SLC500 fault - got idea?

It happened today, and since I am not an SLC500 guru, I would need someone's help. This installation has been running for the past two years with no glitch, and suddenly - "... watchdog scan time has been exceeded."!? In S2 data file the watchdog is [x10ms]S:3(high byte)=160 (1.6sec?). This PLC has one remote IO rack and 4 Quick panels on the DH485 network.

Any help would be appreciated.

Regards,
zif

There are a few things...

Are there any JMP's in the logic that might be conditionally jumping backwards to create a for/next type loop?
Are there any JSR's jumping to subroutines more than once throughout the logic?

Or JSR's within subroutines? Putting JSR's in subroutines can make it really hard to see a reason for a "too long" scan time.

JMP's and JSR's would be the first and easiest to look for. Of course there are other things that could effect scan time.

Would it be possible for you to Zip the .rss file and attach it to your next post?

Go check out what's on Rung 72 of File 2; that's where the Status file is saying your program faulted. It's entirely possible that's the last rung in the program; some faults log themselves at the end of scan.

There's something fishy going on; the default Watchdog value is 10, which with the x10ms timebase is 100 milliseconds (0.10 seconds). If this application requires a Watchdog time of 1.6 seconds, then either it's got some hellacious loops, somebody (or something like the Quickpanels) has been messing with the Status file, or there's a program corruption that was misinterpreted as a watchdog fault.

An any case, I'd reset the Watchdog to a value of 10, reload the program, and watch what happens next.

Thanks for your replies!

I looked through the code, and there is no JMPs of any sort of looping in the logic, it's just plain ladder logic with subroutines listed in the file 2 (each subroutine has return and end statements at their ends), contacts, timers, counters, math (only one division with constant, that is 2), scaling and comparison.

Rung 72 in file 2 is the program end. I was partially involved (programed the Quick panels) at the beginning of the project in 1996 (retrofit was done two years ago), and I remember vaguely something very unusual with Quick panels' addressing: the alarms tags were T4:xx.DN, and they worked only after we had added T4:99.DN (Quick panel's way of AB addressing) tag as "alarm enabler".

I've cleared the fault and switch the PLC to run mode, without reloading the program. Next time I go there, I would change the watchdog timer (as far as I can remeber from other PLCs, watchdog timers are usualy less than 500ms).

Regards,
zif