null  null
GE Fanuc Intelligent Platforms
Programmable Control Products
PACSystems™
Hot Standby CPU Redundancy
User’s Manual, GFK-2308C
March 2009
GFL-002
Warnings, Cautions, and Notes
as Used in this Publication
Warning
Warning notices are used in this publication to emphasize that hazardous voltages,
currents, temperatures, or other conditions that could cause personal injury exist in this
equipment or may be associated with its use.
In situations where inattention could cause either personal injury or damage to equipment,
a Warning notice is used.
Caution
Caution notices are used where equipment might be damaged if care is not taken.
Note
Notes merely call attention to information that is especially significant to understanding and
operating the equipment.
This document is based on information available at the time of its publication. While efforts
have been made to be accurate, the information contained herein does not purport to cover all
details or variations in hardware or software, nor to provide for every possible contingency in
connection with installation, operation, or maintenance. Features may be described herein
which are not present in all hardware and software systems. GE Fanuc assumes no obligation
of notice to holders of this document with respect to changes subsequently made.
GE Fanuc makes no representation or warranty, expressed, implied, or statutory with respect to,
and assumes no responsibility for the accuracy, completeness, sufficiency, or usefulness of the
information contained herein. No warranties of merchantability or fitness for purpose shall
apply.
The following are trademarks of GE Fanuc Intelligent Platforms
Alarm Master
Helpmate
ProLoop
Series Six
CIMPLICITY
CIMPLICITY 90–ADS
Logicmaster
PROMACRO
Series Three
Modelmaster
PowerMotion
VersaMax
CIMSTAR
Motion Mate
PowerTRAC
VersaPoint
Field Control
PACMotion
Series 90
VersaPro
GEnet
Genius
PACSystems
Proficy
Series Five
Series One
VuMaster
Workmaster
©Copyright 2004—2009 GE Fanuc Intelligent Platforms
All Rights Reserved
Contents
Introduction....................................................................................................................1-1
Hot Standby CPU Redundancy ........................................................................................ 1-1
PACSystems HSB Redundancy Feature Summary ......................................................... 1-2
Online Programming......................................................................................................... 1-3
On-Line Repair and System Upgrade............................................................................... 1-3
Definitions ......................................................................................................................... 1-4
Related Publications ......................................................................................................... 1-5
Hot Standby Redundancy Quick Start.........................................................................2-1
System Configuration ...................................................................................................3-1
Components of a Hot Standby Redundancy System ....................................................... 3-1
System Racks .......................................................................................................... 3-1
Redundancy CPU Modules...................................................................................... 3-2
Redundancy Memory Xchange Modules ................................................................. 3-4
Redundant I/O Systems ........................................................................................... 3-5
Local I/O ................................................................................................................... 3-5
CPU Redundancy Using Ethernet NIU Remote I/O ......................................................... 3-6
Dual Controller, Single LAN Systems ...................................................................... 3-6
Dual Controller, Dual LAN Systems......................................................................... 3-8
Genius Hot Standby Operation (RX7i systems only)...................................................... 3-10
Genius Output Control ........................................................................................... 3-10
Basic CPU Redundancy Using Genius I/O ............................................................ 3-11
Configuration Requirements ........................................................................................4-1
Using the Redundancy Wizards ....................................................................................... 4-2
Synchronizing the Hardware Configurations............................................................ 4-3
Hardware Configuration Parameters ................................................................................ 4-4
CPU Parameters ...................................................................................................... 4-4
Redundancy Memory Xchange Module Parameters ............................................. 4-10
Ethernet Interface Parameters............................................................................... 4-11
Rack Module Configuration Parameters ................................................................ 4-12
Genius Bus Configuration (RX7i only) ................................................................... 4-13
Adding Individual Variables to the Transfer Lists ........................................................... 4-14
Using the Variable Transfer List Report................................................................. 4-15
Storing (Downloading) Hardware Configuration ............................................................. 4-16
Run Mode Stores ............................................................................................................ 4-17
Dual RMS with Simultaneous Activation In Redundant Systems .......................... 4-17
Initial RMS Followed by Dual RMS ........................................................................ 4-18
RMS Operational Errors......................................................................................... 4-19
Behavior of EGD in a Dual RMS............................................................................ 4-20
Hardware Configuration and Logic Coupling ......................................................... 4-20
GFK-2308C
iii
Contents
Operation........................................................................................................................5-1
Powerup of a Redundancy CPU....................................................................................... 5-2
Synchronizing Redundant CPUs ...................................................................................... 5-3
Dual Synchronization ............................................................................................... 5-3
Resynchronization.................................................................................................... 5-3
Operation when a Redundancy Link is Removed.................................................... 5-3
%S References for CPU Redundancy.............................................................................. 5-4
Scan Synchronization ....................................................................................................... 5-5
Fail Wait Time ................................................................................................................... 5-6
Data Transfer .................................................................................................................... 5-7
Synchronization and Data Transfer Process ........................................................... 5-7
Estimating Data Transfer Time ................................................................................ 5-8
Programming a Data Transfer from Backup Unit to Active Unit
(SVC_REQs 27 and 28)......................................................................................... 5-11
Disabling Data Transfer Copy in Backup Unit (SVC_REQ 43).............................. 5-12
Validating the Backup Unit (SVC_REQ 43) ........................................................... 5-14
Switching Control to the Backup Unit ............................................................................. 5-15
Switching Times and Impact to Sweep Time ......................................................... 5-15
Commanding a Role Switch from the Application Program (SVC_REQ 26) ......... 5-15
Implementing Preferred Master Using SVC_REQ 26............................................ 5-16
STOP to RUN Mode Transition ...................................................................................... 5-17
RUN Disabled Mode ....................................................................................................... 5-17
Error Checking and Correction ....................................................................................... 5-18
Timer and PID Functions ................................................................................................ 5-19
Timed Contacts............................................................................................................... 5-19
Multiple I/O Scan Sets .................................................................................................... 5-19
Genius Bus Controller Switching (RX7i only) ................................................................. 5-20
Redundant IP Addresses................................................................................................ 5-21
Ethernet Global Data in an HSB Redundancy System................................................... 5-23
Ethernet Global Data Production ........................................................................... 5-23
Ethernet Global Data Consumption ....................................................................... 5-24
Fault Detection...............................................................................................................6-1
Fault Detection.................................................................................................................. 6-1
PLC Fault Table Messages for Redundancy.................................................................... 6-2
Redundancy Fault Group (138) ............................................................................... 6-2
Other Fault Groups .................................................................................................. 6-4
Fault Response................................................................................................................. 6-5
Redundancy Link Failures ................................................................................................ 6-6
Redundancy Memory Xchange Module Hardware Failure ...................................... 6-6
Redundancy Link Communications Failures............................................................ 6-6
iv
PACSystems™ Hot Standby CPU Redundancy User’s Manual–March 2009
GFK-2308C
Contents
Fault Actions in a CPU Redundancy System ................................................................... 6-7
Configuration of Fault Actions .................................................................................. 6-7
Configurable Fault Groups ....................................................................................... 6-8
Non-Configurable Fault Groups ............................................................................... 6-9
Fatal Faults on Both Units in the Same Sweep ....................................................... 6-9
Online Repair and System Upgrade ............................................................................... 6-10
On-Line Repair Recommendations........................................................................ 6-10
Hot Swapping of Modules (RX3i Systems Only) ................................................... 6-11
System CPU Upgrade............................................................................................ 6-11
Online Repair of the Genius Bus ........................................................................... 6-12
Repair of a Non-Synchronized Active Unit (NSAU) Split Control System ............. 6-12
GFK-2308C
Contents
v
Chapter Introduction
1
This manual is a reference to the hardware components, configuration, programming and
operation of Hot Standby CPU redundancy for the PACSystems RX3i and RX7i controllers.
The information in this manual is intended to supplement the system installation,
programming, and configuration information contained in the manuals listed under “Related
Publications” on page 1-5.
Hot Standby CPU Redundancy
Hot Standby CPU Redundancy allows a critical application or process to continue operating if
a failure occurs in any single component. A Hot Standby system uses two CPUs; an active
unit that actively controls the process, and a backup unit that is synchronized with the active
unit and can take over the process if it becomes necessary. The two units are synchronized
when both are in Run Mode, the backup unit has received the latest status and
synchronization information from the active unit via a redundancy link, and both are running
their logic solution in parallel.
Each unit must have a redundancy CPU and one or two Redundancy Memory Xchange
(RMX) modules. The redundancy communication paths are provided by one or two pairs of
RMX modules.
Note:
GE Fanuc strongly recommends using two pairs of RMX modules configured as
dual redundancy links. This practice eliminates the possibility of a single point of
failure that using only one pair of RMX modules presents.
Control automatically switches to the backup unit when a failure is detected in the active unit.
You can initiate a switch of control by activating a toggle switch on the RMX module or
activating a service request in the application program. When a user-initiated switch of control
occurs, the CPUs switch roles; the active unit becomes the backup unit and the backup unit
becomes active.
The system runs synchronously with a transfer of all control data that defines machine status
and any internal data needed to keep the two CPUs operating in sync. Critical control data
plus all redundant outputs must be included in the output data transfer. The transfer of data
from the active unit to the backup unit occurs twice per sweep. These CPU-to-CPU transfers
are checked for data integrity.
The Primary and Secondary units in a redundancy system must be in the same controller
family. An RX3i and an RX7i controller cannot function as a redundant pair.
GFK-2308C
1-1
1
PACSystems HSB Redundancy Feature Summary
Feature
RX3i Redundancy System
RX7i Redundancy System
Redundancy CPU
IC695CRU320
IC698CRE020, CRE030 or CRE040
Redundancy links
Two IC695RMX128 modules per link
Two IC698RMX016 modules per link
Two links (four RMX modules)
recommended per system
Two links (four RMX modules)
recommended per system
Single and redundant Ethernet remote
I/O LANs through ENIU
Single and redundant Ethernet remote
I/O LANs through ENIU
Redundancy I/O systems
supported
Single and Dual Genius Bus Networks
Expansion and remote racks
Failure recovery
Supported
Supported
Survives any one single point of failure.
Survives any one single point of failure
(excluding failures of Genius devices
and bus stubs)
Online repair of failed component
Online repair of failed component
Manual toggle switch for switching
control between active and backup units
Manual toggle switch for switching
control between active and backup units
Application-initiated role switching
Application-initiated role switching
Synchronized CPUs
Synchronized CPUs
One-scan switching
One scan-switching
Configurable transfer data size up to
2Mbytes
Configurable transfer data size up to
2Mbytes
RMX128 module has five redundancy
status LEDs (Link OK, Local Ready,
Local Active, Remote Ready, Remote
Active)
RMX016 module has five redundancy
status LEDs (Link OK, Local Ready,
Local Active, Remote Ready, Remote
Active)
Redundancy status bits and message
logging
Redundancy status bits and message
logging
Online programming
Supported
Supported
Diagnostics
Background diagnostics
Background diagnostics
Memory error checking and correction
(ECC) with single bit corrections and
multiple bit checking
Memory error checking and correction
(ECC) with single bit corrections and
multiple bit checking
1000 feet (304.8 meters)
1000 feet (304.8 meters)
Role switching
Bumpless switching from active
unit to backup unit
Redundancy status monitoring
Maximum fiber optic cable
distance supported between two
RMX modules used in
redundancy link
1-2
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
1
Online Programming
On-line changes to the application program are permitted in both the active unit and the
backup unit. The programming device must be connected to the unit in which changes are to
be made in order to make any on-line changes.
PACSystems releases 5.5 and later support run mode store (RMS) of the redundancy
transfer list. This capability allows you to add, delete or modify transfer list entries without
stopping the controllers.
Run mode stores are performed independently on both controllers. However, in a
synchronized system, the optional Dual RMS with Simultaneous Activation feature can be
used to defer activation of the newly stored application data until an RMS has been
performed on both units. Because the PLC sweeps are synchronized, both units will activate
the new logic and transfer lists on the same sweep. For additional information about the use
of this feature, refer to “Run Mode Stores” in Chapter 4.
On-Line Repair and System Upgrade
A Hot Standby CPU Redundancy system permits online repair of failed components without
disrupting the control application. A failed component can be replaced in either unit after first
removing power from the rack in which it is installed.
After replacing the component, returning power to the rack, and placing the CPU in Run
mode, the repaired unit synchronizes with the currently active unit. Upon successful
synchronization, the repaired unit becomes the backup unit.
RX7i Systems Only
The Redundancy CPU in each unit can be replaced with a different model in a similar
manner. For example, you may want to replace the CRE020 models with CRE030 or
CRE040 models, or CRE030 models with CRE040 models.
Caution
During normal operation, the primary and secondary units in an HSB
redundancy system must have the same CPU model type. Extended
operation with dissimilar CPU types is not allowed. Continued use of
dissimilar CPU types may result in timing issues during
synchronization.
The primary and secondary units with dissimilar CPU model types can be synchronized for a
limited time, for the purpose of system upgrade only. Fail wait times for the higher
performance CPU in a dissimilar redundant pair may need to be increased to allow
synchronization. Either model can be in the primary or secondary unit.
Online repair and system CPU upgrade are described in more detail in chapter 6.
GFK-2308C
Chapter 1 Introduction
1-3
1
Definitions
Active Unit
The unit that is currently controlling the process.
Backup Unit
The unit that is synchronized with the active unit and able to take over the process.
CPU Redundancy
A system with two PLC CPU units cooperating to control the same process.
Critical
Component
Components that acquire or distribute I/O data or that are involved in execution of the
control logic solution.
Genius Dual Bus
The use of two Genius busses to control the same I/O devices. The busses are linked to
the I/O devices by one or more Bus Switching Modules (BSMs). A BSM will automatically
switch to the other bus if the active bus has a failure.
Genius Hot
Standby
A feature of Genius devices whereby the device prefers output data from the Bus
Controller at SBA 31. When outputs from that Bus Controller are not available, the device
takes output data from the Bus Controller at SBA 30. If outputs from neither Controller are
available, the device places its outputs in the designated default state.
Hot Standby
A system where the backup (standby) unit is designated before any critical component
failure takes place, and any necessary state/control information is passed to this
designated backup unit so that it can take control quickly in the event of a critical
component failure.
Non-Synchronized A unit in a Redundancy System that is in Run mode but not synchronized with a backup
Active Unit
unit. The backup unit is either offline (in Stop mode, powered off, or failed), or there are
(NSAU)
no functional redundancy links between the units.
1-4
Primary Unit
The preferred unit to control the process in a Redundancy System. For redundant Genius
I/O, the Genius Bus Controllers in the primary unit are configured for serial bus address
(SBA) 31.
Redundancy
The use of multiple elements controlling the same process to provide alternate functional
channels in case of failure.
Redundancy Link
A complete communications path between the two CPUs, consisting of one RMX in the
primary unit, one RMX in the secondary unit, and a high-speed fiber optic cable
connecting them to each other.
Redundant IP
Address
An IP address that is assigned to the pair of Ethernet interfaces in the primary and
secondary units. All data sent to the redundant IP address (including EGD produced to
the redundant IP address) is handled by the active unit.
Role Switch
User-initiated switch of control, where the active unit becomes the backup unit and the
backup unit becomes the active unit.
Secondary Unit
The unit configured to control the process in a Redundancy System when the primary unit
is unavailable or otherwise marked as not controlling the process. For redundant Genius
I/O, the Genius Bus Controllers in the secondary unit are configured for SBA 30.
Synchronized
Condition where both units are in Run Mode and the backup unit has received the latest
status and synchronization information from the active unit via a redundancy link. When
the two units are synchronized, they run their logic solution in parallel. If the active unit
goes offline, control of the redundancy outputs is switched bumplessly (without
interruption) to the backup unit.
Transfer List
The ranges of references that will be transferred from the active unit to the backup unit.
The transfer list is selected in the hardware configuration for the Redundancy CPU.
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
1
Related Publications
PACSystems CPU Reference Manual, GFK-2222
PACSystems RX7i Installation Manual, GFK-2223
TCP/IP Ethernet Communications for PACSystems, GFK-2224
PACSystems RX7i User's Guide to Integration of VME Modules, GFK-2235
PACSystems Memory Exchange Modules, GFK-2300
PACSystems RX3i System Manual, GFK-2314
PACSystems RX3i Ethernet NIU User’s Manual, GFK-2439
Series 90-30 Ethernet NIU User’s Manual, GFK-2296
Genius® I/O System User's Manual, GEK-90486-1
Genius Discrete and Analog Blocks User's Manual, GEK-90486-2
Series 90™-70 Genius Bus Controller User's Manual, GFK-2017
Proficy™ Machine Edition Logic Developer-PLC Getting Started, GFK-1918
For the most recent versions of PACSystems and related documentation, visit the GE Fanuc
website: http://www.gefanuc.com/.
GFK-2308C
Chapter 1 Introduction
1-5
Chapter Hot Standby Redundancy Quick Start
2
This chapter provides an overview of the steps needed to configure and operate a basic RX3i
or RX7i Hot Standby (HSB) CPU Redundancy system with one Ethernet Remote IO (ENIU)
using a ten-ENIU Machine Edition template.
Notes: The Primary and Secondary units in a redundancy system must be of the same type.
An RX3i and an RX7i controller cannot function as a redundant pair.
1. Install one Redundancy CPU, one or two RMX modules and three Ethernet
modules each into two rack systems.
One Rack system will be designated the Primary rack and the other will be designated
the Secondary rack.
2. With the CPU battery disconnected, apply power to the racks.
When power is applied to the RMX module an internal loopback test occurs; the
OWN DATA and SIGNAL DETECT indicators turn on briefly during this test. When the
RMX module and the CPU are powered up and functioning properly, the RMX module’s
OK indicator is on.
3. Connect a battery to each redundancy CPU.
The redundancy CPUs support Error checking and correction (ECC) memory, which must
be initialized at least one time with the battery disconnected. Once ECC memory is
initialized, the CPU can be power cycled with the battery connected.
4. Download and unzip the appropriate template set for your system.
Templates for redundancy systems are available from the GE Fanuc Support website,
http://support.gefanuc.com. On the website, select Downloads, then select the
Developer Files category.
For a list of available template sets, refer to the PACSystems RX3i Ethernet NIU User’s
Manual, GFK-2439. Each template set consists of a controller template and an ENIU
template.
5. Using the Machine Edition Logic Developer software, restore the controller project
from the appropriate ten-ENIU template set.
GFK-2308C
2-1
2
6. Open the restored project. Assign IP addresses to all the Ethernet modules.
In assigning IP addresses, consider the following functions:
RX3i Configuration
Ethernet Interface
Function
ETM001 in Slot 6
Programmer connection to your PC
Requires a Redundant IP address, which should be the same IP Address
for both the Primary and Secondary rack systems.
ETM001 in Slot 7
Private network, LANA for Ethernet IO exchanges
ETM001 in Slot 8
Private network, LANB for Ethernet IO exchanges
RX7i Configuration
Ethernet Interface
Function
Embedded CPU
Ethernet Port
Programmer connection to your PC
Requires a Redundant IP address, which should be the same IP Address
for both the Primary and Secondary rack systems.
ETM001 in Slot 5
Private network, LANA for Ethernet IO exchanges
ETM001 in Slot 6
Private network, LANB for Ethernet IO exchanges
The hardware configuration should appear similar to the following figure, which shows an
RX3i configuration.
Hardware Configuration provided by the Ten-ENIU Controller Template Project
2-2
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
2
7. Use Fiber Optic cable to connect each RMX module in the Primary Rack to the
corresponding RMX module in the Secondary Rack (the module in the same Slot
number) as described below.
Using an LC- compatible multimode
fiber optic cable, connect the RMX
module’s TX connector to the RX
connector of the other RMX module.
Connect the fiber optic cable from
other RMX module’s TX to the RX
connector (see diagram at right).
TX
RMX Module
in Primary Rack
When the fiber optic transceiver
detects a signal on the network, the
SIGNAL DETECT indicator will
be on.
TX
RX
RMX Module
in Secondary
Rack
RX
8. In PME, close the Controller project and restore the ENIU project from the ten-ENIU
template set:
Open the project and on target ENIU_01 open the Hardware Configuration. Set the IP
addresses of the ETM001 modules, taking into consideration that the ETM001 in Slot 4
will be on a private network called LANA (connected to LANA of the Redundancy CPUs)
and the ETM001 in Slot 5 will be on a private network called LANB (connected to LANB
of the Redundancy CPUs).
The hardware configuration should appear similar to the following figure, which shows an
RX3i configuration.
Hardware Configuration provided by the Ten-ENIU Template Project
GFK-2308C
Chapter 2 Hot Standby Redundancy Quick Start
2-3
2
9. Add IO loopback logic to confirm data transfer between ENIU and Redundancy
CPUs
Under the Logic node in PME, open the Program Block “Local_User_Logic”. Add the
logic shown below to loop outputs %Q1-%Q16 back to inputs %I1-%I16.
10. Install a Power supply, RX3i ENIU (IC695NIU001) and two ETM001 modules into an
RX3i backplane as shown the hardware configuration in step 6. Apply power to the
system.
11. Connect your PC to the ENIU via a Serial cable from the ENIU module’s COM1 or
COM2 port to one of your PC’s COM ports or install an additional ETM001 module
to the ENIU rack to provide connectivity via Ethernet. With the template folder
open in PME, connect to the ENIU either by a COM port or by Ethernet.
Store the ENIU_01 application to the ENIU and put the ENIU into run mode.
12. Connect Ethernet cables between the Redundancy CPUs and the ENIU rack
system.
RX3i Connections
Connect one Ethernet cable from ETM001 in Primary Rack Slot 7 to ETM001 in ENIU
Rack Slot 4. Connect one Ethernet cable from ETM001 in Primary Rack Slot 8 to
ETM001 in ENIU Rack Slot 5.
Connect one Ethernet cable from ETM001 in Secondary Rack Slot 7 to ETM001 in ENIU
Rack Slot 4. Connect one Ethernet cable from ETM001 in Secondary Rack Slot 8 to
ETM001 in ENIU Rack Slot 5.
RX7i Connections:
Connect one Ethernet cable from ETM001 in Primary Rack Slot 5 to ETM001 in ENIU
Rack Slot 4. Connect one Ethernet cable from ETM001 in Primary Rack Slot 6 to
ETM001 in ENIU Rack Slot 5.
Connect one Ethernet cable from ETM001 in Secondary Rack Slot 5 to ETM001 in ENIU
Rack Slot 4. Connect one Ethernet cable from ETM001 in Secondary Rack Slot 6 to
ETM001 in ENIU Rack Slot 5.
2-4
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
2
13. Connect Ethernet cables between an Ethernet switch connected to your PC and
the ETM001 modules assigned as Programmer connections in both the Primary
and Secondary units.
14. Close _10ENIU_CRU_DLDI_ENIUs_1_10 project in PME and again open project
_10ENIU_CRU_DLDI_Controller.
Right click on the Primary Hardware Configuration node and select “Set as Selected
HWC.” Connect to the Primary CPU, store the application and put the CPU in run mode.
Disconnect from the Primary CPU. Right click on the Secondary Hardware Configuration
node and select “Set as Selected HWC.” Connect to the Secondary CPU, store the
application and put the CPU in run mode.
Right click on the “Reference View Tables” node and select “New.” Double click on the
“RefViewTable10” node just created. In the address box, enter %Q1. In the next
address box below %Q00001, enter %I1. Right click into the Values area just to the left
of the Address boxes and select “Format View Table.” Check the box labeled “Apply to
Whole Table”, Select “Word” for the Display Type, select “Hex” for the Display Format
and click OK. Enter values into the %Q00001 values area and notice that the same
values are displayed at %I00001 because of the loopback logic in the ENIU.
This quick start procedure demonstrates setup of a PACSystems Redundancy Controller pair
controlling one ENIU remote IO station. This basic setup can be used to learn about other
CPU Redundancy features such as Role Switching, Transfer Lists, Non-Synchronized Active
Unit (NSAU) and Redundant IP. For details on the operation of CPU Redundancy systems,
refer to the other chapters in this manual.
GFK-2308C
Chapter 2 Hot Standby Redundancy Quick Start
2-5
Chapter System Configuration
3
This chapter describes the hardware components for a Hot Standby CPU Redundancy
system and describes system configurations for the basic redundancy schemes supported by
PACSystems controllers.
For installation instructions, refer to
PACSystems RX7i Installation Manual, GFK-2223
PACSystems RX3i System Manual, GFK-2314
Components of a Hot Standby Redundancy System
System Racks
Redundancy CPU
Redundancy Memory Xchange modules
Redundant I/O System
Genius Bus Controller and Genius Devices (RX7i Systems only)
System Racks
RX3i Systems
In an RX3i redundancy system, an RX3i (IC695CHS0xx) Universal Backplane must be used
as the CPU rack, which is also referred to as Rack 0. For specific backplane versions
required, refer to the Important Product Information document provided with your
redundancy CPU.
Any RX3i expansion rack or any Series 90-30 expansion rack that is supported by RX3i can
be used in an RX3i redundancy system.
RX7i Systems
In an RX7i redundancy system, any RX7i (IC698CHSxxx) rack can be used as Rack 0.
Any Series 90-70 expansion rack that is supported by RX7i can be used, except for the VME
Integrator racks, IC697CHS782 and IC697CHS783.
GFK-2308C
3-1
3
Redundancy CPU Modules
To use the features described in this manual, an RX7i Redundancy CPU module must be
installed in rack 0, slot 1 of both the primary and secondary units. RX3i Redundancy CPUs
can be installed in any slot in rack 0.
Note:
A given feature may not be implemented on all PACSystems CPUs. To determine
whether a feature is available on a given CPU model and firmware version, please
refer to the Important Product Information (IPI) document provided with the CPU.
The CPU provides configurable reference memory limits for %AI (Analog Input), %AQ
(Analog Output), %R (Register), and %W (bulk memory area) reference memory, as well as
symbolic discrete reference memory and symbolic non-discrete reference memory. For
additional CPU features and performance specifications, refer to the PACSystems CPU
Reference Manual, GFK-2222.
Operation of the CPUs can be controlled by the three-position RUN/STOP switch or remotely
by an attached programmer and programming software. Program and configuration data can
be locked through software passwords. The LEDs on the front of the module indicate CPU
and Ethernet interface status.
The CPUs have two configurable ports: COM 1 (RS-232) and COM2 (RS-485). The RX7i
CPUs contain an embedded Ethernet interface board that controls two
10 BASE T/100 BASE TX ports and a configurable Station Manager (RS-232) port.
PACSystems CPUs support the following Ethernet interface features:
3-2
■
Redundant IP address
■
Production of selected EGD exchanges in backup mode
■
RX7i controller data monitoring over the web. Supports a combined total of up to 16 web
server and FTP connections.
■
Up to 255 Ethernet Global Data (EGD) exchanges with up to 100 variables per exchange.
■
EGD upload and selective consumption of EGD exchanges.
■
Upload and download of an Advanced User Parameter (AUP) file, which contains user
customizations to internal Ethernet operating parameters.
■
Run mode store of EGD (PACSystems releases 5.5 and later), which allows you to add,
delete or modify EGD exchanges without stopping the controller. For details on using this
feature, refer to TCP/IP Ethernet Communications for PACSystems, GFK-2224.
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
3
Redundancy CPUs Compared to Other PACSystems CPUs
The following features are not available:
■
I/O and module interrupts: This includes the single edge triggered interrupts from the
discrete input modules, the high alarm and low alarm interrupts from the analog input
modules, and interrupts from VME modules. A program that declares I/O Interrupt
triggers cannot be stored to a Redundancy CPU.
■
Interrupt Blocks (I/O, timed, module): Logic that contains interrupt blocks cannot be
stored to the CPU.
■
Stop I/O Scan mode: If an attempt is made to place the PLC in this mode, the PLC will
reject the selection and return an error.
■
#OVR_PRE %S reference, which indicates whether one or more overrides are active, is
not supported and should not be used.
■
RX3i redundancy controllers do not support Genius Bus Controller (IC694BEM331).
■
RX3i redundancy controllers do not support the PACMotion module (IC695PMM335).
■
RX7i redundancy controllers do not support the 14-point interrupt module
(IC697MDL671).
■
RX7i redundancy controllers do not support VME integrator racks.
The following features operate differently with the redundancy CPUs than they do with other
PACSystems CPUs:
■
Error checking and correction (ECC) is enabled.
■
RUN/DISABLED mode. This is explained in Chapter 5, “Operation.”
■
User-configurable fault actions are not used when the CPUs are synchronized.
■
STOP to RUN mode transition. For details, see “Synchronizing Redundant CPUs” in
chapter 5.
■
Background Window Timer (in Normal Sweep mode) default is 5ms. It is highly
recommended that the Background Window Timer be set to the same value for both
CPUs making up a redundancy pair.
■
By default, Ethernet Global Data (EGD) is produced only by the active unit. The backup
unit can produce individual EGD exchanges that are configured for production in backup
mode.
Also, be aware that instance data associated with IEC transitionals (PTCOIL, NTCOIL,
PTCON, and NTCON) is not synchronized between the two CPUs. For details, refer to “Data
Transfer” in chapter 5.
Using the Redundancy CPU for Non-redundant Operation
The Redundancy CPU can be used for both redundant and non-redundant applications. The
functionality and performance of a Redundancy CPU configured for non-redundant operation
is the same as for a unit that is configured for redundant operation with no backup available.
This includes the redundancy informational messages such as those generated when a unit
goes to Run mode. Refer to "Configuring the Redundancy CPU for Non-redundant
Operation" in Chapter 4.
GFK-2308C
Chapter 3 System Configuration
3-3
3
Redundancy Memory Xchange Modules
The RMX modules provide a path for transferring data between the two redundancy CPUs. A
complete communications path consists of one RMX in the primary unit, one RMX in the
secondary unit, and two high-speed fiber optic cables connecting them to each other. This
must be a two-node ring: no other reflective memory nodes are allowed to be part of this fiber
optic network.
GE Fanuc strongly recommends that two redundancy links (for a total of four RMX
modules) be configured and installed. Optionally, systems can be configured for a single
redundancy link (for a total of two RMX modules).
RMX modules must be installed in the main rack (rack 0).
The RMX module has a toggle switch that can be used to manually request a role switch.
Eight LEDs, described in the following table, provide indication of module status.
Note:
For RX7i systems, it is recommended that the RMX modules be installed in slots 3
and 4 of the main rack. This gives VME interrupt request priority to the RMX
modules. Although this configuration is recommended, it is not required that the RMX
modules be located in slots 3 and 4.
Note:
The RX3i RMX128 module supports hot insertion and removal. However, the
redundancy communication link associated with a hot swapped RMX module will not
be restored automatically. The LINK OK indicator on both RMX modules in the link
will be OFF. To restore the link, refer to “Online Repair” in Chapter 6.
RMX LEDs
LED Label
3-4
Description
OK
ON indicates the module is functioning properly.
LINK OK
When used as a redundancy link, ON indicates the link is
functioning properly.
LOCAL READY
ON indicates the local unit is ready.
LOCAL ACTIVE
ON indicates the local unit is active.
REMOTE READY
ON indicates the remote unit is ready.
REMOTE ACTIVE
ON indicates the remote unit is active.
OWN DATA
ON indicates the module has received its own data packet from
the network at least once.
SIGNAL DETECT
SIG DETECT
ON indicates the receiver is detecting a fiber optic signal.
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
3
Redundant I/O Systems
Ethernet Network Interface Unit (ENIU)
CPU-based ENIU modules can be used to interface the RX7i or RX3i Redundancy CPU to
remote I/O stations through Ethernet LANs. These devices, which include IC695NIU001 and
IC693NIU004, make it possible to use PACSystems RX3i and Series 90-30 I/O remotely on
an Ethernet network.
An identical set of EGD exchange definitions is downloaded to both the primary and
secondary controllers. An ENIU can consume EGD exchanges from two controllers
simultaneously. However, when used with redundant controllers, the ENIU automatically
switches to the standby controller if the active controller becomes unavailable.
For sample redundancy systems using EGD, see page 3-6. For details on EGD operation in a
redundancy system, see “Ethernet Global Data in an HSB Redundancy System” in chapter 5.
For details on the operation of ENIUs, see the PACSystems RX3i Ethernet NIU User’s
Manual, GFK-2439
Genius Bus Controller and Genius Devices – RX7i Only
The Genius Bus Controller (IC697BEM731 or IC687BEM731) interfaces the RX7i
Redundancy CPU to a Genius I/O bus. The bus controller scans Genius devices
asynchronously and exchanges I/O data with the CPU.
An HSB CPU Redundancy system can have multiple Genius I/O bus networks. Any Genius
device can be placed on the bus (Genius blocks, Field Control, Remote I/O Scanner,
VersaMax I/O, etc.). The Genius outputs are determined by the active unit. The Genius Bus
Controller in the primary unit has a Serial Bus Address of 31; the Genius Bus Controller in the
secondary unit has a Serial Bus Address of 30. For sample redundancy systems using
Genius I/O, see page 3-9.
Local I/O
Local I/O can be included in either unit; however, it is not part of the redundant I/O system. A
failure in the Local I/O system will affect the unit as described in the PACSystems CPU
Reference Manual, GFK-2222.
GFK-2308C
Chapter 3 System Configuration
3-5
3
CPU Redundancy Using Ethernet NIU Remote I/O
This section discusses sample system architectures using Ethernet remote I/O with CPU
hot-standby redundancy systems.
These sample system architectures support both general communications (such as a
programmer connection) and remote I/O data transfers. Remote I/O data transfers use EGD
to and from the ENIUs.
For general communication in a hot-standby redundancy system, the Redundant IP feature
must be enabled for the Ethernet interface. In general communication, only the active CPU
produces EGD exchanges. When a redundancy role switch occurs, the backup CPU
becomes active and begins producing EGD. The formerly active CPU switches to backup and
stops producing EGD.
For remote I/O operation, the active and backup CPUs simultaneously process remote I/O
EGD exchanges for each ENIU. For architectures using redundant remote I/O LANs, the
CPUs process separate remote I/O EGD exchanges on each LAN. All EGD exchanges that
can simultaneously occur on a network must have unique Exchange IDs. Hence remote I/O
exchanges that are produced by both the primary and secondary units must have different
Exchange ID values. Remote I/O EGD production continues across CPU role switches. The
application logic in the ENIU selects which EGD remote I/O output exchanges to consume for
controlling outputs.
If the active controller transitions to Run IO Disabled mode, it continues to receive inputs from
the ENIU. However the ENIU no longer receives outputs from the controller. The ENIU’s
status words can be monitored to detect communication activity. For details on the status
words, refer to PACSystems RX3i Ethernet NIU User’s Manual, GFK-2439.
Note:
These architectures are based on the template sets provided for use with Proficy
Machine Edition and Proficy Process Systems programmers. The templates are set
up with coordinated references and coordinated parameters for 10, 20, or 24 ENIUs.
For systems with other numbers of ENIUs, select the template with the next larger
number of ENIUs and delete the extra ENIUs.
For details about the ENIU configuration and operation and use of the ENIU
templates, refer to the PACSystems RX3i Ethernet NIU User’s Manual, GFK-2439.
Dual Controller, Single LAN Systems
The following template sets are available to configure these architectures.
Architecture
Templates for
Proficy Machine Edition
Templates for
Proficy Process Systems
Dual RX7i CRE Controllers,
Single LAN
NA
10 ENIUs,
20 ENIUs
In this architecture, general communications and remote I/O data transfer exist on the same
Ethernet LAN and thus contend for network bandwidth. This contention may degrade remote
I/O performance.
The Redundant IP feature is enabled for the Ethernet interface in both controllers to permit
general communications. Any EGD exchanges used for general CPU communications are
not produced in backup mode.
The produced EGD exchanges that are used for remote I/O data transfer are configured as
“Produce in backup mode” so that they will be produced in both active and backup mode.
3-6
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
3
For easier configuration, each EGD exchange marked as “Produce in backup” is configured
with the Exchange ID value used by the Primary unit. The Programmer automatically
generates a unique Exchange ID value for the Secondary unit by adding the configured
“Secondary Produced Exchange Offset” value to the configured Exchange ID value. For
details on the exchange offset, see “Ethernet Global Data Production” in chapter 5.
RX7i Dual Controller, Single LAN System
Secondary Unit
RMX
RMX
ETM
CRE
RMX
RMX
ETM
Power Supply
Required fan
assembly
not shown
CRE
RX7i
(Rack 0)
Power Supply
Primary Unit
High Speed Fiber Optic Link
High Speed Fiber Optic Link
Ethernet LAN 1
Ethernet LAN 2
RX3i ENIU Remote IO
N
I
U
GFK-2308C
E
T
M
Chapter 3 System Configuration
RX3i ENIU Remote IO
N
I
U
E
T
M
3-7
3
Dual Controller, Dual LAN Systems
The following template sets are available to configure these architectures.
Architecture
Templates for
Proficy Machine Edition
Templates for
Proficy Process Systems
Dual RX3i CRU Controllers,
Dual LAN
10 ENIUs,
20 ENIUs
NA
Dual RX7i CRE Controllers,
Dual LAN
10 ENIUs,
24 ENIUs
10 ENIUs,
20 ENIUs
In this system architecture, the remote I/O stations each have two Ethernet modules to
provide the stations with redundant LAN connections to the controllers. LAN 3 acts as a
backup to LAN 2.
The Redundant IP feature is enabled for the Ethernet interfaces on LAN 1 because it handles
general communications. EGD exchanges used for general CPU communications are not
produced in backup mode.
Each controller uses a separate Ethernet interface for communication on each remote I/O
LAN (one for LAN 2 and another for LAN 3). The remote I/O EGD exchanges are configured
on the Ethernet interfaces for the appropriate LAN.
The following diagrams show the RX7i and RX3i Dual Controller, Dual LAN architectures.
RX7i Dual Controller, Dual LAN System
ETM 2’
ETM 1’
CRE
RMX
RMX
ETM 2
ETM 1
CRE
Secondary Unit
Power Supply
Required fan
assembly not
shown
RMX
RMX
RX7i
(Rack 0)
Power Supply
Primary Unit
High Speed Fiber Optic Link
High Speed Fiber Optic Link
Ethernet LAN 1
Ethernet LAN 2
Ethernet LAN 3
RX3i ENIU Remote IO
ETM B’
ETM A’
NIU
ETM B
NIU
3-8
ETM A
RX3i ENIU Remote IO
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
3
RX3i Dual Controller, Dual LAN System
ETM1'
ETM3'
ETM2'
RMX
RMX
PWS
CRU
Secondary Unit
ETM1
ETM3
RMX
ETM2
CRU
RMX
PWS
Primary Unit
High speed Fiber Optic Link
High Speed Fiber Optic Link
Ethernet LAN-1
Ethernet LAN-2
Ethernet LAN-3
GFK-2308C
Chapter 3 System Configuration
ETM
NIU
ETM
RX3i ENIU Remote I/O
PWS
ETM
NIU
ETM
PWS
RX3i ENIU Remote I/O
3-9
3
Genius Hot Standby Operation (RX7i systems only)
In a Genius Hot Standby CPU redundancy system, the Genius outputs are controlled by only
one unit (the active unit). The inputs are shared between both units. One unit is the Primary
unit and the other is the Secondary unit. The Primary unit contains all externally redundant
Genius Bus Controllers at SBA 31; the Secondary unit contains all externally redundant
Genius Bus Controllers at SBA 30.
The Genius output devices are normally configured for Genius Hot Standby redundant
operation. With this configuration, the devices choose between outputs from the Genius Bus
Controller at SBA 31 and the Genius Bus Controller at SBA 30. If outputs from both Genius
Bus Controllers are available, the devices will use outputs from SBA 31. If there are no
outputs from SBA 31 for three consecutive Genius I/O bus scans, the devices will use the
outputs from SBA 30. If outputs are not available from either SBA 31 or 30, the outputs go to
their configured default (OFF or hold last state).
Genius Output Control
In a Genius Hot Standby CPU Redundancy system, the active unit determines the values of
the Genius outputs.
Both the primary and secondary units send outputs regardless of which one is active. The
user is responsible for ensuring that all redundant Genius outputs are included in the output
data transfer. Because the same output values will then be sent to the GBCs in both units,
the devices will receive the same output values from SBA 31 and SBA 30. There is no data
interruption on switchover because both units are always sending Genius outputs.
GBC
RMX
CRE
RMX
Power Supply
Primary Unit
GBC
RMX
CRE
RMX
RX7i
(Rack 0)
Power Supply
Secondary Unit
Required fan assembly
not shown.
30
31
High-speed fiber optic link
High-speed fiber optic link
Genius Bus
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
IO
IO
IO
IO
IO
IO
IO
PS
3-10
Scanner
IO
Remote Drop
GFK-2308C
3
Basic CPU Redundancy Using Genius I/O
Hot Standby CPU Redundancy supports two types of bus schemes for the Genius networks:
■
Single bus networks
■
Dual bus networks
PACSystems CPU Redundancy implements a floating master algorithm. If an application
requires a preferred master algorithm, see “Implementing Preferred Master” in chapter 5.
Single Bus Networks
This type of network uses a single bus with one Genius bus controller in each PLC.
Required fan
assembly
not shown
30
GBC
CPU
RMX
RMX
Power Supply
Primary Unit
GBC
CPU
RMX
RMX
RX7i
(Rack 0)
Power Supply
Secondary Unit
31
High Speed Fiber Optic Link
High Speed Fiber Optic Link
Genius Bus
Legend
CPU -
RX7i CPU
RMX -
Redundancy Memory Xchange Module
GBC -
Genius Bus Controller
Genius Blocks
The single bus setup is suitable if the application does not require redundant I/O busses.
When using single-bus Genius networks in a Hot Standby CPU Redundancy system, one
Genius Bus Controller for the bus must be located in the primary unit and one in the
secondary unit. There can be multiple Genius busses in the system. The bus controllers in
the primary unit are assigned Serial Bus Address 31. The bus controllers in the secondary
unit are assigned Serial Bus Address 30.
Genius output devices will use outputs from Serial bus Address 31 in preference to outputs
from Serial bus Address 30. Outputs are determined by the active unit, regardless of which
bus controller provides the outputs since all redundant Genius outputs are transferred from
the active unit to the backup unit.
Any type of Genius device can be connected to the network. Each Genius network can have
up to 30 additional Genius devices connected to it. You may want to reserve one Serial Bus
Address for the Hand-Held Monitor.
As a safety feature, a watchdog timer protects each Genius I/O link. The bus controller
periodically resets this timer. If the timer expires, the bus controller stops sending outputs. If
this happens in a Single Bus Genius network of a CPU Redundancy system, the paired GBC
in the other unit drives the outputs of the Genius devices. The cause of the failure must be
remedied to re-establish communications.
GFK-2308C
Chapter 3 System Configuration
3-11
3
Hardware Configuration for Single Bus Networks
The hardware configuration for single bus networks can be created by selecting Redundant
Controllers, Two PLCs in the Redundancy Wizard.
The GBCs must be configured with the following settings.
Redundant Mode: Redundant Controllers
Paired GBC: External
Serial Bus A: 31 (primary unit) or 30 (secondary unit)
The redundant devices must be configured for Hot Standby mode. For example, use the
following settings for a Genius block:
(Programming software) Redundancy = YES
(Hand-Held Monitor) CPU Redundancy = HOT STBY MODE
(Hand-Held Monitor) BSM Present = NO
Dual Bus Networks
This option provides redundancy of both the PLC and the I/O bus. This type of system uses
dual busses with bus controllers in each PLC. A Bus Switching Module (BSM) is required to
connect the initial block in the Genius block daisy chain to the dual bus.
GBC
GBC
CPU
30
Required fan
assembly
not shown.
RMX
RMX
GBC
30
Primary Unit
Power Supply
GBC
CPU
RMX
RMX
RX7i
(Rack 0)
Power Supply
Secondary Unit
31
31
High-speed fiber optic link
High-speed fiber optic link
Genius Bus
Legend
CPU RMX GBC -
RX7i CPU
Redundancy Memory Xchange Module
Genius Bus Controller
Bus Switching
Module
Genius I/O Blocks
The Dual Bus network is suitable if the application requires redundancy of the PLC and the
I/O bus.
When using dual bus Genius networks in a Hot Standby CPU Redundancy system, two Bus
Controllers for the bus pair must be located in the primary unit and two more in the secondary
unit. There can be multiple dual bus pairs. The bus controllers in the primary unit are
assigned Serial Bus Address 31. The bus controllers in the secondary unit are assigned
Serial Bus Address 30.
3-12
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
3
Genius output devices will use outputs from Serial bus Address 31 in preference to outputs
from Serial bus Address 30. Outputs are determined by the active unit, regardless of which
bus controller provides the outputs since all redundant Genius outputs are transferred from
the active unit to the backup unit.
Any type of Genius device can be connected to the network. Each Genius network can have
up to 30 additional Genius devices connected to it. You may want to reserve one Serial Bus
Address for the Hand-Held Monitor.
As a safety feature, a watchdog timer protects each Genius I/O link. The bus controller
periodically resets this timer. If the timer expires, the bus controller stops sending outputs. If
this happens in a Dual Bus Genius network of a CPU Redundancy system, the paired GBC in
the other unit drives the outputs of the Genius devices. If the GBC in the other unit is not
available, the BSMs switch to the other bus. The cause of the failure must be remedied to reestablish communications.
Hardware Configuration for Dual Bus Network
The hardware configuration for this type of network can be created by selecting Dual Bus,
Redundant Controllers in the Redundancy Wizard.
The GBCs must be configured with the following settings
Redundant Mode: Dual Bus_Redundant Controllers
Paired GBC = Internal and External
Serial Bus Address = 31 (primary unit) or 30 (secondary unit)
The redundant devices must be configured for Hot Standby and dual bus mode. For example,
use the following settings for a Genius block:
(Programming Software) Redundancy = YES
(Hand-Held Monitor) CPU Redundancy = HOT STBY MODE*
(Hand-Held Monitor) BSM Present = YES
(Hand-Held Monitor) BSM Controller = YES (if BSM is mounted) or NO
Location of GBCs and Blocks
For fastest switching, all Genius Bus Controllers in the Hot Standby CPU Redundancy
system should be in the main rack. This will cause the Genius Bus Controller to lose power at
the same time that the CPU loses power and allow the backup unit to gain full control of the
I/O as soon as possible. Each GBC has an output timer that it resets during every output
scan. If the GBC determines that the CPU in its PLC has failed, it will stop sending outputs to
its Genius devices. This allows the other GBC to take control of the I/O.
For single and dual bus Genius networks, the Genius bus controllers should be placed at the
same end of the bus, as shown on page 3-12. In particular, the secondary unit should be
placed at one end of the bus and the primary unit must be placed between the secondary unit
and the Genius devices. No I/O blocks or other devices should be located on the bus
between the bus controllers.
In the case of dual bus networks, placing the bus controllers and devices in this manner
minimizes the risk of a bus break between the two units. A bus break between the units could
result in only some devices switching busses, and make the other devices inaccessible to
GFK-2308C
Chapter 3 System Configuration
3-13
3
one of the units. It also allows the primary unit to continue to control the I/O in bus failure
conditions that might otherwise result in loss of inputs and unsynchronized control of outputs.
Since the recommended configuration for single and dual bus networks still has the possibility
of a bus breaking between the two CPUs, you may want to program the application to
monitor the status of the busses from the unit configured at the end of the busses and
request a role switch or bus switch (dual bus network only) if loss of bus is detected.
Duplex Genius Output Mode
Although it is not common, you can configure your Genius I/O system for duplex mode,
meaning that they will receive outputs from both bus controllers 30 and 31 and compare
them. Only devices that have discrete outputs can be configured for Duplex mode.
If the controllers at SBAs 30 and 31 agree on an output state, the output goes to that state. If
the controllers at SBAs 30 and 31 send different states for an output, the device defaults that
output to its pre-selected Duplex Default State. For example:
Commanded State
from Device
Number 31
Commanded State
from Device
Number 30
Duplex Default
State in the Block
or I/O Scanner
Actual Output
State
On
On
Don’t Care
On
Off
On
Off
Off
Off
Off
Don’t Care
Off
On
Off
On
On
If either controller 30 or 31 stops sending outputs to the device, outputs will be directly
controlled by the remaining controller.
3-14
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
Chapter Configuration Requirements
4
This chapter defines the special configuration requirements of a Hot Standby CPU
Redundancy system.
When the program logic will be the same for both units, it is recommended that you use a
Dual HWC Target. When you select a Redundancy CPU, the programming software
automatically presents the Dual HWC Target. The remainder of this chapter assumes a Dual
HWC Target.
If you do not want to use the same logic in both units, you should create two separate targets
and set the target property, Dual HWC to FALSE in each target.
CAUTION
If both units are configured as primary or as secondary, they will not
recognize each other. If this happens, the Genius Bus Controllers
report SBA conflict faults and flash their LEDs. Correct the
configuration of both units before placing either unit in Run mode.
Note:
GFK-2308C
The Redundancy CPU can be used for redundant and non-redundant applications.
For non-redundant applications, set the Dual HWC for the Target to False and do not
configure any redundancy links.
4-1
4
Using the Redundancy Wizards
Machine Edition software provides redundancy wizards to create a hardware configuration
with the correct parameter settings for the redundancy scheme that you choose. See
“Configuration Parameters” for details on parameters specific to redundancy systems. To
launch the wizard, go to the Navigation window, right click Hardware Configuration, point to
Redundancy, and then choose Wizard.
To configure a Hot Standby CPU Redundancy system using the wizards:
1. Run the Set up Primary Hardware Configuration for CPU Redundancy wizard. This
wizard configures a redundancy CPU in slot 1 of the main rack and allows you to select
the location of the RMX modules used for redundancy links.
2. RX7i only — Run the Add GBCs for Genius Redundancy wizard to configure Genius bus
controllers in the primary unit.
3. Complete configuration of all parameters for the primary unit.
4. When you have finished configuring the primary unit, run the Generate Secondary
Hardware Configuration from the Current Configuration wizard. This wizard copies the
primary hardware configuration to the secondary configuration and adjusts appropriate
parameters for the secondary configuration.
5. Edit the configuration parameters for each item in the secondary unit’s hardware
configuration that is unique for the secondary unit (for example, the secondary unit’s
direct IP address and the CPU’s SNP ID).
4-2
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
4
Synchronizing the Hardware Configurations
To synchronize the two configurations (after making changes to the primary configuration or
uploading a different primary configuration), right click Hardware Configuration, choose
Redundancy, and Mirror to Secondary Hardware Configuration. This command copies the
primary hardware configuration to the secondary configuration and adjusts appropriate
parameters for the secondary configuration.
Note:
You can control whether the contents of specific slots in the primary configuration are
copied to the secondary configuration. If the Mirror to Secondary property for a slot is
set to True (default), the configured module in that slot in the primary configuration
overwrites the corresponding slot in the secondary configuration. I/O variables
associated with a module in the primary configuration are copied to the
corresponding module in the secondary configuration.
To prevent a slot from being mirrored, set this property to False.
GFK-2308C
Chapter 4 Configuration Requirements
4-3
4
Hardware Configuration Parameters
CPU Parameters
This section discusses only the parameters that apply to redundancy systems. For
information on all the CPU parameters, see the PACSystems CPU Reference Manual,
GFK-2222.
Settings
Parameter
Default
Choices
Stop-Mode I/O
Scanning
Disabled
N/A
Watchdog
Timer (ms)
200
10 through 1000, in
increments of 10ms
Requires a value
that is greater than
the program sweep
time.
Description
Always Disabled for a Redundancy CPU.
The watchdog timer, which is designed to detect "failure to complete
sweep" conditions, is useful in detecting abnormal operation of the
application program, which could prevent the PLC sweep from
completing within a specified time period. The CPU restarts the
watchdog timer at the beginning of each sweep. The watchdog timer
accumulates time during the sweep.
Note: In a CPU redundancy system, the watchdog timer should be
set to allow for the maximum expected scan time plus two
fail wait times. (The Fail Wait parameter is set on the
Redundancy tab.) Furthermore, the watchdog timer setting
must allow enough time for the CPU to complete one input
data transfer and two output data transfers.
Scan Parameters
Communications Window Considerations
The redundancy CPU supports the use of high-speed communications modules such as the
Ethernet Interface. Requests from devices attached to these communications modules are
handled in the Controller and Backplane Communications windows. Because these requests
can be sent in large volumes, there is the potential for either of these windows to be
processing requests for a significant amount of time.
One way to reduce the risk of one CPU failing to rendezvous at a synchronization point with
the other CPU is to configure the Controller and Backplane Communications windows for
Limited Window mode. This sets a maximum time for these windows to run.
Other options are to configure the CPU sweep mode as Constant Window or Constant
Sweep. The CPU will then cycle through the communications and background windows for
approximately the same amount of time in both units.
4-4
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
4
Parameter
Default
Choices
Description
Sweep Mode
Normal
Normal
Constant Window
Constant Sweep.
For details on sweep modes, refer to the
PACSystems CPU Reference Manual,
GFK-2222.
Controller
Communications
Window Mode
Limited
Limited: Time sliced. The
maximum execution time for
the Controller Communications
Window per scan is specified in
the Controller Communications
Window Timer parameter.
Complete: The window runs to
completion. There is no time
limit.
(Available only when Sweep Mode is set to
Normal.) Execution settings for the
Controller Communications Window.
Controller
Communications
Window Timer
Controller
Communications
Window Mode is:
Limited: 10
Complete: There
is no time limit.
Controller Communications
Window Mode is:
Limited: 0 through 255 ms.
Complete: Read only. There is
no time limit.
The maximum execution time for the
Controller Communications Window per
scan.
Backplane
Communications
Window Mode
Limited
Limited: Time sliced. The
maximum execution time for
the Backplane
Communications Window per
scan is specified in the
Backplane Communications
Window Timer parameter.
Complete: The window runs to
completion. There is no time
limit.
(Available only when Sweep Mode is set to
Normal.) Execution settings for the
Backplane Communications Window.
Backplane
Communications
Window Timer (ms)
10ms for Limited
mode
Limited: Valid range: 0 through
255 ms.
Complete: Read only. There is
no time limit.
(Available only when Sweep Mode is set to
Normal.) The maximum execution time for
the Backplane Communications Window
per scan. This value can be greater than
the value for the watchdog timer.
It is highly recommended that this
parameter be set to the same value for
both CPUs in a redundancy pair.
Background
Window Timer
5ms
0 through 255ms
Sweep Timer (ms)
100ms
5 through 2550ms, in
increments of 5. If the value
typed is not a multiple of 5ms, it
is rounded to the next highest
valid value.
GFK-2308C
Chapter 4 Configuration Requirements
Setting the background window time to
zero disables the background RAM tests.
(Available only when Sweep Mode is set to
Constant Sweep.) The maximum overall
PLC scan time. This value cannot be
greater than the value for the watchdog
timer.
Some or all of the windows at the end of
the sweep might not be executed. The
windows terminate when the overall PLC
sweep time has reached the value
specified for the Sweep Timer parameter.
4-5
4
Parameter
Default
Choices
Description
Window Timer (ms)
10
3 through 255, in increments
of 1.
(Available only when Sweep Mode is set to
Constant Window.) The maximum
combined execution time per scan for the
Controller Communications Window,
Backplane Communications Window, and
Background Communications Window.
This value cannot be greater than the value
for the watchdog timer.
Number of Last
Scans
0
0-5 (Should be set to 0.)
The number of scans to execute after the
PACSystems CPU receives an indication
that a transition from Run to Stop mode
should occur.
Note: In a redundancy system, this
parameter should be set to 0
(default). Using a non-zero value
would allow a unit to stay in RUN
mode for a few sweeps after
detecting a fatal fault.
Fault Parameters
Parameter
Recoverable Local
Memory Error
4-6
Default
Diagnostic
Choices
Diagnostic
Fatal
Description
Redundancy CPUs only. Determines
whether a single-bit ECC error causes the
CPU to stop or allows it to continue
running.
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
4
Redundancy Parameters
Parameter
Default
Choices
Description
Redundancy
Mode
Primary
Primary
Secondary
(Read-only when the Dual
HWC target property is set to
True.)
Specifies whether the current Hardware
Configuration is Primary or Secondary.
Note: When the Dual HWC target property is
set to True, one Hardware Configuration is
automatically set to Primary, and the other to
Secondary.
Control
Strategy
HSB
HSB
Selects the HSB control strategy.
Fail Wait
Time
60
60 through 400 ms, in
increments of 10 ms.
The maximum amount of time this CPU waits
for the other CPU to reach a synchronization
point.
For recommendations on setting Fail Wait time,
see chapter 5.
Redundancy
Links
Determined by
number of
redundancy links
configured for this
unit.
Read-only
0: The CPU behaves as a
redundancy CPU without a
backup.
1: The CPU behaves as a
redundancy CPU with one
redundancy link.
2: The CPU behaves as a
redundancy CPU with two
redundancy links. – Strongly
Recommended
The number of redundancy links configured for
this unit. Each redundancy link is a pair of RMX
modules (one in each unit) that have the
Redundancy Link parameter set to Enabled.
--- Redundancy Link 1 --Rack Number
0
(Read only) 0
The rack location of the first RMX module.
(Shown only if the Redundancy Links
parameter is 1 or 2.)
Slot Number
Determined by slot
location of RMX
module.
(Read-only)
The slot location of the first RMX module.
(Shown only if the Redundancy Links
parameter is 1 or 2.)
--- Redundancy Link 2 --Rack Number
0
(Read-only) 0
The rack location of the second redundancy
link. (Shown only if the Redundancy Links
parameter is 2.)
Slot Number
Determined by slot
location of RMX
module.
(Read-only)
The slot location of the second redundancy
link. (Shown only if the Redundancy Links
parameter is 2.)
GFK-2308C
Chapter 4 Configuration Requirements
4-7
4
Transfer List
This is where you select the ranges of references that will be transferred from the active unit
to the backup unit. If the program logic requires identical input values for the two units, those
references must be included in the input transfer list.
A maximum of 2Mbytes of data can be included in the transfer list. The amount of data
transferred is also limited by the amount of user memory consumption. Overrides and
Legacy-style Transitions are transferred for any specified discrete transfer data, as well as
point fault information for transferred discrete and analog data if Point Faults are enabled.
Transferred data, along with user program, configuration, and reference memory size, etc. all
count against the user memory size and contributes to the CPU scan time.
Because the redundancy transfer list is part of hardware configuration, the transfer lists in
both units must be the same for synchronization to occur.
Note:
Individual variables can also be configured as transferred variables in either or both
the input and output transfer lists. For details, see page 4-14.
To view the amount of memory used for transfer data (redundancy memory usage), go online
and store the configuration. Then right click the Target, choose Online Commands, and
select Show Status. In the status dialog box, select the Redundancy tab.
Genius HSB (Available only on RX7i Systems)
If the program logic requires identical input values for the two units, those references
(including Genius inputs) must be included in the input transfer list.
You must include all redundant Genius outputs, i.e. those %Q and %AQ references tied to
redundant Genius devices, in the output transfer list. Failure to do so will result in the primary
unit always determining the output values, even when it is the backup unit. By default,
Machine Edition generates an error and prevents storing of the configuration if an output is
not included in the transfer list. For special situations, you can adjust the Target property,
Genius Output, to generate a warning instead.
4-8
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
4
Parameter
Default
Choices
Description
Input/Output Transfer Point
%I
Reference
%I00001
This address must be byte-aligned, that is,
it must have a value of 8n + 1. Example:
%I00025, where 25= (8 * 3) + 1.
The starting address for the range of %I
references that are synchronized between
the redundant CPUs.
%I Length
0
0 through (32,768 - Iref + 1), in increments
of 8, where Iref = the value set in the %I
Reference parameter.
The number of %I references that are
synchronized between the redundant CPUs.
%Q
Reference
%Q00001
This address must be byte-aligned, that is,
it must have a value of 8n + 1. Example:
%Q00049, where 49 = (8 * 6) + 1.
The starting address for the range of %Q
references that are synchronized between
the redundant CPUs.
%Q Length
0
0 through (32,768 - Qref + 1), in
increments of 8, where Qref = the value
set in the %Q Reference parameter.
The number of %Q references that are
synchronized between the redundant CPUs.
%M
Reference
%M00001
This address must be byte-aligned, that is,
it must have a value of 8n + 1. Example:
%M00121, where 121 = (8 * 15) + 1.
The starting address for the range of %M
references that are synchronized between
the redundant CPUs.
%M Length
0
0 through (32,768 - Mref + 1), in
increments of 8, where Mref = the value
set in the %M Reference parameter.
The number of %M references that are
synchronized between the redundant CPUs.
%G
Reference
%G00001
This address must be byte-aligned, that is,
it must have a value of 8n + 1. Example:
%G00081, where 81 = (8 * 10) + 1.
The starting address for the range of %G
references that are synchronized between
the redundant CPUs.
%G Length
0
0 through (7,680 - Gref + 1), in increments
of 8, where Gref = the value set in the %G
Reference parameter.
The number of %G references that are
synchronized between the redundant CPUs.
%AI
Reference
%AI00001
The limit configured for %AI references is
based on values provided in the Memory
tab. The value of the beginning references
plus the value of the length must be less
than, or equal to, the configured limit.
The starting address for the range of %AI
references that are synchronized between
the redundant CPUs.
%AI Length
0
0 through (AIul - AIref + 1), where AIul =
the upper limit of %AI memory configured
on the Memory tab, and AIref = the value
set in the %AI Reference parameter.
The number of %AI references that are
synchronized between the redundant CPUs.
%AQ
Reference
%AQ00001
The limit configured for %AQ references is
based on values provided in the Memory
tab. The value of the beginning reference
address plus the value of the length must
be less than, or equal to, the configured
limit.
The starting address for the range of %AQ
references that are synchronized between
the redundant CPUs.
%AQ
Length
0
0 through (AQul - AQref + 1), where AQul
= the upper limit of %AQ memory
configured on the Memory tab, and AQref
= the value set in the %AQ Reference
parameter.
The number of %AQ reference addresses
that are synchronized between the
redundant CPUs. The limit configured for
%AQ references is based on values
provided in the Memory tab. The value of the
beginning reference plus the value of the
length must be less than, or equal to, the
configured limit.
GFK-2308C
Chapter 4 Configuration Requirements
4-9
4
Parameter
Default
Choices
Description
%R
Reference
%R00001
The limit configured for %R references is
based on values provided in the Memory
tab. The value of the beginning references
plus the value of the length must be less
than, or equal to, the configured limit.
The starting address for the range of %R
references that are synchronized between
the redundant CPUs.
%R Length
0
0 through (Rul - Rref + 1), where Rul = the
upper limit of %R memory configured on
the Memory tab, and Rref = the value set
in the %R Reference parameter.
The number of %R reference addresses that
are synchronized between the redundant
CPUs. The limit configured for %R
references is based on values provided in
the Memory tab. The value of the beginning
address plus the value of the length must be
less than, or equal to, the configured limit.
%W
Reference
%W00001
The limit configured for %W references is
based on values provided in the Memory
tab. The value of the beginning reference
address plus the value of the length must
be less than, or equal to, the configured
limit.
The starting address for the range of %W
references that are synchronized between
the redundant CPUs.
%W Length
0
0 through (Wul - Wref + 1), where
Wul = the upper limit of %W memory
configured on the Memory tab, and Wref =
the value set in the %W Reference
parameter.
The number of %W references that are
synchronized between the redundant CPUs.
The limit configured for %W references is
based on values provided in the Memory
tab. The value of the beginning reference
address plus the value of the length must be
less than, or equal to, the configured limit.
Redundancy Memory Xchange Module Parameters
Parameter
Redundancy
Link
4-10
Default
Enabled
Choices
Enabled
Disabled
Description
If the RMX module is being used as a
redundancy link, this parameter must be set
to Enabled. An RMX module being used as
a redundancy link cannot be used as a
general-purpose reflective memory module.
All the reflective memory parameters are
unavailable, and the Interrupt parameter is
set to Disabled.
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
4
Ethernet Interface Parameters
Each unit contains at least one Ethernet interface that is assigned a direct IP address used to
directly access the specific unit. A third, redundant, IP address can be assigned to the pair of
Ethernet interfaces in both the primary and secondary units. The redundant IP address is
active on the Ethernet interface in only one of the units at a time, the active unit. All data sent
to the redundant IP address (including EGD produced to the redundant IP address) is
handled by the active unit. When active, the Ethernet interface always initiates
communications using the redundant IP address. When the unit is not active, all
communications are initiated through the direct IP address. For more information about the
Redundant IP address, refer to “Redundant IP Addresses” in chapter 5.
You can have up to four Ethernet interfaces in each rack, including the embedded Ethernet
interface in an RX7i CPU. Each Ethernet interface can be set up as part of a pair for the
purposes of redundant IP. (You can also include Ethernet interfaces in the unit that are not
part of a redundant IP pair.)
When an Ethernet Interface is configured to produce Ethernet Global Data (EGD), you must
configure a redundant IP address in addition to the direct IP address. For more information
about using EGD in a redundancy system, see chapter 5.
Parameter
Default
Choices
Description
IP Address
0.0.0.0
x.x.x.x where x
ranges from 1 to 255
This IP address, also known as the direct IP address, always applies
only to this unit. The IP Address should be assigned by the person
responsible for your network. TCP/IP network administrators are familiar
with these sorts of parameters and can assign values that work with your
existing network. If the IP address is improperly set, your device may not
be able to communicate on the network and could disrupt network
communications.
Redundant
IP
Disable
Disable
Enable
Enabling this feature allows the Ethernet Interface to share an IP
address with the corresponding Ethernet Interface in the other unit.
When this parameter is enabled, a Redundant IP Address must be
entered.
Redundant
IP Address
0.0.0.0
x.x.x.x where x
ranges from 1 to 255
(Available only when the Redundant IP parameter is set to Enable.) The
IP address shared by two Ethernet Interfaces that are connected to the
same network and reside in separate units (one in the primary unit and
the other in the secondary unit). Although the redundant IP address is
shared by both Ethernet Interfaces, only the Interface in the active unit
responds to this IP address. This IP address is assigned in addition to
the device’s primary IP address.
For a pair of Ethernet Interfaces, the redundant IP address must be the
same value on the primary and secondary units.
Note: The redundant IP address must not be the same as the direct
IP address of either Ethernet Interface. The redundant IP
address must be on the same sub-network as the direct IP
address and Gateway IP address, if used.
For more information about Ethernet redundancy, see TCP/IP Ethernet
Communications for PACSystems, GFK-2224.
GFK-2308C
Chapter 4 Configuration Requirements
4-11
4
Rack Module Configuration Parameters
I/O Interrupts
Interrupts cannot be ENABLED when the configured CPU is a Redundancy CPU. When a
redundant CPU is configured, any interrupts enabled in the configuration are DISABLED.
I/O Variables
An I/O variable is a symbolic variable that is mapped to a terminal in the hardware
configuration for individual modules. A terminal can be one of the following: a physical
discrete or analog I/O point on a PACSystems module or on a Genius device, a discrete or
analog status returned from a PACSystems module, or Global Data. The use of I/O variables
allows you to configure hardware modules without having to specify the reference addresses
to use when scanning their inputs and outputs. Instead, you can directly associate variable
names with a module’s inputs and outputs.
I/O variables can be used any place that other symbolic variables are supported, such as in
logic as parameters to built-in function blocks, user defined function blocks, parameterized
function blocks, C blocks, bit-in-word references, and transitional contacts and coils. For
additional information on the use of I/O variables, see the PACSystems CPU Reference
Manual, GFK-2222.
Mapping Hardware I/O Variables Example
Using I/O Variables in a Redundancy System
In a redundancy system, the mapping of I/O variables must be the same in both units. It is
possible to have different modules configured in each unit, as long as the modules that differ
do not have I/O variables assigned to them.
When an I/O variable is added, moved or deleted in one hardware configuration, Machine
Edition performs the same action on the other hardware configuration. If you move a module
with I/O variables to a different rack location, the variables in the corresponding module in the
other hardware configuration are disassociated, causing an I/O Variable Mismatch error. If an
I/O variable is assigned to a module in one unit without a corresponding I/O variable on a
module of the same type in the other unit, an I/O Variable Mismatch error will be generated
upon validation.
I/O variables can be configured as transferred variables in either or both the input and output
transfer lists. For details, see “Adding Individual Variables to the Transfer Lists” on page 4-14.
4-12
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
4
Genius Bus Configuration (RX7i only)
Bus Controller Configuration Parameters
■
When configuring the PRIMARY PLC, all Genius Bus Controllers configured for external
redundancy must have Serial Bus Address 31.
■
When configuring the SECONDARY PLC, all Genius Bus Controllers configured for
external redundancy must have Serial Bus Address 30.
Note:
It is possible to configure Genius networks in which there is not a redundant bus
controller in the other unit. It is not necessary for the serial bus addresses to be 31 in
the primary unit and 30 in the secondary unit for such networks.
■
For single Genius bus networks, the Genius Bus Controllers must be configured for RED
CTRL Redundancy with the redundant pair set to EXTERNAL.
■
For Dual Bus Genius networks, the Genius Bus Controllers must be configured for Dual
Bus/Redundant Controller.
Note:
Genius Bus Controllers for networks that are connected to just one unit may have
any setting.
Genius Device Configuration Parameters
All Genius devices that are connected to both units must be configured as redundant.
Note:
GFK-2308C
Devices that are connected to just one unit may use any available setting.
Chapter 4 Configuration Requirements
4-13
4
Adding Individual Variables to the Transfer Lists
Individual variables can be configured as transferred variables in the input transfer list and/or
the output transfer list. Mapped, managed (symbolic and I/O), and function block instance
variables can be transferred. This is the only way that managed and function block instance
variables can be transferred.
The following types of variables cannot be transferred:
■
Mapped BOOL variables with bit-in-word addresses
■
Elements of BOOL arrays that are mapped to word memories (%R, %W, %AI, %AQ)
■
Aliases to variables
The Input Transfer List and Output Transfer
List properties for a variable are set to False
by default. To add or remove a variable to or
from the variable transfer list, edit the
Properties for that variable.
In most cases, a variable should be part of
the input or output transfer, but not both. In
some unusual cases, where there is a need
to update a variable at both transfer points in
the sweep, the variable may be configured
for both lists.
Mapped Variables
An advantage of configuring mapped
variables this way instead of including them
in the CPU’s Transfer List is that the transfer
properties are tied to the variable, not the
memory location. If you need to relocate a
variable, you do not risk accidentally moving
it out of the transfer area.
Mapped variables must be assigned to one of the memory ranges allowed for redundancy
transfer: %I, %AI, %Q, %AQ, %R, %M, %W, or %G.
Note:
If a mapped variable within a range specified in the CPU hardware configuration
Transfer List (page 4-8) is also configured as a transferred variable, it will be
transferred twice.
Arrays
Arrays can be configured as Mixed transferred variables, allowing individual elements to be
included in the input transfer list and/or the output transfer list. If the top level of the array
variable is set to True or False for either list, all elements in the array are set to the top-level
value for that list.
Instance Data Structure Variables
All elements of instance data structure variables, such as those associated with a function
block, are transferred according to the setting of the head of the data structure.
4-14
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
4
Using the Variable Transfer List Report
The report provides the total number of variable bytes, the total whole bytes, and the total
partial bytes included in the input and the output transfer lists.
To access this report, right click the Target and select Report. In the Available Reports list,
select Variable Transfer List Report and click OK.
GFK-2308C
Chapter 4 Configuration Requirements
4-15
4
Storing (Downloading) Hardware Configuration
A PACSystems control system is configured by creating a configuration file in the
programming software, then transferring (downloading) the file from the programmer to the
CPU via the Ethernet Interface or serial port. The CPU stores the configuration file in its
non-volatile RAM memory.
In the programming software all online operations, including downloading a folder, are
performed on the PLC that is the selected hardware configuration. You must download the
hardware configuration to each PLC in the redundancy system in a separate operation.
CAUTION
If both units are configured as primary or as secondary, they will not
recognize each other. If this happens, the Genius Bus Controllers
report SBA conflict faults and flash their LEDs. Correct the
configuration of both units before placing either unit in Run mode.
1. Make sure the primary
HWC is selected.
To select a hardware
configuration, right click
on Hardware
Configuration and choose
Set as Selected HWC.
2. If not already done, set
the physical port
parameters for the
primary unit in the Target
properties.
3. Connect to the CPU. Make sure the CPU is in Stop mode.
4. Download.
5. Go offline.
6. Select the secondary HWC.
7. If not already done, set the physical port parameters for the secondary unit in the Target
properties.
8. Connect to the CPU. Make sure the CPU is in Stop mode.
9. Download.
4-16
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
4
Run Mode Stores
PACSystems releases 5.5 and later support run mode store (RMS) of the redundancy
transfer list. This capability allows you to add, delete or modify transfer list entries without
stopping the controllers.
If two redundant units are synchronized, the RMS must be performed as a dual operation.
However, when a redundant unit is not synchronized to another unit, the redundancy transfer
list can be stored in a single RMS. This facilitates the commissioning phase of a redundancy
system, where the redundant partner may not be in place yet.
Caution
Do not attempt to synchronize a unit while an RMS is in progress to a
non-synchronized active unit. If the unit attempting to synchronize in
this case is taken to run mode, both units will be non-synchronized
active units.
An RMS of the transfer list requires two copies of the redundancy configuration to be resident
on the PLC for a short time. During that period both copies of the transfer list are charged
against the user memory limit. If there is not enough user space available for both copies
(along with any new logic or EGD data that is part of the RMS), the store will fail.
Dual RMS with Simultaneous Activation In Redundant Systems
Warning
A synchronous RMS of invalid user logic or configuration, such as
would cause a watchdog or processor exception, could cause both
units to fail. To mitigate the risk of such application errors, the
procedure, “Initial RMS Followed by Dual RMS” on page 4-18 is
recommended.
To modify EGD, application logic and/or the redundancy transfer list using RMS and have the
controllers simultaneously activate the changes, you must perform independent downloads to
both controllers. The two controllers then negotiate when to activate the new items. The initial
store can be done to either the primary or the secondary unit. Note that a dual RMS does not
have to include transfer lists. It may include only EGD and/or logic.
GFK-2308C
Chapter 4 Configuration Requirements
4-17
4
When you command an RMS to one of the units, you will be given the option of selecting
synchronized activation of the redundant controllers.
If you select Do synchronized activation of redundant controllers, the first unit defers
application of the newly stored application data until the following actions have occurred:
1. You disconnect from the first unit, connect to the other unit, and command an RMS to
that unit.
2. The programmer performs the RMS to the second unit.
3. Both units validate that the new application data is compatible in the two units.
Because the PLC sweeps are synchronized, both units will activate the new logic and transfer
lists on the same sweep.
If a power loss occurs on one of the units after activation of the new components begins, but
before it completes, the unit maintaining power will complete the activation and continue as a
non-synchronized active unit. When the other unit is powered back on (assuming a good
battery) it will either have the newly stored application or the original application. If the units
match, they can synchronize without a download. If the unit that lost power does not contain
the new application data, a Primary and secondary units are incompatible fault (fault 9 in
group 138) will be generated.
Initial RMS Followed by Dual RMS
The following procedure is recommended to avoid the risk of both units failing due to logic
errors in a dual RMS:
1. Perform an RMS of the new application data only to the backup controller prior to
modifying the transfer list. (Do synchronized activation of redundant controllers is not
selected.)
2. Perform a role switch to make the modified controller active.
3. Add any variables that require synchronization to the transfer list. (See “Adding Individual
Variables to the Transfer Lists” on page 4-14.
4. Initiate a dual RMS.
5. If necessary, perform a role switch so that the primary unit is the active unit.
The unit whose logic had already been stored in run mode will receive only the new transfer
list. The other unit will receive the new transfer list and new logic/EGD.
4-18
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
4
RMS Operational Errors
Certain operational errors can occur only when performing a dual RMS to two synchronized
controllers and performing simultaneous activation of new application data. The table below
outlines possible modes of failure and the system operation when the failure occurs.
Error Mode
System Operation
User requests a normal store (single RMS, not dual
RMS) when the transfer list has changed.
The programmer will not attempt the run mode store and will
display an error message.
User requests a dual RMS on a controller that is
not synchronized to a redundant partner.
The dual store will not be completed. The programmer will display
the following controller error message:
The requested action could not be completed because the target
is not synchronized with another controller. (0x05, 0x3E)
User requests a dual RMS on a controller whose
redundant partner does not support dual RMS.
The dual store will not be completed. The programmer will display
the following controller error message:
The firmware for the remote redundant controller does not
support the operation. (0x05, 0x3C)
Dual RMS aborted (user commanded, loss of
communications, failed download) to a controller
whose redundant partner does not have a pending
dual RMS.
The controller will abort the RMS and delete any new application
data that had been stored.
Dual RMS aborted (user commanded, loss of
communications, failed download) to a controller
whose redundant partner has a pending dual RMS.
Both controllers will abort the RMS and delete any new application
data that had been stored.
Loss of synchronization in a dual RMS where only
one controller has a pending dual RMS.
The controller will abort the RMS and delete any new application
data that had been stored.
Loss of synchronization in a dual RMS where both
controllers have a pending dual RMS.
Both controllers will abort the RMS and delete any new application
data that had been stored.
The two controllers determine that the newly stored
transfer lists are not compatible.
Both controllers will abort the RMS and delete any new application
data that had been stored.
One or both of the units determine that there is a
problem with one of the components downloaded
during the run mode store.
Both controllers will abort the RMS and delete any new application
data that had been stored.
A loss of synchronization occurs after the
activation of the new components begins, but
before it completes.
Both units complete the activation of newly stored application data
and run as non-synchronized active units.
A fatal error (stop halt) occurs after the activation of
the new components begins, but before it
completes.
Both units complete the activation of newly stored application data.
If only one unit has a fatal error, the other unit will run as a
non-synchronized active unit.
A power loss occurs on one of the units after
activation of the new components begins, but
before it completes.
The unit maintaining power will complete the activation and
continue as a non-synchronized active unit. If the other unit is
powered back on (assuming a good battery) it will either have the
newly stored application or the original. The firmware will attempt
to ensure that this unit has the new application so that it can
synchronize to the other unit without a download, but it will not be
guaranteed.
If the units match, they can synchronize without a download. If the
unit that lost power does not contain the new application data, a
Primary and secondary units are incompatible fault (fault 9 in group
138) will be generated.
User attempts to go to programmer mode on a
PLC that already has a pending dual RMS.
You will be prompted to either abort the dual RMS or stay in
monitor mode.
GFK-2308C
Chapter 4 Configuration Requirements
4-19
4
Error Mode
System Operation
User requests a role switch via logic or the physical
switch on the RMX module.
User commanded role switches do not impact the ability to do a
dual RMS. The role switch may be deferred for one sweep, at
most, if it coincides with the simultaneous activation.
Dual RMS may fail in Normal sweep mode with the
Backplane Communication Window Mode set to
Complete. Synchronization is lost and both units
transition to NSAU operation.
When RMS of a large file is performed with the CPU in this sweep
mode, the CPU tries to complete the RMS in a single scan, causing
the sweep time to exceed the Fail Wait time.
To avoid this failure, set the Backplane Communication Window
Mode to Limited or select the Constant Window or Constant Sweep
mode.
Behavior of EGD in a Dual RMS
Added exchanges will begin consumption/production shortly after the activation of logic that is
part of the RMS. Deleted exchanges will cease consumption/production shortly before the
activation of logic that is part of the RMS. Modified exchanges will be offline for a short time
during the activation of new logic that is part of the RMS. For general information about the
behavior of this feature in a simplex system, refer to “Run Mode Store of EGD” in the TCP/IP
Ethernet Communications for PACSystems, GFK-2224.
Unlike activation of the transfer list and logic, activation of EGD changes is not guaranteed to
be simultaneous between the two units in a dual RMS. Even in cases where hardware
configuration and logic are identical on the two units, it cannot be guaranteed that
production/consumption of deleted or modified exchanges will stop on the same PLC sweep.
Likewise, it cannot be guaranteed that production/consumption of added or modified
exchanges will resume on the same PLC sweep. This is consistent with normal operation of
EGD in a redundancy system.
Hardware Configuration and Logic Coupling
If I/O Variables are used, an RMS must include both logic and hardware configuration.
If I/O Variables are not used, you can choose whether to RMS logic, hardware configuration,
or both. If you choose hardware configuration or both, all portions of hardware configuration
that can be stored in run mode will be included. If there are portions of hardware
configuration that are not equal and cannot be stored in run mode, a warning will be
generated.
4-20
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
Chapter Operation
5
This chapter discusses aspects of PACSystems CPU operation that function differently in a
redundancy system. For general details of CPU operation, refer to the PACSystems CPU
Reference Manual, GFK-2222.
GFK-2308C
■
Powerup of a Redundant CPU
■
Synchronizing Redundant CPUs
■
HSB Control Strategy
■
%S References for CPU Redundancy
■
Scan Synchronization
■
Fail Wait Time
■
Data Transfer
■
Switching Control to the Backup Unit
■
Error Checking and Correction
■
Timer and PID Functions
■
Timed Contacts
■
Multiple I/O Scan Sets
■
Stop to Run Mode Transition
■
RUN Disabled Mode
■
Genius Bus Controller Switching — RX7i systems only
■
Redundant IP Addresses
■
Ethernet Global Data in an HSB Redundancy System
5-1
5
Powerup of a Redundancy CPU
When a redundant CPU is powered up, it performs a complete hardware diagnostic check
and a complete check of the application program and configuration parameters. This causes
the powerup time of a redundancy CPU to be longer than a non-redundancy CPU. If the
primary and secondary units power up together, the primary becomes the active unit and the
secondary unit becomes the backup unit.
When the secondary unit powers up, if it does not detect the primary unit, the secondary unit
waits up to 30 seconds for the primary unit to power up. If the primary unit has not completed
its powerup sequence within 30 seconds, the secondary unit assumes the primary unit is not
present. In this case, if the secondary unit is configured to transition to Run on powerup, it
becomes an active unit without a backup unit.
If the primary unit completes its powerup sequence before the secondary unit, the primary
unit waits a few seconds for the secondary unit to complete its powerup sequence. If the
primary unit is set up to transition to Run on powerup and does not detect the secondary unit
within this time, it becomes an active unit without a backup.
Note:
If the system should be fully redundant upon powerup, the secondary unit must
complete power-up first but no more than 30 seconds before the primary unit. To be
sure that this happens, apply power to the secondary unit first.
If either unit is powered up after the other unit is already in Run mode, communications
between the two units are established. If the unit being powered up goes to Run mode, a
resynchronization occurs.
Synchronization of the Time of Day Clocks
At the point when the two units establish communications, the primary unit’s time of day clock
is copied to the secondary unit.
5-2
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
5
Synchronizing Redundant CPUs
When synchronization is initiated, the CPUs exchange information about their configurations.
If a transitioning CPU detects that the configurations are not in agreement, that CPU will not
transition to RUN mode; if both CPUs are transitioning at the same time, neither CPU
transitions to RUN mode.
The following items must be in agreement in order to synchronize:
1. Both CPUs must be configured for the same redundancy control strategy.
2. Both CPUs must have compatible transfer lists.
3. If %I, %Q, %AI, or %AQ references are included in the transfer list, the Point Fault
References configuration parameter must be identical on both units.
During synchronization, the active unit sends a synchronization request to the backup unit
and waits for a response from the backup unit. If the active unit does not receive a response
from the backup unit within its configured Fail Wait time, it operates as a non-synchronized
active unit (NSAU).
During synchronization, the backup unit waits for a synchronization request from the active
unit. If the backup unit does not receive the request within its configured Fail Wait time, it
transitions to NSAU operation. If the backup unit receives a synchronization request within
the Fail Wait time, it waits to receive the synchronization data. If it receives the data within
60ms, synchronization completes. If it does not receive the data, the backup unit operates as
a NSAU.
Dual Synchronization
Dual Synchronization occurs when both CPUs transition to Run at the same time. The
primary unit becomes the active unit and the secondary unit becomes the backup unit.
Non-retentive data is cleared, and the #FST_SCN reference and #FST_EXE bits are set to 1.
Resynchronization
Resynchronization occurs when one unit is already in Run mode and the other unit is put into
Run mode. The unit already in RUN mode remains the active unit and the transitioning unit
becomes the backup unit. The behavior is the same whether the unit going to RUN is the
primary unit or the secondary unit.
At this point, the active unit sends the output transfer data and the input transfer data to the
backup unit. In addition to the configured redundancy transfer data, the #FST_SCN %S
reference as well as internal timer information and #FST_EXE for each common logic block
are transferred from the active unit to the backup unit. Only the internal timers and
#FST_EXE data for program blocks with the same name are transferred. Therefore, the
#FST_SCN and #FST_EXE bits for common blocks are not set on the first scan of the
transitioning unit.
Operation when a Redundancy Link is Removed
When one of the links in a system with dual redundancy links is lost, for example when the
fiber optic cable is removed from one RMX module, and the CPUs remain synchronized with
one link, the redundancy status LEDs (Local Ready, Local Active, Remote Ready, Remote
Active) on the RMX modules associated with the failed link will continue to be updated.
GFK-2308C
Chapter 5 Operation
5-3
5
%S References for CPU Redundancy
%S33 through %S39 and %SB18 reflect the status of the redundancy units. The table below
describes these %S references, and shows their expected states, assuming the primary unit
is active and the secondary unit is backup.
Expected State
%S Bit
Definition
Name
Description
Primary
Unit
Secondary
Unit
%S33
Primary Unit
#PRI_UNT
Set to 1 if the local unit is configured as the
primary unit: otherwise it is cleared. For any
given local unit, if PRI_UNT is set, SEC_UNT
cannot be set.
ON
OFF
%S34
Secondary Unit
#SEC_UNT
Set to 1 if the local unit is configured as the
secondary unit: otherwise it is cleared. For any
given local unit, if SEC_UNT is set, PRI_UNT
cannot be set.
OFF
ON
%S35
Local Unit Ready
#LOC_RDY
Set to 1 if local unit is in Run mode with outputs
enabled. Other wise set to 0.
ON
ON
%S36
Local Unit Active
#LOC_ACT
Set to 1if local unit is currently the active unit;
otherwise it is cleared. For any given local unit,
if LOC_ACT is set, REM_ACT cannot be set.
ON
OFF
%S37
Remote Unit Ready
#REM_RDY
Set to 1 if remote unit is in Run mode with
outputs enabled. Otherwise set to 0.
ON
ON
%S38
Remote Unit Active
#REM_ACT
Set to 1 if remote unit is currently the active
unit; otherwise it is cleared. For any given local
unit, if REM_ACT is set, LOC_ACT cannot be
set.
OFF
ON
%S39
Logic Equal
#LOGICEQ
Set to 1 if the application logic for both units in
the redundant system is the same. Otherwise
set to 0.
ON
ON
%SB18 Redundancy
Informational
Message Logged
#RDN_MSG Set if a redundancy informational message was logged. It can be cleared
in reference tables, logic, or by clearing the fault tables.
%S references can be read from the application program, but cannot be altered or
overridden. These references are always OFF when no configuration has been stored.
Anytime a configuration is stored, the states of these %S references are updated in both
STOP and RUN modes.
The four redundancy status LEDs on the RMX Module correspond to the %S35, %S36,
%S37, and %S38 references. The programming software summarizes the state of the
redundancy system on the Redundancy tab of the Show Status dialog box, accessed from
Online commands. Additionally, external indicators can be used to monitor the state of any
status reference.
If the two CPUs are in Run mode but lose synchronization (due to Fail Wait time set too short
or failure of both redundancy links), both units log faults and proceed as NSAUs. In this case
both units attempt to control the process independently; both units set their #LOC_ACT
status to 1, and clear the #REM_RDY, #REM_ACT, and #LOGICEQ status flags.
#OVR_PRE %S Reference Not Available
The #OVR_PRE reference, %S00011, which indicates whether one or more overrides is
active, is not supported by the Redundancy CPU and should not be used.
5-4
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
5
Scan Synchronization
The figure below shows the sweep components for the active and the backup CPUs.
Active CPU
Backup CPU
Housekeeping
Housekeeping
Input Scan
Input Scan
Send Input Transfer Data
and
Synchronize
1
Logic Solution
Send Output Transfer
Receive Inputs
and
Synchronize
Logic Solution
2
Receive Outputs
and
Other Data
Output Scan
Output Scan
Windows
and
Run-Time Diagnostics
Windows
and
Run-Time Diagnostics
1
Input data transfer
2
Output data transfer
There are two synchronization points in the sweep. The input transfer point occurs
immediately after the inputs are scanned. At this point in the sweep, the newly read inputs
are sent from the active unit to the backup unit. At the output transfer point, the rest of the
data (outputs, internal references, registers) is sent from the active unit to the backup unit.
These data transfers are automatic; they require no application program logic, but do require
proper configuration.
Data can be transferred on either redundancy link. If one link fails, the transfer switches to the
other link without causing a loss of synchronization.
GFK-2308C
Chapter 5 Operation
5-5
5
Fail Wait Time
The active and backup CPUs synchronize their execution twice each sweep: once before
logic execution and once afterwards. Certain failures of one CPU, such as an infinite loop in
the logic, are detected by the other CPU as a failure to reach the next synchronization point
on time. The maximum time to wait for the other CPU is known as the Fail Wait time. The
duration of this time must be specified during configuration of both the Primary and
Secondary Units and can range from 60 ms to 400 ms (in increments of 10 ms), with the
default being 60 ms.
The configured Fail Wait time for the system must be based on the maximum expected or
allowable difference in the two CPUs reaching a synchronization point. For example, if one
CPU might spend 20ms in the communications phase of the sweep and the other unit might
spend 95ms in communications in the same sweep, the Fail Wait time must be set to at least
80ms (80 > 95 -20) to prevent loss of synchronization. In addition, Fail Wait Time must be
greater than the sum of the Controller Communications Window, Backplane Communications
Window and Background Window timer settings.
Differences in the logic execution time and other phases must also be considered when
selecting a Fail Wait time. Some applications limit the possible difference during the
communications window by using Constant Sweep mode or Constant Window mode, or by
setting the system communications window to Limited and selecting a small window time.
If the Communications Window mode is set to Complete (run to completion), the controllers
could lose synchronization, particularly during RMS using a rack-based Ethernet module.
5-6
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
5
Data Transfer
The data is transferred in blocks. Each block is checked for data integrity. The backup CPU
holds the transferred data in a temporary area until all the data has been received and
verified. Then the backup CPU copies the data into the actual PLC memories. If the full
transfer fails to complete properly, the backup unit becomes an NSAU and discards the data
in the temporary area.
Synchronization and Data Transfer Process
Input Data and Synchronization Data Transfer to the Backup Unit
Immediately after the input scan, the active unit sends the selected input data to the backup
unit. This includes the selected ranges within %I, %Q, %AI, %AQ, %R, %M, %G and %W
memories, as well as transferred variables. For discrete data, the status, override, and
legacy-style transition information is transferred. If point faults are configured, point fault data
is also sent.
Sweep Time Synchronization
During the first transfer, the active unit automatically sends a synchronizing message to the
backup unit. This message contains the Start of Sweep Time. The CPUs stay synchronized
because the active unit waits for the backup CPU to respond to the synchronizing message
before starting its logic execution.
The Start of Sweep Time message transfer repeatedly coordinates the elapsed time clocks
(upon which timers are based) in the redundant CPUs. The system time is continuous as long
as one of the two systems is running. When a switchover occurs, the same time continues to
be kept in the new active unit.
Transition Contacts and Coils
PACSystems supports two types of Transition contacts and coils:
■
Legacy Transition contacts and coils: POSCON, NEGCON, POSCOIL, and NEGCOIL
■
IEC Transition contacts and coils: PTCON, NTCON, PTCOIL, and NTCOIL
The essential difference between the two types of instruction is that each IEC transitional
used in logic has its own associated instance data. The instance data gives the state (ON or
OFF) of the BOOL variable associated with the contact or coil the last time it was executed.
For additional information on Transition contacts and coils, refer to the PACSystems CPU
Reference Manual, GFK-2222.
For any redundant transfer data item placed in a transfer list that is located in a discrete
reference table or in the symbolic discrete reference region, the associated Override and
legacy-style Transition data is transferred as part of that list. However, the instance data
associated with IEC transitionals is not synchronized. For this reason, IEC transitionals
should not be used in redundancy if the application requires that this data be synchronized.
IEC transitionals must be used with symbolic data; no legacy-style transition data exists for
symbolic data.
GFK-2308C
Chapter 5 Operation
5-7
5
Output Data Transfer to the Backup Unit
After the input data transfer, both units operate independently until the end of the program
logic solution. Before the output scan starts, a second automatic data transfer occurs. At this
time, the active unit transfers the output transfer data to the backup unit. This includes the
selected ranges within %I, %Q, %AI, %AQ, %R, %M, %G and %W memories, as well as
transferred variables. For discrete data, the status, override, and legacy transition information
is transferred. If point faults are configured, point fault data is also sent.
After the output data transfer, the active and the backup units independently perform their
output scans and run their communications and background windows. They continue to
operate independently until they synchronize again after the next input scan.
Estimating Data Transfer Time
When a system is synchronized, there are additions to the sweep time (compared to a similar
non-redundant CPU model) for transferring data from one unit to the other. The data transfer
time includes the time for the active unit to read the data from the appropriate reference
memory type as specified in the configured redundancy transfer list, move it from the CPU
memory across the backplane, with appropriate data integrity information, into the RMX onboard memory. The data is then transferred from the RMX module in the active unit to the
backup unit’s RMX module via a high-speed fiber optic link. On the backup unit, the data is
moved from the RMX on-board memory over the backplane into the CPU memory. A data
integrity check is performed, and assuming the integrity checks pass, the transfer data is
written to the appropriate reference memory in the backup unit.
These additions to the sweep time can be estimated using the data and equations given in
this section.
1. Calculate the total number of bytes configured as memory ranges in the CPU
configuration’s Transfer List.
5-8
Reference
Type
Reference Size
If Point Faults are
Disabled:
If Point Faults are Enabled:
%I
Bit
(%I length x 3 ) ÷ 8
(%I length x 4) ÷ 8
%AI
Word
(%AI length x 2)
(%AI length x 3)
%Q
Bit
(%Q length x 3) ÷ 8
(%Q length x 4) ÷ 8
%M
Bit
(%M length x 3) ÷ 8
%G
Bit
(%G length x 3) ÷ 8
%AQ
Word
%R
Word
(%R length x 2)
%W
Word
(%W length x 2)
(%AQ length x 2)
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
(%AQ length x 3)
GFK-2308C
5
2. Use the following formulas to estimate the data transfer time for memory ranges.
RX3i Formulas
Data transfers less
than 56K bytes:
Estimated transfer time for memory
ranges (ms)
=
0.00005705959 x Total Transfer Data Size +
0.212556909
Data transfers greater
than 56K bytes:
Estimated transfer time for memory
ranges (ms)
=
0.00004790867 x Total Transfer Data Size +
0.341614952
RX7i Formulas
Data transfers less
than 28K bytes:
Estimated transfer time for memory
ranges (ms)
=
(0.00018355 x Total Bytes Transferred) + 0.184
Data transfers greater
than 28K bytes:
Estimated transfer time for memory
ranges (ms)
=
(0.00013738 x Total Bytes Transferred) + 1.954
Analysis of the linear curve resulting from the measurement of various data points yielded
a break point around 28K, resulting in the two linear equations stated above. Using the
proper equation for the amount of transfer data will yield a minimum amount of error
when doing the calculation. The actual data transfer time may vary slightly from the
estimated time; most systems will see slightly better performance than the estimated
value. In addition, the estimated data transfer time is based on a redundant system with
two redundancy links in a steady state non-error condition without CPU serial
communications activity, Genius bus faults or other high backplane interrupt activity.
3. Calculate the total number of bytes and number of symbolic variables in the
transfer list.
This information is obtained from the variable transfer list report. For details, see “Using
the Variable Transfer List Report” in chapter 4.
Size of transfer list
=
Number of entries
=
Total Variable Bytes Transferred (in Input List) +
Total Variable Bytes Transferred (in Output List)
Entries Containing Only Whole Bytes (in Input List) +
Entries Containing Partial Bytes (in Input List) +
Entries Containing Only Whole Bytes (in Output List) +
Entries Containing Partial Bytes (in Output List)
4. Use one of the following formulas to estimate the total transfer time for symbolic
variables.
*
GFK-2308C
CRU320
Transfer time for variables*
=
0.00003923 x (size of transfer list) +
0.000177916 x (number of entries) – 0.61871745
CRE020
Transfer time for variables
=
0.000130992 x (size of transfer list) +
0.000376524 x (number of entries) + 2.1
CRE030
Transfer time for variables
=
0.000111019 x (size of transfer list) +
0.000249549 x (number of entries) + 1.9
CRE040
Transfer time for variable
=
0.0000940902 x (size of transfer list) +
0.0000783293 x (number of entries) + 1.4
For a negative result, use a value of 0.
Chapter 5 Operation
5-9
5
5. Add the following quantities:
RX3i Formula
+
+
Synchronization base sweep addition – additional amount of time
required to synchronize the CPUs with 0 Data Transfer (ms)
Total transfer time for memory ranges (step 2)
Total transfer time for transferred symbolic variables (step 4)
=
Total estimated transfer time:
3.238 ms
_____ ms
_____ ms
_____ ms
RX7i Formula
+
+
Synchronization base sweep addition – additional amount of time
required to synchronize the CPUs with 0 Data Transfer (ms)
Total transfer time for memory ranges (step 2)
Total transfer time for transferred symbolic variables (step 4)
=
Total estimated transfer time:
3.234 ms
_____ ms
_____ ms
_____ ms
Tips for Reducing Transfer Time
Transferred BOOL variables and non-byte aligned BOOL arrays will increase transfer time.
For these, you can create an array of BOOLs and transfer the entire array for efficiency. You
can alias individual array elements to make logic more readable.
Data structures that contain non-contiguous members of different data types can be created.
You can also create arrays of these structures. This feature allows you to put individual
members of a data structure or the entire structure on one or both of the transfer lists. Placing
arrays of structures in the transfer list has the potential to significantly increase the number of
entries in the transfer list, which will impact user space charged and transfer time.
5-10
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
5
Programming a Data Transfer from Backup Unit to Active Unit
(SVC_REQs 27 and 28)
The program logic can be used to transfer eight bytes (four registers) of data from the backup
unit to the active unit before the next logic solution.
To initiate this transfer, the backup unit executes SVC_REQ 27 (Write to Reverse Transfer
Area). This command copies eight bytes of data from the reference in the backup unit
specified by the PARM parameter. Note that SVC_REQ 27 only works when its CPU is the
backup unit. When its CPU is the active unit, SVC_REQ 27 has no effect.
The active unit stores the transferred data in a temporary buffer. The program in the active
unit must execute SVC_REQ 28 (Read from Reverse Transfer Area), which copies the eight
bytes of data from the temporary buffer to the reference specified by the PARM parameter.
SVC_REQ 28 only works in the active unit. It has no effect when its CPU is the backup unit.
There is always a one-sweep delay between sending data from the backup unit using
SVC_REQ 27 and reading the data at the active unit using SVC_REQ 28.
This data copied from the buffer is not valid in the following cases:
■
During the first scan after either unit has transitioned to RUN;
■
While the backup unit is in STOP mode;
■
If the backup unit does not issue SVC_REQ 27.
The data should not be used if #REM_RDY is off or if #REM_RDY is transitioning to on.
Reverse Data Transfer Example
The following rungs would be placed in the program logic of both units. In this example, the
backup unit would send %P0001 through %P0004 to the active unit. The active unit would
read the data into %P0005 through %P0008. %P0001 through %P0004 on the active unit
and %P0005 through %P0008 on the backup unit would not change. %T0002 would be set to
indicate that the operation was successful and that the data could be used.
GFK-2308C
Chapter 5 Operation
5-11
5
Disabling Data Transfer Copy in Backup Unit (SVC_REQ 43)
To instruct the backup unit to bypass the copy of the transfer data from the active unit, use
SVC_REQ 43. This operation can be used to determine if the active and backup units are
arriving at the same results.
This function is valid only when issued in the backup CPU. It is ignored if issued when the
units are not synchronized, or if it is issued in the active unit.
SVC_REQ 43 disables the copy of data for one sweep, beginning with the output data
transfer and ending with the input data transfer of the next sweep. The copy can be disabled
for multiple sweeps by invoking SVC_REQ 43 once each sweep for the appropriate number
of sweeps.
The resynchronization data transfer always occurs, even if SVC_REQ 43 is invoked in the
first sweep after synchronization (this data transfer includes all inputs, outputs, and internal
data that must be exchanged) since the resynchronization data transfer occurs before the
start of logic execution.
This service request can be set up to disable the copies for all transfers or just the output
transfers. If just the output copy is disabled, the two units can still use the same set of inputs
on each unit. This makes it possible to test the ability of the two units to derive the same
results from the same inputs.
In all cases, the data is still transferred over the redundancy link every sweep and the
synchronization points are still met. The effect of SVC_REQ 43 is to disable the copy of the
data from the transfer to the actual reference memories on the backup unit.
Warning
When SVC_REQ 43 is in effect, the backup unit still takes control of the system
in event of a failure or role switch. Switches to the backup unit may cause a
momentary interruption of data on the outputs because the two units may not
be generating the exact same results.
While SVC_REQ 43 is in effect, you should consider disabling outputs on the backup unit.
Disabling outputs on the backup unit eliminates the risk of an unsynchronized switch of
control (which can cause a momentary interruption of data in the outputs) if the active unit
fails or loses power while the input/output copies are disabled. If the active unit fails or loses
power while outputs are disabled on the backup unit, the system's outputs will go to their
default settings. A secondary effect of disabling outputs on the backup unit is that the nonsynchronized fault action table is used by the active unit to determine which faults are fatal.
Note:
If the CPU is already in RUN/ENABLED mode, a command to disable its outputs will
not take effect until one sweep after the command is received. Therefore, disable the
outputs at least one sweep before you enable SVC_REQ 43.
SVC_REQ 43 cannot be used to disable output data transfer on the primary unit when
outputs are enabled on the primary unit. If that is attempted, the SVC_REQ 43 is rejected.
The first time SVC_REQ 43 is used, a fault is logged as a warning that the PLCs are not
completely synchronized.
The reverse data transfer, if any, is unaffected by SVC_REQ 43.
5-12
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
5
Enabling logic should be used with SVC_REQ 43. A contact with a non-transferred reference
should be part of this enabling logic. That will allow the service request to be turned on/off
directly without being overwritten by the value from the active unit.
If the service request is invoked multiple times in a single sweep, the last call is the one that
determines the action taken.
Successful execution occurs unless:
The values in the command block are out of range.
The service request is invoked when the two units in a redundant system are not
synchronized.
The service request is issued on the active unit.
The service request is issued on the primary unit while the primary unit’s outputs are
enabled.
If the service request is unsuccessful, it will not pass power flow to the right.
Command Block for SVC_REQ #43
The command block for SVC_REQ #43 has two words:
Address
0
Address +1 1 = Disable input and output copies
2 = Disable output copy only
Example
In the following example, when %T00035 is on, the input and output copies are disabled.
GFK-2308C
Chapter 5 Operation
5-13
5
Validating the Backup Unit (SVC_REQ 43)
SVC_REQ 43 can be used to determine if the backup unit is collecting inputs properly (that
is, validate the input scan). It can also be used to determine whether the backup unit is
calculating outputs and internal variables properly (that is, validate the logic solution).
Validating the Backup Unit’s Input Scan
To determine whether the backup PLC is collecting inputs properly, follow these steps:
1. Activate SVC_REQ 43 on the backup CPU, passing the values 0 and 1 to disable the
input and output data transfer copies.
2. Monitor the backup unit's input references and input variables. The values presented
correspond to the inputs that the backup is currently collecting.
3. Visually compare the backup unit's input references and input variables with those
presented by the active unit. Pay special attention to the references and variables that
are included in the input transfer.
4. When you are satisfied that the backup unit is collecting inputs properly, disable the rung
that calls SVC_REQ 43.
Validating the Backup Unit’s Logic Solution
To determine whether the backup unit is calculating outputs and internal variables properly,
follow these steps:
1. Activate SVC_REQ 43 on the backup CPU, passing the values 0 and 2 to disable the
output data transfer copy.
2. Monitor the backup unit's output references, output variables, and internal variables. The
values presented correspond to the values that the backup is currently calculating.
3. Visually compare the backup unit's output references, output variables, and internal
variables with those presented by the active unit. Pay special attention to the references
and variables that are included in the output transfer.
4. When you are satisfied that the backup unit is calculating outputs and internal variables
properly, disable the rung that calls SVC_REQ 43.
5-14
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
5
Switching Control to the Backup Unit
Control switches from the active unit to the backup unit if:
1. The active unit detects a fatal fault.
2. The active unit is placed in Stop mode.
3. The active unit fails or is powered off.
4. The toggle switch on an RMX module is activated.*
5. A switch is commanded from the application program.*
*These two types of requests are not honored if they occur within 10 seconds of the previous
request.
Switching Times and Impact to Sweep Time
The amount of time needed to switch control from the active unit to the backup unit depends
on the reason for the switch.
There are two ways that the backup unit detects that the active unit has failed or lost power.
A. Failure of all remaining redundancy links. This type of failure has negligible impact on
the controller sweep time.
B. Failure of the active unit to rendezvous at a synchronization point within the Fail Wait
time. An example of this type of failure is the CPU not responding because the user
logic is in an endless loop. If the redundancy links are still operational, the increase to
the sweep time will equal the Fail Wait Time.
For these two cases the switchover occurs immediately.
For all other cases, the switchover occurs just before the next input data transfer. The
maximum delay is 1 sweep. There may be an input and an output scan between detection of
the fatal fault and the switch.
Commanding a Role Switch from the Application Program (SVC_REQ 26)
The application program can use SVC_REQ 26 to command a role switch between the
redundant CPUs (active to backup and backup to active). As long as the units remain
synchronized, the switch occurs just before the input data transfer of the next sweep.
When SVC_REQ 26 receives power flow to its enable input, the PLC is requested to perform
a role switch. Power flow from SVC_REQ 26 indicates that a role switch will be attempted on
the next sweep. Power flow does not indicate that a role switch has occurred or that a role
switch will definitely occur on the next sweep. The role switch request is not valid if it occurs
within 10 seconds of a previous request. The 10-second limitation guarantees that only a
single switch occurs if both units make a request at approximately the same time. SVC_REQ
26 ignores the PARM parameter; however the programming software requires that an entry
be made for PARM. You can enter any appropriate reference here; it will not be used.
GFK-2308C
Chapter 5 Operation
5-15
5
Example
In this example, a pushbutton switch on a control console is wired to input %I0002. In the
program logic, the reference for %I0002 is used as the input to the SVC_REQ 26 function
block. When the button is pressed, logic power flows to SVC_REQ 26, causing a role switch
to be requested. The PRM reference is not used and can have any value.
Implementing Preferred Master Using SVC_REQ 26
The HSB control strategy implements a floating master algorithm. This means that when one
unit is put into Run mode while the other unit is already in Run mode, the transitioning unit
always becomes the backup unit.
If an application requires a preferred master algorithm where the primary unit always
becomes the active unit when placed in Run mode, the logic can use the Role Switch service
request, SVC_REQ 26, as shown in the sample LD rung below. This logic must be included
in the primary unit and may also be included in the secondary unit.
5-16
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
5
STOP to RUN Mode Transition
A resynchronization will occur at all STOP to RUN mode transitions. The time to perform this
resynchronization may be larger than STOP to RUN transitions on non-redundancy CPUs.
The STOP to RUN mode transition has two separate paths.
1. If the unit performing the transition is doing so alone or both units are transitioning to Run
at the same time, a normal STOP to RUN mode transition is performed (clear nonretentive memory and initialize #FST_SCN and #FST_EXE).
2. If the other unit is active when this unit performs a STOP to RUN mode transition, nonretentive references will be cleared followed by a resynchronization with the active unit.
RUN Disabled Mode
RUN/DISABLED mode causes all physical outputs to go to their default state in that PLC.
Inputs are still scanned and logic is solved. A CPU in RUN/DISABLED mode may be the
active unit.
The following guidelines apply to using RUN/DISABLED mode with the HSB control strategy.
1. If a unit is in RUN/DISABLED mode, its #LOC_RDY %S reference and the other unit's
REM_RDY %S reference are not set and the corresponding LEDs on the RMX modules
are OFF. This indicates that the unit (with #LOC_RDY reference off) is not available to
drive outputs.
2. If a unit is in RUN/ENABLED mode and the other unit is in RUN/DISABLED mode, the
unit in RUN/ENABLED mode does not use its synchronized fault action table. Instead, it
uses the user-configurable fault actions since there is no backup available to drive
outputs.
3. If Outputs from Active Unit Only is enabled in an ENIU, placing the active controller in
RUN/DISABLED mode will result in the ENIU outputs being held in the last state.
4. When outputs used for Remote IO are specified in the transfer list, outputs are
transferred from the active unit to the backup unit and if outputs are enabled on either
Redundancy CPU, the Remote IO devices will receive the output values calculated by the
active unit. Note: Genius outputs on RX7i Redundancy CPUs are always included in the
transfer list.
Note:
GFK-2308C
If the backup unit is in RUN/DISABLED mode, the backup unit continues NOT to
drive outputs upon failure of the active unit and therefore is not a complete backup.
Chapter 5 Operation
5-17
5
Error Checking and Correction
Error checking and correction (ECC) allows the CPU firmware to detect errors in memory and
correct some of them on the fly. This added layer of checking differs from parity checking in
that it can correct a single-bit error. If the ECC error is a single-bit corrected error, the CPU
generates a diagnostic fault and sets %SA0006 so that you can know of a possible
impending problem and take corrective action. If the ECC error is a multi-bit error, which
cannot be corrected, the CPU logs a fatal fault and goes to Stop-Halt mode.
The Error Checking and Correction function of the memory controller is enabled on the
redundancy CPU regardless of the Background Window Timer setting. This provides
parity-like checking on the contents of every RAM location: the ECC bits are set on every
non-cached memory write and checked on every non-cached memory read. If you are
comfortable with the level of integrity checking that the ECC function provides, you may
chose to disable the additional background RAM tests entirely by setting the Background
Window Timer value to 0.
5-18
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
5
Timer and PID Functions
Timer and PID function blocks remain in lock step between two synchronized units provided:
A. Enabling logic for each function is identical on both units. This includes power flow,
how often the block is called, and so forth.
B. The block in which the function occurs has the same name in both units. Note that
_MAIN is always common.
C. Reference registers (3 for timers, 40 for PID), enabling references, and reset
references for each timer and PID function block are included in the data transfer
lists.
For example, if the following ladder logic appears in the _MAIN block on both units,
%M100, %R250, %R251, and %R252 must all be included in the output data transfer
list to keep this timer synchronized between the two units:
Timed Contacts
When both systems are synchronized, timed contacts (%S3, %S4, %S5, %S6) have exactly
the same value in both units. For example, whenever T_SEC is on in one unit, it also is on in
the other unit as long as both units are synchronized.
Multiple I/O Scan Sets
The Redundancy CPU supports the configuration of multiple scan sets. However, it is
strongly recommended that the redundant I/O be configured in the default scan set (Scan set
1), which is scanned every sweep. The I/O scan set feature allows the scanning of I/O points
to be more closely scheduled with its use in user logic programs.
If an I/O Scan set is not scanned every sweep, it is not guaranteed to be scanned in the
same sweep in the Primary and Secondary CPUs. For example, if the Primary and
Secondary CPUs each have a scan set that is scanned every other sweep (that is,
PERIOD=2), the Primary CPU might scan its scan set in one sweep and the Secondary CPU
scan its scan set in the next.
Use of non-default scan sets can cause variance in the time the units get to the rendezvous
points. This should be considered when determining the Fail Wait time.
GFK-2308C
Chapter 5 Operation
5-19
5
Genius Bus Controller Switching (RX7i only)
In the HSB control strategy, Genius outputs are always enabled for both units (unless
explicitly disabled) so that bumpless switching is possible regardless of which unit is currently
the active unit. All redundant Genius outputs must be included in the output transfer lists so
that both units are configured the same.
Genius Bus Controllers stop sending outputs to Genius devices when no output data has
been received from the PLC CPU for a period equal to two times the configured watchdog
timeout.
If the CPU in the primary unit becomes inoperative in an uncontrolled fashion (for example,
because of a power failure), the Genius Bus Controllers detect this within twice the watchdog
setting, and stop sending outputs to the Genius devices. After three Genius I/O bus scans of
not receiving data from the Genius Bus Controllers at Serial Bus Address 31, the Genius
devices start driving data from Serial Bus Address 30 (the secondary unit) if available.
For example, if the system has a 200ms watchdog timeout and 5ms Genius bus scan time,
and the primary unit main rack loses power, the Genius Bus Controllers in expansion racks
will wait 400ms and then stop updating outputs on Genius devices. After 15ms, the devices
will begin driving outputs based on data from the secondary unit. Note that any Genius Bus
Controllers in the main rack would stop driving outputs immediately since they would also
lose power. Genius devices on these buses would begin driving data from the secondary unit
within 15ms.
Note:
For fastest switching, all Genius Bus Controllers in the Hot Standby CPU
Redundancy system should be installed in the main rack. This causes the Genius
Bus Controllers to lose power at the same time that the CPU loses power. This, in
turn, allows the secondary unit to gain full control of the I/O as soon as possible.
For single bus Genius networks, if outputs are not available on Serial Bus Address 30 or 31,
the devices’ outputs revert to default or hold last state (as configured).
For dual bus networks, if outputs are not available on Serial Bus Address 30 or 31, then the
BSM will switch to the other bus. If outputs are not available on either bus, then the block’s
outputs revert to default or hold last state (as configured).
5-20
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
5
Redundant IP Addresses
Each unit contains at least one Ethernet interface that is assigned a direct IP address, which
is used to directly access the specific PLC. A third, redundant, IP address can be assigned to
the pair of Ethernet interfaces in the primary and secondary PLC units. All data sent to the
redundant IP address (including EGD produced to the redundant IP address) is handled by
the active PLC. When active, the Ethernet interface always initiates communications using
the redundant IP address. When the PLC is in the backup state, all communications are
initiated through the direct IP address.
Each Ethernet interface in the system can be set up as part of a pair that shares a redundant
IP address. Each unit can also include Ethernet interfaces that are not part of a redundant
IP pair.
Immediately after configuration, neither Ethernet interface responds to the redundant IP
address. When notified by the CPU that the unit has become active, the Ethernet interface
determines whether the redundant IP address is in use on the network. If the address is not
in use on the network, the Ethernet interface activates the redundant IP address and sends
out an address resolution protocol (ARP) message to force all other Ethernet devices on the
network to update their ARP cache. This ARP message is sent so that communications to the
redundant IP address will be directed to the newly active unit. At this point the Ethernet
interface responds to both the redundant IP address and its direct IP address. When
commanded to begin EGD production by the CPU, the Ethernet interface in the active unit
verifies that it has successfully obtained the redundant IP address. EGD production does not
begin until the Ethernet interface obtains the redundant IP address.
If the redundant IP address is in use by another device on the Ethernet network, the Ethernet
interface periodically attempts to verify that the address is not in use. The Ethernet interface
attempts to verify the redundant IP address until it determines the redundant IP address is no
longer in use on the network or until the Ethernet interface transitions to backup due to either
a notification from the CPU that the unit has become the backup unit or a failure that results
in the Ethernet interface transitioning to backup. This means that if all redundancy links
between the two units fail and the units become non-synchronized active, both units will
attempt to use the redundant IP address, but only one will succeed. If one of the two units
was already active and responding to the redundant IP address, it will continue to do so; the
unit that was backup will not be able to activate the redundant IP address.
CAUTION
When using the redundant IP feature, the application should take steps
to ensure that the CPU that owns the redundant IP address is the same
CPU that maintains control of the outputs. This becomes an issue
when both CPUs are operating as NSAUs (known as split control),
since both units attempt to control the process independently.
Running both CPUs as NSAUs is not recommended and should be
fixed as soon as possible. Refer to “On-Line Repair
Recommendations” in chapter 6.
GFK-2308C
Chapter 5 Operation
5-21
5
The Ethernet interface monitors the status of the CPU. If the Ethernet interface determines
that it can no longer communicate with the CPU, it deactivates the redundant IP address. The
Ethernet interface also deactivates the redundant IP address when notified by CPU that the
active unit has transitioned to backup. When the Ethernet interface deactivates the redundant
IP address, it transitions to the backup state. In the backup state, the Ethernet interface no
longer responds to the redundant IP address, but forwards any packets received by the
interface destined for the redundant IP to the Ethernet interface in the active PLC. If the
backup unit continues to receive packets destined for the redundant IP address, it will send
additional ARP messages on behalf of the active unit and after a number of time periods, it
will log an exception that will be recorded in the PLC CPU fault table as a LAN System
Software Fault.
Additional details on the operation of the Ethernet Interface can be found in TCP/IP Ethernet
Communications for PACSystems, GFK-2224.
5-22
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
5
Ethernet Global Data in an HSB Redundancy System
Note that two redundant units are not guaranteed to consume a given exchange on the same
PLC sweep when using redundant IP. When using Produce In Backup mode, the backup unit
is not guaranteed to produce data on the direct IP at exactly the same time the active unit
produces data on the redundant IP for a given exchange.
Ethernet Global Data Production
By default, only the active unit produces EGD exchanges. This reduces the amount of traffic
on the Ethernet network and simplifies the handling of the exchanges by the consumer. In
particular, the consumer is able to consume exchanges from the redundant system in the
same way it consumes exchanges from non-redundant systems.
Individual exchanges can be configured for “produce in backup mode.” The backup unit
produces these exchanges through the Ethernet module’s direct IP address.
If the PLC is set to Stop-IO Disabled mode, outputs are disabled on the active unit, and
neither unit produces EGD.
In an Ethernet Interface pair with Redundant IP enabled, a newly active Ethernet interface
arbitrates for the redundant IP address and delays EGD production accordingly. If both
redundant units become non-synchronized active units (this can occur if no redundancy links
are functioning), for each redundant pair, the Ethernet Interface that owns the redundant IP
address will produce exchanges through the Redundant IP address.
If Redundant IP is not enabled, the Ethernet Interfaces in both units produce exchanges
through their direct IP addresses.
The Producer ID as well as all production exchanges should be identical for both units. This
allows the consumer to continue consuming exchanges from the redundant system when the
backup unit becomes active.
Configuring Exchanges to be Produced in Backup Mode
In Machine Edition, to configure a production exchange to be produced in backup mode, go
to the Project view, expand the Ethernet Global Data folder, select the exchange and set its
Produce in Backup Mode property to True.
To change the offset from the default value of 1000, select the Ethernet Global Data folder
and set the Secondary Produced Exchange Offset property to the desired value.
GFK-2308C
Chapter 5 Operation
5-23
5
For exchanges that are produced in backup
mode, an offset must be added to the
Exchange ID. This ensures that the
Exchange ID is unique for those exchanges
that are produced simultaneously by the
active and backup controllers.
For an HSB system using dual HWC, one
set of EGD configuration data is used to
create EGD configuration files for both the
primary and secondary controllers. When
Machine Edition creates the EGD exchange
files for download to the secondary
controller, it adds the secondary offset to the
Exchange ID for each exchange configured
to produce in backup.
Note:
For non dual HWC systems, it is the
user’s responsibility to ensure that
the same offset value is specified in
both the primary and secondary
target projects.
Dual HWC Redundancy Target
Ethernet Global Data
Local Producer ID = (a.b.c.d)
Secondary Produced Exchange
Offset = offset
Produced Exchanges
Name = exchgX
Exchange ID = X
Produce in Backup = FALSE
Name = exchgY
Exchange ID = Y
Produce in Backup = TRUE
Download EGD
Configuration
to Primary
Download EGD
Configuration
to Secondary
PLC - Primary
PLC - Secondary
Exchange ID = X
Exchange ID = Y
Exchange ID = X
Exchange ID = Y + offset
Exchange ID Offset in Dual HWC HSB System
Ethernet Global Data Consumption
Both the active and backup units consume EGD exchanges in RUN mode, regardless of
whether or not the units are synchronized.
It is recommended that all consumption exchanges be configured identically for both units. In
addition, these exchanges must be configured as multicast or directed to the Redundant IP
address.
The consumption of multicast exchanges occurs independently on the two units. The
Ethernet modules obtain a copy of multicast exchanges at the same time, but reading of that
exchange in the two CPUs may be phased by one sweep. This can result in the two units
seeing different values for the same exchange in a given sweep. Only the active unit
consumes exchanges directed to the Redundant IP address.
If data from the exchanges must be seen identically on the two units, the reference data for
the exchanges can be transferred from the active unit to the backup unit during the input data
transfer. That transfer occurs shortly after the EGD consumption portion of the CPU sweep.
Exchange variables transferred must be placed into %I or %AI memory to participate in the
input data transfer.
5-24
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
Chapter Fault Detection
6
This chapter describes how faults are handled in a Redundancy system.
■
Fault Detection
■
PLC Fault Table Messages for Redundancy
■
Fault Response
■
Redundancy Link Failures
■
Fault Actions in a CPU Redundancy System
■
Online Repair
Fault Detection
The detection of faults and failures falls into three basic categories:
1. Faults and failures that are detected immediately
2. Faults and failures that are detected as soon as possible, but not necessarily within the
current sweep
3. Faults and failures that are detected in the background
Faults and failures that are detected immediately are those that are identified within the
current sweep. These faults include I/O data corruption, single and multiple bit memory
failures, power supply failures, processor failures, and VME transfer failures.
Faults and failures that are detected as soon as possible, but not necessarily within the
current sweep, include a group of faults that are not detected by the CPU itself. These faults
are typically detected within one second. Genius faults (circuit faults, loss of device, and so
forth) fall into this category.
During the background window, additional memory tests are continuously performed. These
tests can also detect single and multiple bit memory failures.
GFK-2308C
6-1
6
PLC Fault Table Messages for Redundancy
The following table lists messages, descriptions, and corrective actions for error codes
associated with the redundancy fault group. These error codes can be viewed in the Fault
Tables provided by Machine Edition. The entire fault data (including these error codes) can
also be accessed using SVC_REQ 15 and 20.
Redundancy Fault Group (138)
Error
Code
1
Message
Fault Description
Corrective Action
The primary and secondary units have
switched roles, the secondary
transitioned to Run after the primary, or
both units transitioned to Run at the
same time.
Secondary unit is active The secondary and primary units have
and primary unit is
switched roles, or the primary
backup.
transitioned to Run after the secondary.
Primary unit is active; no The primary unit has transitioned to Run
backup unit available.
mode or secondary unit was put into
Stop mode. The primary unit is running
without a backup.
Secondary unit is active; The secondary unit has transitioned to
no backup unit available. RUN mode or primary unit was put into
Stop mode. The secondary unit is
running without a backup.
None required.
5
Primary unit has failed;
secondary unit is active
w/o backup.
The primary unit has recorded a fatal
fault or the secondary has lost
communications with the primary. The
secondary unit is running without a
backup.
6
Secondary unit has
failed; primary unit is
active w/o backup.
The secondary unit has recorded a fatal
fault, or the primary unit has lost
communications with the secondary.
The primary unit is running without a
backup.
If primary unit has also logged the fault
“Secondary Unit Has Failed: Primary
Unit is Active w/o Backup”, then
communications is broken between the
two units and must be repaired. If a
fatal fault has been logged in the
primary unit, the indicated fault must
be repaired. Power may have to be
cycled on one of the units in order to
re-establish communications and
return to a synchronized system.
If secondary unit has also logged the
fault “Primary Unit Has Failed:
Secondary Unit is Active w/o Backup”,
then communications has been broken
between the two units and must be
repaired. If a fatal fault has been
logged in the secondary unit, the
indicated fault must be repaired. Power
may have to be cycled on one of the
units in order to re-establish
communications and return to a
synchronized system.
8
Unable to Switch
Redundancy Roles
An attempt to switch redundancy roles
was made when it was not possible to
perform the switch.
2
3
4
6-2
Primary unit is active
and secondary unit is
backup.
None required.
To have a synchronized system, the
secondary unit must be placed in RUN
mode with a compatible configuration.
To have a synchronized system, the
primary unit must be placed in RUN
mode with a compatible configuration.
None required.
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
6
Error
Code
Message
Fault Description
Corrective Action
9
Primary and secondary
units are incompatible
This unit could not be placed into RUN
mode because the configurations were
not compatible.
Correct the configurations so that the
CPUs have compatible transfer lists
and the same point faults enabled
setting.
10
CPU to CPU
communications
terminated
Synchronization protocol has been
violated.
Contact GE Fanuc technical support.
If the fault is accompanied by a Loss of
Module fault, see corrective action for
“Loss of Module”’ fault.
The link can be restored to service by
power cycling either unit or storing
configuration to either unit.
The CPU has timed out while waiting on Contact GE Fanuc technical support.
communications from the other unit.
The link can be restored to service by
power cycling either unit or storing
configuration to either unit.
Due to actions taken by the user, the
Disable the logic that executes
two units in a CPU redundant system
SVC_REQ 43.
are not fully synchronized. This means
the backup unit is not executing with the
same inputs and/or outputs as the active
unit while the units are synchronized
due to data transfers being disabled.
11
Redundant Link has
timed out
12
Units Are Not Fully
Synchronized
14
Redundant link
communication failure
Communications with the other CPU
over this link has failed.
15
Fail Wait time exceeded
The other CPU failed to rendezvous at a Increase the configured Fail Wait time.
synchronization point within the Fail
Wait time.
17
Could not synchronize
with remote.
The remote unit is unable to synchronize Attempt to synchronize after the
with the local unit because it is
remote unit completes its RMS.
performing an RMS.
GFK-2308C
Chapter 6 Fault Detection
If the other unit failed or lost power,
power cycle it.
Verify one CPU is configured for
primary and the other for secondary.
Check the cable connections between
the two RMX modules.
If the fault is accompanied by a Loss of
Module fault, see corrective action for
“Loss of Module”’ fault.
Otherwise, contact GE Fanuc technical
support.
6-3
6
Other Fault Groups
The following table lists messages, descriptions, and corrective actions for error codes
associated with redundancy in other fault groups.
Error
Code
Group
Loss of Option
Module (4)
various
Message
Loss of or
missing option
module
Fault Description
Install the missing
module or correct the
configuration.
Otherwise, replace
the module and
contact GE Fanuc
technical support.
The bus controller has detected
that another device on the
Genius network is already
using the same serial bus
address.
Verify that one CPU
is configured for
primary and one for
secondary.
or
Redundant link
hard failure
occurred
I/O Bus Fault (9)
none
SBA conflict.
Corrective Action
The module is missing or the
CPU has determined that the
module has failed.
Correct the
configuration of the
Genius devices.
PLC Software (135)
148
Units contain
mismatched
firmware; update
recommended.
The redundant CPUs have
different firmware revision
levels. Having different
revisions of firmware in the
CPUs is intended for short-term
synchronization only as some
change in the behavior of the
system may be experienced
when mixing revisions.
Upgrade the CPUs
so that they have the
same revision of
firmware according
to the firmware
upgrade procedure.
Configuration Mismatch (11)
75
ECC jumper
should be
enabled, but is
disabled
When redundancy firmware is
installed in the CPU module,
the ECC jumper must be in the
enabled position.
Set the ECC jumper
to the enabled
position (jumper on
both pins). See the
instructions provided
with the firmware
upgrade kit.
Recoverable Local Memory
Error (26)
1
Recoverable local A single-bit error was
memory error
encountered and corrected.
%SA00006 is set.
The CPU may need
to be replaced.
Fatal local
memory error
Replace the CPU
and contact GE
Fanuc technical
support.
CPU Hardware (13)
6-4
169
Multiple bit ECC error.
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
Contact GE Fanuc
technical support.
GFK-2308C
6
Fault Response
The Hot Standby CPU Redundancy system detects and reports failures of all critical
components so that appropriate control actions may be taken. All components that acquire or
distribute I/O data or that are involved in execution of the control logic solution are considered
critical components.
A FATAL fault in the active unit causes a switch of control to the backup unit. A DIAGNOSTIC
fault allows the currently active system to continue operating as the active system.
Faults within the unit may be such that:
1. The CPU has a controlled shutdown,
2. The CPU has an uncontrolled shutdown, or
3. The CPU continues to operate.
If the CPU detects an internal fault and has a controlled shutdown, it logs a fault, goes to
Stop/Fault mode, and notifies the other CPU. If the fault was detected on the active unit, the
switchover does not normally occur until the next sweep. The exception is when the active
unit detects a fatal fault during the input scan. In that case, the two units switch roles just
before performing the input data transfer.
If the CPU has an uncontrolled shutdown, the CPU logs a fault if it can and proceeds as
described above. When the backup CPU detects that the active CPU has failed (either by
receiving notification, by detecting that both redundancy links have failed, or by detecting
failure of the active CPU to rendezvous at the next synchronization point within the Fail Wait
time) it becomes an unsynchronized active unit.
If the two CPUs lose synchronization for other reasons (due to Fail Wait time set too short or
failure of both redundancy links), both units log faults and proceed as non-synchronized
active units. In this case both units attempt to control the process independently. The
redundant Genius outputs will prefer the output values sent by the primary unit.
CAUTION
When using the redundant IP feature, the application should take steps
to ensure that the CPU that owns the redundant IP address is the same
CPU that maintains control of the outputs. This becomes an issue
when both CPUs are operating as non-synchronized active units
(NSAU), since both units attempt to control the process independently.
Running both CPUs as NSAUs is not recommended and should be
fixed as soon as possible. Refer to “On-Line Repair
Recommendations” on page 6-10.
GFK-2308C
Chapter 6 Fault Detection
6-5
6
Redundancy Link Failures
There are distinct differences between losing a redundancy link and faulting an RMX module.
Redundancy Memory Xchange Module Hardware Failure
Failures such as VME bus errors are considered hardware failures of the RMX module. The
following actions are taken when such an error is detected:
■
Either a Loss of or Missing Option Module or a Redundant Link Hard Failure Occurred
fault is logged in the PLC Fault Table
■
A Redundant Link Communications Failure fault is logged in both units.
■
All LEDs on the RMX module are turned OFF.
■
The fault locating references that correspond to the module are set (i.e. the SLOT_00XX
fault contact is set, where XX is the slot number for the RMX module).
■
The corresponding redundancy link is no longer used. If the other link is still operating,
that link is used for all further data transfer, and the units can remain in synchronization. If
the other redundancy link is not available and either unit is in Run mode, that unit
operates as a non-synchronized active unit.
Power must by cycled on the rack to restore a faulted RMX module to service.
Redundancy Link Communications Failures
The following errors are reported as failures of the redundancy link:
■
The other unit has lost power or failed such that it can no longer communicate.
■
One or both cables between the two RMX modules have failed or are disconnected.
■
A network error was detected on the fiber optic link that connects the two RMX modules.
(This includes data checks on mismatches, protocol errors, and rogue packets.)
■
Failure of the other CPU to rendezvous at the next synchronization point within the Fail
Wait time.
The following actions are taken when a redundancy link communications failure occurs:
1. Either a Redundant Link Communications Failure or Fail Wait Time Exceeded fault is
logged in the PLC Fault Table of both units.
2. The LINK OK LEDs on both RMX modules are turned off.
3. The fault locating references that correspond to the module are set (i.e. the SLOT_00XX
fault contact is set, where XX is the slot number for the RMX module).
4. The corresponding redundancy link is no longer used. If the other link is still operating,
that link is used for all further data transfer, and the units can remain in synchronization. If
the other redundancy link is not available and either unit is in Run mode, that unit
operates as a non-synchronized active unit.
If the RMX modules’ OK LEDs are still ON, the link can be restored to service by power
cycling either unit or storing a hardware configuration to either unit. If either OK LED is OFF,
power must be cycled on the rack to restore that RMX module to service.
6-6
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
6
Fault Actions in a CPU Redundancy System
Fault actions in the Hot Standby CPU Redundancy System are handled differently than fault
actions in a non-redundant system. Whenever the units are synchronized, the types of faults
that are considered to be FATAL (i.e., cause the CPU to stop) are not configurable. The
following types of faults are considered FATAL when the units are synchronized:
■
Any fault that causes loss of control of I/O
■
Any fault that degrades performance
Note:
In a CPU redundancy system a Fatal fault from a Genius Bus Controller causes a
synchronized unit to transition to STOP/FAULT mode. All Diagnostic faults allow the
CPU to remain in Run mode.
Configuration of Fault Actions
You can configure whether certain faults are considered fatal when the CPUs are not
synchronized.
The following should be considered when configuring the fault actions for a redundancy CPU.
For a given fault that is fatal for the synchronized case, if you set the non-synchronized fault
action to be diagnostic, there is a chance that a less healthy unit could remain the active unit
even after a more healthy backup unit is placed in Run mode. For example, if you were to
configure "Loss of or Missing Rack" failures as diagnostic, the following sequence of events
could occur:
1. If an expansion rack fails when the units are synchronized, the unit with the rack failure
will transition to STOP/FAULT mode and the other unit will become a non-synchronized
active unit.
2. If an expansion rack fails in the non-synchronized active unit, a diagnostic fault will be
logged but the unit will stay in RUN mode and continue to control the process.
3. If the first unit is repaired and then transitions to Run, the second unit with the failed
expansion rack will stay in RUN mode and will remain in control of the process.
To prevent this situation, you may want to include logic to shut down the less healthy unit or
request a role switch.
Also, a unit with the fault actions set to diagnostic may be placed in RUN mode and become
the active unit even though it may have a diagnostic fault, which would be logged as fatal in a
synchronized system.
For example, if an expansion rack fails while in STOP mode or while transitioning to RUN
mode, a diagnostic fault is logged. However, the unit will still transition to RUN. In addition, if
you have programmed a Preferred Master algorithm, this unit will become the active unit. To
prevent this situation, you may want to include logic to shut down the less healthy unit or
modify the role switch logic.
GFK-2308C
Chapter 6 Fault Detection
6-7
6
Configurable Fault Groups
The table below shows the configurable faults and their fault actions. There are three fault
actions: Fatal, Diagnostic, and Conditionally Fatal. Fatal always stops the PLC, Diagnostic
never stops the PLC, and Conditionally Fatal stops the PLC depending on other information
in the fault.
Table
Type
Fault Group
Non-Synchronized
Fault Action
Description
Default
Synchronized Fault Action
(fixed)
Configurable
LOSS_RACK (1)
PLC
Loss of or Missing Rack
Diagnostic Yes
Fatal
LOSS_IOC (2)
I/O
Loss of or Missing I/O
Controller
Diagnostic Yes *
Fatal
LOSS_IO_MOD (3)
I/O
Loss of or Missing I/O
Module
Diagnostic Yes
Diagnostic
LOSS_OTHR_MOD (4)
PLC
Loss of or Missing Option
Module
Diagnostic Yes
Diagnostic
SYS_BUS_ERROR (12)
PLC
System Bus Error
Fatal
IOC_FAULT (9)
I/O
IOC or I/O Bus Fault
Diagnostic Yes
Conditionally Fatal**
CNFG_MIS_MTCH (11)
Both
System Configuration
Mismatch
Fatal
Diagnostic
IOC_SOFTWR (15)
I/O
IOC Software Failure
Diagnostic Uses LOSS_IOC
setting
Conditionally Fatal**
OVER_TMP (24)
PLC
CPU Over Temperature
Diagnostic Yes
Fatal
LOC_MEM_ERROR
PLC
Recoverable Local Memory Diagnostic Yes
Error
(38)
6-8
Yes
Yes
Fatal
Diagnostic
*
Even if the non-synchronized fault action for the LOSS_IOC fault group is configured as Fatal, the
PLC will not go to STOP/FAULT mode unless both Genius Bus Controllers of a dual bus pair fail.
**
Conditionally Fatal: When the units are synchronized, the two fault groups IOC_FAULT and
IOC_SOFTWR faults are fatal if the Genius Bus Controller reports the fault as Fatal. When a GBC
logs one of these faults, it notifies the PLC whether or not it can continue by placing Fatal or
Diagnostic in the fault action of the fault entry.
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
6
Non-Configurable Fault Groups
The table below shows the non-configurable faults and their fault actions. There are two fault
actions: Fatal and Diagnostic. Fatal always stops the PLC and Diagnostic never stops the
PLC.
Fault Group
SYS_BUS_FAIL
Table
Type
PLC
Description
System bus failure.
Fault
Action
Fatal
NO_USER_PRG
PLC
No User's Program on Power-up.
BAD_USER_RAM
PLC
Corrupted User RAM detected on Power-up.
Diagnostic
WIND_CMPL_FAIL
PLC
Window Completion Failure in Constant Sweep Mode (i.e., all
windows failed to receive their allotted time).
PASSWD_FAIL
PLC
Password Access Failure.
Diagnostic
NULL_SYS_CNFG
PLC
NULL System Configuration for RUN Mode.
Diagnostic
CPU_SOFTWR
PLC
PLC CPU Software Failure.
Fatal
SEQ_STORE_FAIL
PLC
Communication failure during a store operation by the programmer.
This fault results when the start-of-store sequence was received but
not an end-of-store sequence.
Fatal
ADD_RCK
PLC
Addition of Extra Rack
Diagnostic
Fatal
Diagnostic
ADD_IOC
I/O
Addition of or Extra IOC
Diagnostic
ADD_IO_MOD
I/O
Addition of or Extra I/O Module
Diagnostic
Addition of, Reset of, or Extra Option Module
Diagnostic
I/O Module Fault
Diagnostic
ADD_OTHR_MOD
PLC
IO_MOD_FAULT
I/O
CPU_HARDWR
PLC
CPU Hardware Failure
MOD_HARDWR
PLC
Module Hardware Failure (for example, Serial Port Failure on PCM)
Diagnostic
MOD_OTHR
SOFTWR
PLC
Option Module Software Failure
Diagnostic
PRG_BLK_CHKSUM
PLC
Program Block Checksum Mismatch
LOW_BATTERY
PLC
Low Battery in the System
Diagnostic
CNST_SW_EXCD
PLC
Constant Sweep Exceeded
Diagnostic
PLC_FTBL_FULL
PLC
PLC System Fault Table Full
Diagnostic
Fatal
Fatal
IO_FTBL_FULL
PLC
I/O Fault Table Full
Diagnostic
APPLICATION_FLT
PLC
User Application Fault
Diagnostic
Fatal Faults on Both Units in the Same Sweep
It is very unlikely that a fatal fault would occur on both units in the same sweep. If that should
happen, however, the first CPU to detect a fatal fault will use the synchronized fault action
table. The other CPU will use the non-synchronized fault action table. This allows one of the
units to stay in Run mode when the synchronized fault action is Fatal and the nonsynchronized fault action is diagnostic.
GFK-2308C
Chapter 6 Fault Detection
6-9
6
Online Repair and System Upgrade
With a Hot Standby CPU Redundancy system, most system component failures can be
repaired by replacing the failed component while the system is online. You may choose to
replace components for other reasons, such as upgrading to a new model of a module. CPUs
in both units must have the same model types and firmware version.
On-Line Repair Recommendations
Note:
If the LOCAL ACTIVE LEDs are ON and the REMOTE ACTIVE LEDs are OFF on
both units, the system is operating under split control, that is, with both units
operating as NSAUs. Do not use this procedure if this condition exists, since neither
unit has the backup role. Additionally, in a system that uses ENIU I/O, there is no
guarantee that all ENIUs are taking outputs from the same controller. See “Repair of
a Split Control System” on page 6-12.
To replace a component online, it is strongly recommended that you follow this procedure:
1. Make sure the unit to be repaired is the backup unit. (The LOCAL ACTIVE LED should
be OFF and the REMOTE ACTIVE LED should be ON. You can also confirm this by
viewing the Redundancy tab of the programmer’s online status dialog box.) If the unit to
be repaired is already in Stop mode, skip this step. If the unit to be repaired is active,
activate the Role Switch on the RMX module.
2. Power-off the unit to be repaired.
3. Replace the defective component.
4. On the CPU of the repaired unit, place the Run/Stop switch in the Stop position.
5. Power on the repaired unit.
6. After several seconds, verify that the LINK OK LEDs are ON for all RMX modules in both
units. If the LINK OK LEDs are not on, see the PLC Fault Table.
7. If the repaired CPU is in Stop/Fault mode, verify that there are no unexpected faults and
then clear the Fault Tables.
8. Place the repaired unit into RUN mode by putting the Run/Stop switch in the Run
position.
6-10
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
6
Hot Swapping of Modules (RX3i Systems Only)
RX3i redundancy systems support hot swapping of modules to the same extent allowed in
non-redundant systems. Modules that support hot swapping can be removed and replaced in
the RX3i main rack and in ENIU remote racks while the rack is powered up.
Hot Swapping RMX 128 Modules
The RX3i RMX128 module supports hot insertion and removal. However, the redundancy
communication link associated with a hot swapped RMX module will not be restored
automatically. The LINK OK indicator on both RMX modules in the link will be OFF.
To restore the link while the system is in operation, first determine which unit is the backup
unit, and if possible, cycle power or store hardware configuration to that unit.
If either RMX module’s OK indicator is OFF, power must be cycled on the rack to restore the
RMX module to service.
System CPU Upgrade
If you are upgrading your redundancy system with new CPU models, you will need to replace
the CPUs in both units. To replace the CPUs in your redundancy system, follow the steps in
“On-Line Repair Recommendations.” When you have replaced the CPU in the backup unit
and returned it to RUN mode, activate the Role Switch on the RMX module and repeat steps
1—8 for the other unit.
Caution
During normal operation, the primary and secondary units in an HSB
redundancy system must have the same CPU model type. Extended
operation with dissimilar CPU types is not allowed. Continued use of
dissimilar CPU types may result in timing issues during
synchronization.
The primary and secondary units with dissimilar CPU model types can be synchronized for a
limited time, for the purpose of system upgrade only. Fail wait times for the higher
performance CPU in a dissimilar redundant pair may need to be increased to allow
synchronization. It does not matter whether the newer model is in the primary or
secondary unit.
GFK-2308C
Chapter 6 Fault Detection
6-11
6
Online Repair of the Genius Bus
Single Bus Networks
The Genius bus of a single bus network can be repaired without disturbing power to either
unit. However, repairing the bus without taking the entire Hot Standby CPU Redundancy
system offline is not recommended because all devices on that bus will be disconnected from
the controllers while the bus is being repaired.
Dual Bus Networks
The Genius bus of a dual bus network can be repaired without disturbing power to either unit.
It is recommended that you disconnect the failed bus from the GBCs before you attempt to
repair it.
Repair of a Non-Synchronized Active Unit (NSAU) Split Control System
When Redundancy CPUs lose all redundancy links and become NSAUs, there is a possibility
of split control or of a failed rack controlling outputs.
In a split control situation, some of the Remote IO devices are taking outputs from one
Redundancy CPU and the other Remote IO devices are taking outputs from the other CPU.
In this situation turning off one of the controllers could result in defaulting the outputs of some
of the Remote IO devices.
A situation where a failed rack controls the outputs occurs when the failed RMX module is
contained in the same rack as the CPU that is currently controlling Remote Device outputs.
The procedures given in this section discusses ways to reduce the chance of defaulting
outputs on some of the Remote IO devices controlled by the Redundancy CPU pair.
Although these procedures might prevent defaulting outputs, they might also involve a short
disruption in the outputs as the Remote IO devices switch to taking outputs from the other
CPU. It is incumbent on the user to know which CPU is controlling outputs on a specific
Remote IO Device and determine whether it is acceptable to allow those outputs to default or
to be disrupted.
Initial Steps for All Systems
Determine the source of the Redundancy link failure, which can either be the fiber optic cable
or a failed RMX module.
1. Check the OK LEDs on the RMX modules. If the RMX’s OK LED is off, the RMX module
has failed.
If there is a failed RMX module, the rack containing the module will have to be taken
offline in order to do the repair.
2. If all RMX OK LEDs are on, check the Signal Detect LEDs on the RMX modules. If the
Signal Detect LED is off, it might indicate that the fiber optic cable connected to the RX
input has failed.
If there is a failed fiber optic cable, you will need to choose which CPU to take offline to
recover the redundancy link(s). Before taking one of the Redundancy CPUs offline, follow
the steps given below for the particular I/O system.
6-12
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
6
Genius I/O Systems (RX7i only)
If the Genius Bus Controllers on both the Primary and Secondary CPUs are OK and actively
sending outputs to the Genius devices, it is preferable to power off the Secondary CPU rack
because the Genius devices prefer the Primary CPU.
If an RMX module has failed the rack containing the failed module must be powered
off, even if it is the Primary CPU rack.
If it has been determined that the problem is due to a failed fiber cable only, you can
choose to take the Secondary CPU offline.
Note:
If there is a problem with Genius Bus Controller connectivity to any of the Genius I/O
Devices, this should be fixed before proceeding to the next steps.
Caution
Since the Redundancy CPUs are not synchronized, taking a CPU offline
may cause a disruption in the outputs. You must be prepared to handle
this condition.
ENIU I/O Systems
1. Using the ENIU status data, you should determine whether all ENIUs have network
connectivity to both Redundancy CPUs. For details on using the ENIU status
information, refer to the PACSystems RX3i Ethernet NIU User’s Manual, GFK-2439.
Note:
If there is a problem with network connectivity to either CPU from any ENIU, this
should be fixed prior to proceeding to the next steps.
2. Using the ENIU status data, determine which CPU is controlling outputs on each ENIU.
If all ENIUs are taking outputs from one CPU (normally it will be the Primary on
LAN A), it is preferable to take the Redundancy CPU that is not currently controlling
outputs offline.
If it has been determined that the problem is due only to a failed fiber cable, you can
choose to take the CPU not controlling outputs offline.
If there are some ENIUs taking outputs from one CPU and some taking outputs from
the other CPU or you need to take the CPU that is currently controlling outputs
offline, for example if it contains the failed RMX module, take the desired CPU offline.
Caution:
Since the Redundancy CPUs are not synchronized, taking a CPU offline
may cause a disruption in the outputs. You must be prepared to handle
this condition.
GFK-2308C
Chapter 6 Fault Detection
6-13
6
Final Steps for All Systems
RX7i Systems: When a module has failed, the CPU will have to be taken offline by powering
off the rack.
RX3i Systems: Because the RX3i system supports Hot Swap of modules, the CPU can be
taken offline by either powering off the rack or by stopping the CPU.
After taking the Redundancy CPU offline, replace the defective RMX module or cable
and bring the CPU back online.
If the CPU was powered off and retained its logic and configuration and is configured
to Run after a power cycle, the Redundancy CPUs will automatically re-establish the
redundancy links and resynchronize.
If the CPU was stopped, use the programmer to download logic and configuration
and put the CPU into Run mode. This will cause the CPUs to re-establish the
redundancy links and resynchronize. After the CPUs are resynchronized, the steps
given in “On Line Repair Recommendations” on page on page 6-10 can be followed
to fix any other failed modules in the Redundancy CPU racks.
6-14
PACSystems™ Hot Standby CPU Redundancy User’s Manual – March 2009
GFK-2308C
Index
#
#OVR_PRE, 3-3, 5-4
%
%S references, 5-4
#OVR_PRE, 3-3
A
Active unit
defined, 1-4
non-synchronized, defined, 1-4
B
Background Window timer
configuring, 4-5
different for redundancy CPUs, 3-3
disabling, 5-18
Backup Unit
defined, 1-4
switching control to, 5-15
commanding from program, 5-15
switching times, 5-15
validating the input scan, 5-14
validating the logic solution, 5-14
Bus Controller, Genius, 3-5
configuring, 4-13
dual GBCs at same end of bus, 3-13
switching, 5-20
C
Communications
terminating, 6-6
Configurable fault groups, 6-8
Configuration
hardware, 4-4
storing (downloading), 4-16
Constant Sweep mode, 4-4
Contacts, timed, 5-19
CPU parameters, 4-4
Faults, 4-6
Redundancy, 4-7
Settings, 4-4
Transfer List, 4-8
CPU Redundancy
defined, 1-4
Critical component
defined, 1-4
D
Data transfer, 5-7
from backup to active unit, 5-11
inputs, 5-7
outputs, 5-8
ranges, configuring, 4-8
time, 5-8
variables, configuring, 4-14
Definitions, 1-4
Disable data transfer copy in backup unit,
5-12
Downloading configuration, 4-16
Dual Bus
defined, 1-4
Duplex Genius output mode, 3-14
E
Error checking and correction (ECC), 3-3,
5-18
fault configuration, 4-6
Ethernet controller
configuring communications window, 4-4
Ethernet global data
consumption, 5-24
produce in backup mode, 5-24
production, 5-23
redundant IP addresses, 5-21
Ethernet Interface
parameters, 4-11
Ethernet network interface unit (NIU), 3-5
Ethernet remote I/O, 3-6
F
Fail Wait time, 4-7, 5-6
Fault actions, 6-7
configuration, 6-7
configured differently for redundancy CPUs,
3-3
Fault detection, 6-1
Fault groups
configurable, 6-8
non configurable, 6-9
Fault messages for redundancy, 6-2
Fault response, 6-5
G
Genius blocks
configuring, 4-13
installing on same end of bus, 3-13
Genius bus controller, 3-5
Genius Duplex output mode, 3-14
GFK-2308C
Index-1
Index
Genius HSB operation, 3-10
output control, 3-10
Genius I/O, 3-11
description, 6-10
Output control
Genius, 3-10
Output data transfer, 5-7
H
P
Hot Standby (HSB)
CPU redundancy, 1-1
features, 1-2
defined, 1-4
Genius operation, 3-10
Genius output control, 3-10
Hot Swap, RX3i, 3-4, 6-11
RMX modules, 3-4
Parameters, 4-4
PID function blocks, 5-19
Powerup, 5-2
Preferred master, 5-16
Primary unit
defined, 1-4
powerup sequence, 5-2
Produce in backup mode, 5-24
Programming
I
online, 1-3
I/O scan sets, 5-19
I/O systems
Q
description, 3-5
IEC Transitionals, 5-7
Input data transfer, 5-7
Interrupts
Quickstart, 2-1
R
not available with Redundancy CPUs, 3-3
Racks
L
LEDs
operation when link is removed, 5-3
RMX, 3-4
Links
Redundancy
failing, 6-6
removed, 5-3
Local I/O, 3-5
M
configuration wizards, 4-2
defined, 1-4
memory usage, 4-8
parameters
CPU, 4-7
Redundancy CPUs
Multiple I/O scan sets, 5-19
N
description, 3-2
differences from other CPUs, 3-3
powerup, 5-2
Redundancy link
Non configurable fault groups, 6-9
Non redundant operation, 3-3
configuring, 4-1
Non-synchronized active units (NSAU)
conditions, 4-20, 5-3, 5-4, 5-7
defined, 1-4
split control, 5-21, 6-10
repairing, 6-12
defined, 1-4
failures, 6-6
removed, 5-3
Redundancy Memory Xchange (RMX)
module
description, 3-4
faulting, 6-6
parameters, 4-10
Redundant IP addresses, 5-21
defined, 1-4
O
References, 1-5
Repair
Offset, 5-24
Online programming, 1-3
Online repair, 1-3
Index-2
backplane version required for RX3i
redundancy, 3-1
for redundancy systems, 3-1
RX7i redundancy does not support VME
racks, 3-1
online, 1-3
Resynchronization, 5-3
PACSystems™ Hot Standby CPU Redundancy –March 2009
GFK-2308C
Index
Role switch
Transfer time
commanding, 5-15
defined, 1-4
sweep impact, 5-15
estimating, 5-8
reducing, 5-10
Transferred variables, 4-14
Transition contacts and coils, 5-7
Run mode store (RMS), 1-3, 4-17
Run/Disabled mode, 5-17
W
different for redundancy CPUs, 3-3
S
Watchdog timer
Genius bus, 3-11, 3-13
setting, 4-4
Scan sets
multiple, 5-19
Wizards, 4-2
Scan synchronization, 5-5
Secondary unit
defined, 1-4
powerup sequence, 5-2
Service requests
26, Implementing preferred master, 5-16
26, Role switch from program, 5-15
27, Write to reverse transfer area, 5-11
28, Read from reverse transfer area, 5-11
43, Backup qualification, 5-14
43, Disable data transfer copy in backup
unit, 5-12
Split control, 5-21, 6-10
repair of, 6-12
Status bits
redundancy, 5-4
Stop I/O Scan mode
not available with Redundancy CPUs, 3-3
Stop to Run mode transition, 5-17
different for redundancy CPUs, 3-3
Storing configuration, 4-16
Sweep time synchronization, 5-7
Switchover time, 5-15
Synchronization
operation, 5-3
scan, 5-5
Synchronized
defined, 1-4
System Communications Window, 4-4
System upgrades
CPU types, 1-3, 6-11
T
Templates, 2-1, 3-6
Timed contacts, 5-19
Timer function blocks, 5-19
Transfer List
data transfer, 5-7
defined, 1-4
individual variables, 4-14
memory used, 4-8
ranges, 4-8
report, 4-15
GFK-2308C
Index
Index-3
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement