, or home, position (often called machine zero). Even if the
origin of the axis for programming purposes has been offset (often called program zero), the physical
position of these position limits does not change; they maintain their reference to the machine zero point.
Turbo PMAC checks the actual position for each motor as the trajectory is being executed against these
limits; if a limit is exceeded, the program is aborted and the motors are decelerated at the rate set by
Ixx15.
Variable Ixx41 for each Motor xx defines the distance between the actual position limits explained above,
and the desired position limit that can be checked at move calculation time, even in lookahead. That is, if
the calculated desired move position is greater than (Ixx13 – Ixx41), or less than (Ixx14 + Ixx41), this
will constitute a desired position limit violation. Desired position limits are checked only if bit 15 of
Ixx24 is set to 1.
310
Writing and Executing Motion Programs
Turbo PMAC User Manual
In this mode, if the lookahead algorithm, while scanning ahead in the programmed trajectory, determines
that any motor in the coordinate system would exceed one of its desired position limits, it will suspend the
program and force a stop right at that limit. It will then work backwards through the buffered trajectory
segments to bring the motors to a stop along the path at that point in the minimum time that does not
violate any motor’s Ixx17 acceleration constraint.
However, if bit 14 of Ixx24 is also set to 1, the program does not stop at the limit. Instead, it will
continue, with the offending motor saturating at the limit value.
When stopped on a desired position limit within lookahead, the program is only suspended, not aborted.
The action is equivalent to issuing a \ quick-stop command. It is possible to “retrace” the path coming
into the limit, or even to resume forward execution after changing the limit value. An “abort” command
must be issued before another program can be started.
Note, however, that if an actual position limit is also tripped during the deceleration to a stop at the
desired position limit, the program is aborted, so retracing and resuming are not possible. For this reason,
if the possibility of retracing and resuming is important, Ixx41 should be set to a large enough value so
that the actual position limit is never tripped during a desired position limit stop.
This technique permits these software position limits to be placed just within the hard stops of the
machine. Without the desired position limits, the software position limits cannot be detected until the
actual trajectory actually passes the limit. This requires that these limits be placed far enough within the
hard stops so that the motors have enough distance to stop after they pass the limits. (When a motor hits a
software position limit without lookahead, the deceleration of motors is controlled by Ixx15, not Ixx17,
and deceleration is not necessarily along the programmed path.)
Velocity Limits
Variable Ixx16 for each Motor xx defines the magnitude of the maximum velocity permitted for the
motor. These variables are defined in the raw PMAC units of counts per millisecond, so a quick
conversion must be calculated from the user units (e.g. millimeters per minute).
If the algorithm, while looking ahead in the programmed trajectory, determines that any motor in the
coordinate system is being asked to violate its velocity limit, it will slow down the trajectory at that point
just enough so that no limit is violated. It will then work backwards through the buffered trajectory
segments to create a controlled deceleration along the path to this limited speed in the minimum time that
does not violate any motor’s Ixx17 acceleration constraint.
Note:
During the initial move-block calculations, before move data is sent to the
lookahead function, a couple of factors can result in commanded velocities lower
than what is programmed. First, if the vector feedrate commanded in the motion
program with the F command exceeds the maximum feedrate parameter Isx98,
then Isx98 is used instead. Second, if the move-block time, either specified
directly with the TM command, or calculated as vector-distance divided by vectorfeedrate, is less than the programmed acceleration time (the larger of TA or 2 *
TS), the programmed acceleration time is used instead. This results in a speed less
than what was programmed. The lookahead function can further slow these
moves, but it cannot speed them up.
Acceleration Limits
Variable Ixx17 for each Motor xx defines the magnitude of the maximum acceleration permitted for the
motor. These variables are defined in the raw PMAC units of counts per (millisecond-squared), so a
quick conversion must be calculated from the user units (e.g. in/sec2, or g’s).
Writing and Executing Motion Programs
311
Turbo PMAC User Manual
If the algorithm, while looking ahead in the programmed trajectory, determines that any motor in the
coordinate system is being asked to violate its acceleration limit, it will slow down the trajectory at that
point just enough so that no limit is violated. It will then work backwards through the buffered trajectory
segments to create a controlled deceleration along the path to this limited speed in the minimum time that
does not violate any motor’s Ixx17 acceleration constraint.
Calculating the Segmentation Time
Turbo PMAC’s lookahead function operates on intermediate motion segments calculated from the
programmed trajectory. An intermediate point for each motor is computed once per segment from the
programmed path, and then a fine interpolation using a cubic spline to join these segments is executed at
the servo update rate. Therefore, the user settable segmentation time is an important parameter for
optimization of the lookahead function.
Variable Isx13 for each Coordinate System x defines the time for each intermediate segment in the
programmed trajectory, in milliseconds, before it is possibly extended by the lookahead function. Isx13
is an integer value; if a non-integer value is sent, Turbo PMAC will round to the next integer. If Isx13 is
set to 0, the coordinate system is not in segmentation mode; no intermediate segments are calculated, and
the lookahead function cannot be enabled.
Several issues must be addressed in setting the Isx13 segmentation time. These include its relationship to
the maximum block rate, the small interpolation errors it introduces, and its effect on the calculation load
of the Turbo PMAC. Each of these is addressed in turn, below.
Block Rate Relationship
In most applications, the Isx13 segmentation time will be set so that it is less than or equal to the
minimum block (programmed move) time. Put another way, usually the segmentation rate defined by
Isx13 is set greater than or equal to the maximum block rate. For example, if a maximum block rate of
500 blocks per second is desired, the minimum block time is 2 milliseconds, and Isx13 is set to a value no
greater than 2.
This relationship holds because blocks of a smaller time than the segmentation time are skipped over as
Turbo PMAC looks for the next segment point. While this does not cause any errors, there