PACSystems RX3i CPUs, IC695CPU310-CC v2.90, GFK

PACSystems RX3i CPUs, IC695CPU310-CC v2.90, GFK
December 22, 2004
PACSystems™ RX3i CPU
IC695CPU310-CC with Firmware Version 2.90
This document contains information that is not available in any other publication;
therefore, we recommend that you read and save it for future reference.
This is an update to the CPU firmware. Support for the following modules has been added:
The new RX3i PROFIBUS Master module (IC695PBM300)
Genius Bus Controller Modules: IC694BEM331 and IC693BEM331
Serial IO Processor Modules: IC694APU305 and IC693APU305C
This update also resolves the problems described in “Problems Resolved by this Version“ on
page 3.
Version Identification
Hardware Identification
Firmware Identification
Catalog Number
Circuit Board ID
Firmware ID
CPU Primary
2.90 Build 50A4
BOC Build 28A1
CPU Boot
2.50 Build 25A3
CPU Programmable Parts
Part ID
IC695CPU310-AA is field-upgradeable to firmware version 2.90 by ordering the upgrade kit
below or downloading from the services web site. The hardware is not field upgradeable.
Upgrade Kit: 44A752290-G03
Important Product Information
Product Documentation
PACSystems CPU Reference Manual, GFK-2222
PACSystems RX3i Hardware and Installation Manual, GFK-2314
TCP/IP Ethernet Communications for the PACSystems, GFK-2224
C Toolkit for PACSystems User’s Manual, GFK-2259
Proficy™ Machine Edition Getting Started, GFK-1868
Proficy Logic Developer – PLC Programming Software Getting Started, GFK-1918
Datasheet, PACSystems RX3i CPU, GFK-2316
IPI, PACSystems RX3i CPU, GFK-2329D (this document)
CPU Functional Compatibility
Machine Edition Logic Developer 5.0 or later must be used to configure and program the RX3i.
Service Pack 2 is required to support the new features in release 2.90.
C Toolkit
The C Toolkit for PACSystems is distributed with Machine Edition Logic Developer 5.0 or greater.
Toolkit build 50A1 or later is required for use with the RX3i. Please note: The Series 90 Toolkit
(IC641SWP709/719) is not compatible with PACSystems.
Series 90-30
The PACSystems RX3i supports the following Series 90-30 modules:
Discrete Input Modules: IC693ACC300,
Discrete Output Modules:
Discrete Combinational: IC693MAR590, IC693MDR390
Analog I/O Modules: IC693ALG220/221/222/223/390/391/392/442
High Speed Counter: IC693APU300
FANUC I/O Link: IC693BEM320, IC693BEM321
Motion Control: IC693DSM314, IC693DSM324
Genius Bus Controller Module: IC693BEM331
Serial IO Processor: IC693APU305
All other Series 90-30 modules are not currently supported.
IC694 Module
The PACSystems RX3i supports the following IC694 modules:
IC694ACC300, IC694MDL230/231/240/241/632/634/645/646/654/655
Discrete Output Modules:
Analog I/O Modules: IC694ALG220/221/222/223/390/391/392/442
High Speed Counter: IC694APU300
Fanuc I/O Link Modules: IC694BEM320, IC694BEM321
Motion Control: IC694DSM314, IC694DSM324
Genius Bus Controller Module: IC694BEM331
Serial IO Processor: IC694APU305
Expansion Backplanes: IC694CHS392/398
Expansion Power Supplies: IC694PWR321/330/331
IC695 (PCI)
Ethernet Interface: IC695ETM001
Universal Analog Input Module: IC695ALG600
Profibus Master Module: IC695PBM300
Series 90-30
Expansion Rack
The PACSystems RX3i supports local and remote Series 90-30 expansion racks.
PACSystems RX3i CPUs do not operate in a Series 90-30 Rack.
Important Product Information
Series 90-30
Main Rack
Series 90-30 Main Racks cannot be used in a PACSystems RX3i system.
Series 90-30 CPUs do not operate in PACSystemsRX3i Racks.
Isolated 24V
In applications that use the IC69xALG220/221/222, consult PACSystems RX3i Hardware and
Installation Manual, GFK-2314 for details of wiring the 24V power.
Problems Resolved by this Version
Ref Address field of
I/O Fault Table for
Universal Analog
The I/O Fault Table Ref Address field now always displays the correct value for the Universal
Analog Input module and circuit faults. (This problem was reported in the IPI for the Universal
Analog Module IC695ALG600, GFK-2348.)
Error with a Flash
Write or Clear
The CPU firmware no longer reports faults for large flash writes or clears.
Faults with Hot
Swap and DO_IO
Previously, hot-inserting a module that does not support DO_IO into a slot configured for
DO_IO, or vice versa, caused the CPU to record faults. This has now been fixed.
Hot insertion of
Series 90-30 smart
module to ETM slot
If a Series 90-30 smart module is hot-inserted into a slot that has been configured for an RX3i
Ethernet module (ETM001), the fault table does not list multiple Extra Module faults and the
module is not repeatedly reset.
CPU goes to
Module Fault
If an unsupported module is hot-inserted into the RX3i system, an Unsupported Module fault is
generated, and the CPU goes to STOP-HALT mode.
“Lost” or late
interrupt handling
for the Universal
Analog Module
This CPU pulls all interrupts from a Universal Analog Input module (IC695ALG600) when the
module sends many interrupts in a short period of time. It will not appear to have lost interrupts
or interrupts that execute late.
Module fault for
DSM module
After a power cycle, a DSM 314 or DSM324 module will not generate an Unsupported Module
New CPU Features and Enhancements
Feature Description
This version of the RX3i CPU supports:
The new RX3i PROFIBUS Master module (IC695PBM300)
Genius Bus Controller Modules: IC694BEM331 and IC693BEM331
Serial IO Processor Modules: IC694APU305 and IC693APU305C
CPU Restrictions and Open Issues
Ethernet Disconnect During
Word for Word Change
If the Ethernet connection is broken during a word-for-word change, the programmer
may not allow a subsequent word-for-word change after reconnecting due to the fact
that it thinks another programmer is currently attached. If this occurs, you should go
offline and then back online again.
Simultaneous Clears, Loads
and Stores Not Supported
Currently, PACSystems CPUs do not support multiple programmers changing CPU
contents at the same time. The programming software may generate an error during
the operation.
Power Cycle During Online
If the user stores a folder to flash that is configured to power up from flash and then
subsequently power is cycled in the middle of a Online Edit session, the programmer
will still indicate that the Online Edit session is in progress after the power cycle. The
user should cancel the Online Edit session to continue.
Important Product Information
Power Cycle During Write to
If the CPU is power cycled during the process of writing to flash, and is configured to
power up from flash, a fault will be generated on power up.
Hardware Configuration Not
Equal After Changing Target
If the user stores a hardware configuration to flash which indicates that “Logic/Config
Power up Source” is set to “Always Flash” or “Conditional Flash” and then
subsequently changes the name of the target in the programming software, the
hardware configuration will go Not Equal and will not Verify as equal.
PLC and IO Fault Tables
May Need to be Cleared
Twice to Clear Faulted State
Both PLC and IO fault tables may need to be cleared to take the CPU out of
Stop/Fault mode. If one of the tables contains a recurring fault, the order in which the
tables is cleared may be significant. If the CPU is still in Stop/Fault mode after both
tables are cleared, try clearing the fault tables again.
Setting Force On/Off by
Storing Initial Value
Once a force on or force off has been stored to the PLC, the user cannot switch from
force on to force off or vice-versa directly by downloading initial values. The user can
turn off the force by doing a download, and then change the force on or off by another
Storing Large Numbers of
Blocks to Flash
Currently, storing logic with approximately 200 blocks or more to flash may fail, due to
the programmer timing out.
Number of Active Programs
Returned as Zero
The SNP request Return Controller Type and ID currently returns the number of
active programs as zero.
Serial I/O Failure at 115K
During Heavy Interrupt Load
Rare data corruption errors have been seen on serial communications when running
at 115K under heavy interrupt load on the PLC. Under heavy load applications, users
should restrict serial communications to 57K or lower.
RAND_MAX and rand()
function incompatible
In the C Toolkit, the RAND_MAX system variable is defined as a 32-bit integer.
However, the rand() function returns a 16-bit integer. By definition, rand() should
return an integer between 0 and RAND_MAX.
Incorrect Commreq Status
For Invalid Program Name
The program name for PACSystems is always "LDPROG1". When another program
name is used in a CommReq accessing %L memory, an Invalid Block Name (05D5)
error is generated.
SNP ID not always provided
Unlike the Series 90-30, the RX3i CPU’s SNP ID will not appear in the Machine
Edition programmer Show Status display. Service Request 11 will always return
Second programmer can
change logic while in Test &
Edit mode
While currently active in a Test and Edit session using Machine Edition on one PC,
Machine Edition running on another PC is not prevented from storing new logic to the
FANUC I/O Master and
Slave operation
Scan sets on the master do not work properly for the first operation of the scan set
after entering RUN mode. They do work properly for subsequent scans.
After downloading a new hardware configuration and logic, a power cycle may be
required to resume FANUC I/O operation.
Use PLCs of similar performance in FANUC I/O networks. If a master or slave is
located in an RX3i system, the other PLCs should be RX3is or Series 90-30
Must Have Logic If
Powering-Up From Flash
If the application will configure the CPU to retrieve the contents of flash memory at
power-up, be sure to include logic along with hardware configuration when saving to
flash memory.
Avoid Ethernet module
Resetting the Ethernet module, either by Service Request 24 or by the restart
pushbutton, will cause some of the CPU’s internal memory to be consumed if EGD is
configured. The problem becomes more pronounced with larger EGD configurations.
Power-cycling the system will recover the internal memory.
Configuration mismatch with
unsupported module causes
PLC sequence store failure
If you attempt to store a hardware configuration to the CPU that has a module
configured for a slot that physically contains an unsupported module, the store will fail
with a sequence store failure.
Two loss of modul faults for
Universal Analog mocule
Occasionally the hot removal of the Universal Analog Input Module (IC695ALG600)
results in two “Loss of I/O Module” faults instead of one.
CPU logs incorrect fault for
When the CPU detects an overtemp condition, the overtemp fault causes the CPU to
go to STOP/HALT instead of STOP/FAULT.
Important Product Information
Powerup of High-speed
Counter may take up to 20
As power is applied to a Series 90-30 High-Speed Counter module, the "module
ready" bit in the status bits returned each sweep may not be set for as long as 20
seconds after the first PLC sweep, even though there is no "loss of module"
indication. I/O data exchanged with the module is not meaningful until this bit is set
by the module. c.f., pages 4-3 to 4-5 of GFK-0293C
Lost count at powerup for
Serial IO Processor
A Serial IO Processor module will lose the first count after every powerup or every
time the module receives a configuration.
Info fault at power up
Intermittently during powerup, an Informational non-critical CPU software fault may be
generated with fault extra data of 01 91 01 D6. This fault will have no effect on the
normal operation of the PLC. But, if the hardware watchdog timer expires after this
fault and before power has been cycled again, the outputs of I/O modules may hold
their last state, rather than defaulting to zero.
Serial Port Lock-up after
Bad Modbus Message
If a badly-formed Modbus RTU message is sent ot the PLC, the serial port will lock
up. Power must be cycled to recover proper port operation.
When power is cycled on a rack in a system containing a Series 9030 smart option
module or if the module is hot-swapped in an RX3i rack, the module may
intermittently fail in one of three ways:
1) There is a "loss of module" fault,
2) There is a "System Configuration Mismatch" fault, or
3) There is an "Unsupported hardware" fault.
In all three cases, the module will not be scanned until power is cycled on the module.
In the latter two cases, the PLC fault table must be cleared before the PLC can be
placed into RUN mode. These failures are much less likely if the module is in the
main RX3i rack or if the module is in an expansion rack and only the expansion rack
power is cycled. If operation of the module is critical in the application, then the user
logic should monitor the fault locating reference for the module, as described on
pages 12-11 and 12-12 of GFK-2222.
On rare instances, Machine Edition may disconnect from the PLC after a download.
Faults possible when power
cycling or hot-swapping
smart modules in an RX3i
Lost Connection to Machine
Possible Machine Edition
disconnect when multiple
Genius Bus Controllers are
present in expansion/remote
Timed interrupt response
time increased
Extended Memory Types for
IO Triggers
If a system contains multiple Genius Bus Controllers in expansion/remote racks, the
Machine Edition software can time out its connection to the PLC on a clear operation
or a store of configuration. For each Bus Controller in an expansion/remote rack
there is a 3-second delay added to the time required for a clear/store of configuration.
The default connection timeout is 10 seconds and the default request timeout is 16
seconds. These values should be increased by at least 3 seconds per Genius Bus
Controller in an expansion/remote rack. This is true for both Ethernet and serial
programmer connections.
A Genius Bus Controller in the system may impact response time for timed interrupts.
The worst case interrupt response time for a PLC system with a Genius Bus
Controller and no other Genius devices is 0.5 milliseconds. The worst-case interrupt
response time for a PLC system with a Bus Controller and maximum amount of
Genius data is 50 milliseconds.
%R, %W and %M cannot be used as IO triggers for the Universal Analog Input
module (IC695ALG600).
Important Product Information
CPU Operational Notes
power supply
placement and
The A/C Power-Supply (IC695PSA040) for the RX3i is a doublewide module whose connector
is left justified as viewed when installed in a rack. It cannot be located in Slot 11 of a 12-slot
rack nor Slot 15 of a 16-slot rack. No latch mechanism is provided for the last (right-most) slot in
a rack, therefore it is not possible to place the power-supply in the second to last slot.
2. The RX3i CPU (IC695CPU310) is a doublewide module whose connector is right justified as
viewed when installed in a rack. It is referenced for configuration and by user logic applications
by the leftmost slot that it occupies. For example, if the RX3i CPU has its physical connector
inserted in to slot 4, which means it occupies slots 3 and 4, the CPU is referenced as being
located in slot 3. The referenced location of the CPU is not determined by what slot the physical
connector is located in, but rather by the left most slot occupied by the entire module.
3. Due to item #2 above, the RX3i CPU may be located in Slot 0 of a rack (physical connector in
Slot 1). In addition the CPU cannot be located in Slot 11 of a 12-slot rack nor Slot 15 of a 16-slot
rack, since doing so would require the physical connector to be located in the slot reserved for an
expansion module.
4. When migrating a Series 90-30 CPU system to a PACSystems RX3i CPU, be aware that to
maintain the Slot 1 location of the CPU, only a singlewide power-supply may be used in Slot 0.
Currently, the only available singlewide power-supply is a DC power-supply (IC695PSD040).
Therefore, if the application using an existing Series 90-30 system must maintain a Slot 1 CPU
and uses an AC power-supply, the RX3i system must have the RX3i AC power-supply located in
a slot to the right of the RX3i CPU in Slot 1.
5. In deciding to place the CPU in slots other than Slot 1, the user should be aware of the
possible application migration issues that could arise. The following lists the areas that could be
affected when migrating an application from one CPU slot to another.
Item Affected
How Affected
User Logic
Service Request #15
Location of CPU faults will not be the standard 0.1
(Read Last-Logged
location, but will reflect the slot the CPU is located in.
Fault Table Entry)
User logic that decodes fault table entries retrieved
by these service requests may need updating.
Service Request #20
(Read Fault Tables)
Commreqs directed to the CPU (e.g. those directed
Request (Commreq)
to the serial ports of the CPU) will need to be updated
with the correct CPU slot reference.
CPU Slot location
Slot location of the CPU must be updated in the HW
Configuration to reflect the CPU’s true location.
Fault Tables
Faults logged for the
The location of faults logged for the CPU in the fault
table will not be the standard 0.1 (rack.slot) location,
but will reflect the CPU’s actual slot.
Series 90 PLCs
Remote Series 90 PLCs that use SRTP Channels COMMREQs expect the CPU
to be in slot 1. In order to support communications with Series 90 SRTP clients
such as Series 90 PLCs using SRTP Channels, the RX3i internally redirects
incoming SRTP requests destined for {rack 0, slot 1} to {rack 0, slot 2}, provided
that the CPU is located in rack 0 slot 2 (and the remote client has not issued an
SRTP Destination service on the connection to discover the rack and slot of the
CPU). This special redirection permits Series 90-30 applications that expect the
power supply to be located leftmost and the CPU to be located to the right of the
power supply to function. Attempts to establish channels with CPUs in slots
other than 1 or 2 will fail if initiated from Series 90 PLCs.
HMI and External Communication Devices
All external communication devices that interact with the CPU should be
checked for compatibility with CPU slot locations other than slot 1. Problems
may arise with, but are not limited to, initial connection sequences and fault
reporting. Machine Edition View customers should select “GE SRTP” as their
communications driver – it can communicate with a CPU in any slot.
Host Communications Toolkit (HCT)
Applications that utilize the Host Communications Toolkit may require updated
Important Product Information
Instructions for
A battery is shipped with the CPU unit behind the battery door on the faceplate but it is not
connected. Do not connect the battery until the CPU is installed in the rack and the rack powered
on. The battery may then be attached to either of the two terminals in the battery compartment.
Once that is done, the CPU may be powered down and normal battery back up operation will
begin. To save battery life, do not connect the battery for the first time until the CPU is powered up.
Machine Edition LD-PLC no longer supports a function that connects to the PLC, downloads, then
disconnects from the PLC. The connect and download functions are now separate. To perform a
download to the PLC, you must first connect to the PLC.
Logic Executed
in Row Major
Instead of
Column Major
Logic execution in PACSystems RX3i is performed in row major order (similar to the Series 90-30).
This is different from the Series 90-70, that executes in column major order. This means that some
complicated rungs may execute slightly differently on PACSystems RX3i and Series 90-70. For
specific examples, see the programming software on-line help.
NaN Handled
Than in 90-30
The PACSystems RX3i CPU may return slightly different values for Not A Number as compared to
Series 90-30 CPUs. In these exception cases (e.g., 0.0/0.0), power flow out of the function block is
identical to Series 90-30 operation and the computed value is still Not A Number.
PID Algorithm
The PID algorithm used in PACSystems has been improved and therefore PID will function slightly
differently on PACSystems RX3i than on the Series 90-30. The differences are that the elapsed
time is computed in 100 µS instead of 10 mS units. This smooths the output characteristic,
eliminating periodic adjustments that occurred when the remainder accumulated to 10mS.
Also, previous non-linear behavior when the integral gain is changed from some value to 1
repeat/second was eliminated.
Some Service
different from
Series 90-30 or
no longer
Service Requests 6, 15, and 23 have slightly different parameters. Refer to GFK-2222.
PACSystems PLCs support Service Request 26/30 functionality via fault locating references.
Service Request 13 requires a valid value in the input parameter block (Refer to GFK-2222 for
details). On the Series 90-30 and Series 90-70 the parameter block value was ignored.
Service Requests 48 and 49 are no longer supported (there is no auto-restart) because most
faults can be configured to be not fatal.
IL and SFC
IL and SFC are not available.
The Series 90-30 Enhanced DO I/O instruction is converted to a standard DO I/O instruction (the
ALT parameter is discarded and ignored.)
The Series 90-30 END instruction is not supported. Alternate programming techniques should be
MCR, &
Non-nested JUMPs, LABELs, MCRs, & ENDMCRs are translated to the corresponding nested
JUMPs, LABELs, MCRs, & ENDMCRs when converting from Series 90-30 to PACSystems RX3i.
Changing IP
Address of
Interface While
Storing a hardware configuration with a new IP address to the RX3i while connected via Ethernet
will succeed, then immediately disconnect because the RX3i is now using a different IP address
than the Programmer. You must enter a new IP address in the Target Properties in the Machine
Edition Inspector window before reconnecting.
Address for
Modbus Will
Conflict with
Other Nodes
The default serial protocol for the RX3i is Modbus RTU. The default Station Address is 1. If the
PLC is added to a multi-drop network, care must be taken that the PLC is configured with a unique
Station Address. Nodes with duplicate Station Addresses on the same network will not work
Important Product Information
Care should be taken when timers (ONDTR, TMR, and OFDTR) are used in program blocks that
are NOT called every sweep. The timers accumulate time across calls to the sub-block unless they
are reset. This means that they function like timers operating in a program with a much slower
sweep than the timers in the main program block. For program blocks that are inactive for large
periods of time, the timers should be programmed in such a manner as to account for this catch up
Related to this are timers that are skipped because of the use of the JUMP instruction. Timers that
are skipped will NOT catch up and will therefore not accumulate time in the same manner as if they
were executed every sweep.
Constant Sweep time, when used, should be set at least 10 milliseconds greater than the normal
sweep time to avoid any over-sweep conditions when monitoring or performing on-line changes
with the programmer. Window completion faults will occur if the constant sweep setting is not high
Large Number
Sent to Module
in One Sweep
Causes Faults
A large number of COMMREQs (typically greater than 8) sent to a given board in the same sweep
may cause Module Software faults to be logged in the PLC fault table. The fault group is
MOD_OTHR_SOFTWR (16t, 10h) and the error code is COMMREQ_MB_FULL_START (2). When
this occurs, the “FT” output of the function block will also be set. To prevent this situation,
COMMREQs issued to a given board should be spread across multiple sweeps so that only a
limited number (typically 8 or less) of COMMREQs are sent to a given board in each sweep. In
addition, the FT output parameter should be checked for errors. If the FT output is set (meaning an
error has been detected), the COMMREQ could be re-issued by the application logic.
C Block
Standard Math
Functions Do
Not Set errno
In C Blocks, standard math functions (e.g. sqrt, pow, asin, acos) do not set errno to the correct
value and do not return the correct value if an invalid input is provided.
The process of upgrading the CPU firmware with the WinLoader utility may fail when multiple IO
modules are in the main rack, due to the time it takes to power cycle the rack system. If the
upgrade process fails, move the CPU to a rack without IO modules and restart the upgrade
Winloader initial connect baud rate is fixed at 19200 baud. Note that the firmware download will
occur at 115.2K baud by default.
Note that if you have hyperterm open on a port, and then try to use Winloader on the same port,
Winloader will often say “Waiting for Target” until the hyperterm session is closed.
Hot Swap
Hot Swap of power supplies or CPUs is not supported in this release.
Single Power
Use only one power supply in an RX3i main rack.
Serial Port
With the following combination of circumstances, it is possible to render serial communications with
the CPU impossible:
User configuration disables the Run/Stop switch
User configures the power up mode to Run or Last
Logic is stored in FLASH and user configures CPU to load from FLASH on power up
User application issues COMMREQs that set the protocol on both of the serial ports to
something that does not permit communications to the ME programmer.
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