PLC Software Trap?

Just an off the cuff question. Ive heard of some programmers maybe experiencing an intermittant or random problem in a process, when they cant narrow it down to a single item causing the problem they sometimes create a "trap" so they can see whats causing the problem. I just wondered how you guys go about doing this, i suppose one simple way would be to set a bit or pass a value when a condition is met.

Thanks

LH

What I do, and I'm sure it's pretty much the same everywhere, is use Counters, Latch, Set, RTOs, etc. Anything that is not going to automatically reset itself once false.

Say I dunno what momentary bit is causing the fault to happen, because by the time I get to check it, it's already happened and gone. I just have the bits that I think the problem is each latch (Set) a seperate bit. This lets me see that yes, that bit was true during this time period, even if it was only true for 10milli seconds.

I'll use counters in the program to catch the number of times something happens during an offshift, since operators rarely communicate their issues across shifts, and it's easier for me to have the PLC tell me exactly how many times something occured.

I'll use retentative timers if I feel like knowing for how long certain things are happening, like an analog input inexplicably falling to 0, even though everything appears fine, for only a few milliseconds.

there are different ways to stop program execution in right condition. the most equipment out there is run by some sort of sequencer and this makes it very easy to set the trap and create step mode for example.

What I've done in the past is to write a little block of code that grabs and stores a set of data on the rising edge of a trigger bit. The data was stored as a record in a data block; when a new record was added, older records were rippled down and the oldest record discarded. This method can be left running and the results can be examined at leasure. I last used this system to diagnose a problem by modem with a machine 4000 miles away.

Nick

As stated above counters are great ways to trap events or bits that don't stay on long enough for the screen to refresh.
Sometimes I add rungs that are self latching with no reset, this lets me monitor those intermittent problems and I can check back later and see if that rung ever became true.
Just use logical thinking as to what might be happening and then try and figure out out to catch it.

Dido all of the above plus I've used CLx trends to capture errors.
The most complex of these had a trigger programmed and stored trend values several seconds before the error occurred.

Steve.

Sorry about the delayed thanks.....but thanks for the replies. very helpful.


LH

Quick Trap I used Yesteday

so I could do some fabricating instead of chasing wire