[swbrown]

We had a PLC 540 fault today - Major error 95 Coprocessor reported fault

Knowledege base come up with:

"16986 - Coprocessor Reported Fault.
Access Level: TechConnect
Date Created: 12/08/1998 04:14 PM
Last Updated: 04/20/2010 08:37 AM

Coprocessor Reported Fault
The control coprocessor and PLC5 combination can produce a PLC5 major error 95 (Coprocessor reported fault) when the copro goes into a system reset state. This state is usually caused by customer programs that have a pointer problem in their 'C' code that steps on something in OS-9 that forces the CPU halt condition of the copro. When the copro CPU goes into this state, it prevents an internal software watchdog from being updated. When the watchdog fires, it pulls a hardware fault line between the copro and the PLC low, to indicate a failure. When the fault line goes low, the major fault of 95 is generated in the PLC.

The task "pit_irq" in the copro is related to the watchdog. If that task is killed, it will cause a System State Exception of the copro, which halts the CPU and causes the watchdog to fire.

If you don't do a power cycle to clear it, then the PLC5 will fault again with an error 98, (Failed last 16 byte test of RAM) after everything is thought to be reset."




I didn't even know the PLC5 had a coprocessor, never mind trying to find C code and OS-9 conflicts. I was able to clear the fault, but I would like to be informed if it happens again. Where do I start?

This post has been edited by swbrown: 08 August 2012 - 01:08 PM

[Mickey]

swbrown, on 08 August 2012 - 01:07 PM, said:

We had a PLC 540 fault today - Major error 95 Coprocessor reported fault

Knowledege base come up with:

"16986 - Coprocessor Reported Fault.
Access Level: TechConnect
Date Created: 12/08/1998 04:14 PM
Last Updated: 04/20/2010 08:37 AM

Coprocessor Reported Fault
The control coprocessor and PLC5 combination can produce a PLC5 major error 95 (Coprocessor reported fault) when the copro goes into a system reset state. This state is usually caused by customer programs that have a pointer problem in their 'C' code that steps on something in OS-9 that forces the CPU halt condition of the copro. When the copro CPU goes into this state, it prevents an internal software watchdog from being updated. When the watchdog fires, it pulls a hardware fault line between the copro and the PLC low, to indicate a failure. When the fault line goes low, the major fault of 95 is generated in the PLC.

The task "pit_irq" in the copro is related to the watchdog. If that task is killed, it will cause a System State Exception of the copro, which halts the CPU and causes the watchdog to fire.

If you don't do a power cycle to clear it, then the PLC5 will fault again with an error 98, (Failed last 16 byte test of RAM) after everything is thought to be reset."




I didn't even know the PLC5 had a coprocessor, never mind trying to find C code and OS-9 conflicts. I was able to clear the fault, but I would like to be informed if it happens again. Where do I start?

Do you have one of these in the chassis?

http://literature.ro...um676_-en-p.pdf

[swbrown]

No, only an ethernet card next to the processor.

[BobLfoot]

swbrown, on 08 August 2012 - 04:23 PM, said:

No, only an ethernet card next to the processor.

I can tell you I have seen this fault on probably three or four PLC5xx {40/80} processors in the past 6 years and every time the Ethernet Daughtercard was in the process of failing or had not bee properly secured to the main PLC. I use the following approach.

1. First CoPro Fauilt - Make Note and Reset
2. Second CoPro Fault - Remove PLC and DaughterCard ENET and Dissassemble and reassemble. Blow Off any Dust and make sure all pins seat securely.
3. Third COPro Fault - repalce Daughtercard with another New One.
4. Fourth CoPro Fault - repalce PLC with a New One.

never saw a fifth fault.