PLC 5/40 E offline bench test

[shockalock 0]

Recently I have had the need to test PLC 5/40E processors when they come back from repairs. I am looking for a great way to test the processor on the bench testing area. Currently I have created a simple routine which loads up 2000 integer elements, with ones. A timer and counter is used to increment the test. When the 1999th element is loaded, the process resets. I am only checking this operation to be completed for what ever the accumulator has been set to. Also the timer is made faster to complete the test quicker and so on. Any help with ideas for doing a great PLC bench test would be great....

[TimWilborne 108]

I guess I just don't quite get it, what you are testing for?

[shockalock 0]

A reliable test to insure that the processor is worthey of production machinery, before placing it on the store room shelf as good. The real test is to place the processor in production. This is not possible because production is running 24/6. Currently I am just testing basic memory, timers, counters, and logic operations. I want to create a much deeper bench test to use before placing the processor in stock as spare.

[TimWilborne 108]

Sorry, but I still don't get it. It turns on...it works...you can't make a PLC "think" any harder by putting some complex code in it. I would be much more worried about it operating under varying voltages and the vibrations of an industrial environment. If you sent it back to AB then it should be fine. If it is a critical application and you are worried about the rebuild units then you may be better off purchasing a new spare.

[shockalock 0]

Thanks for the responses and the positive AB comment. I feel very confident with there products and capabilities as well. As always I want to push things to the envelopes, and am looking to stress the processor performance to find operational issues.

[Mike Dyble 0]

If the CPU can process one simple rung of logic then it will carry out any of the other predefined instructions. All that increasing the number of instructions does is increase the scan time. PLC's are 'interpretted' code (generally) and process each instruction in turn with the CPU on a fixed clock cycle. So increasing the complexity of the loaded program will not make the test any more demanding. PLC either work or dont in this context, and if they dont they wont even enter run, or even communicate with the programming terminal (usually). The only issues which it would be worth testing are temperature deviations (High and low) and as mentioned by another contributor vibration. This could identify a previously undisclosed fault which would then become apparent. Other than that it is worth leaving the processor powered and running for a period of time (soak test). One place I worked we used to have a rack CPU and I/O offline from the process into which cards returned from repair spend anything up to 1 week before being put away. This did uncover a problem on one or two occasions. Remember to remove batteries before storing, otherwise they will be flat when you get the card out in 2 years time. Better to fit a new battery when you put the card into service. I have never seen a CPU failure where the CPU can only process certain instructions or access certain memory locations. (in 28 years working with PLC's). The nearest to this scenario was on an S5 115U, which was running from an EPROM, on a hot day the processor would go to stop intermittently, it turned out that one bit in the memory module was toggling its state when it got hot, and this changed the meaning of the instruction and caused an error. This was many years ago and replacing the EPROM cured the problem. And I have never had a similar problem before or since. Mike

[kaiser_will 65]

I suggest you build up a spare PLC rack with some I/O and some physical inputs. Pop the spare CPU into the rack, power up, load in a test program (something simple coming from the inputs), put the rack into run mode and verify proper operation online. Nuff said. I use our test rack for training techs and new controls engineers. Fresh meat. On a related note, beware of the plastic bands (or clips) in the PLC-5 rack CPU slot inserts. I did have one occasion in which one of the bands got stuck (on the CPU I was pulling out) so the new CPU that I slid in pushed the orphan band into the card edge fingers and everything went to crap on power up. It is possible that the plastic bands was a PLC-2 thing that eventually went away.

[shockalock 0]

I did not mention that I have made five outputs toggle on with counter operation 1 to 5. Each out put then goes to a input. It was a few months back since I threw together the routine. It works simple and sweet. Thanks for the advice.

[BobLfoot 302]

I've read most of the posts and I agree that the CPU is good if it turns on and can execute a single instrction. That said if it is an AB PLC and you want to do a thorough bench test prior to restocking be sure you check each and every communication port. I've had PLC 5's that lost one of their comm boards thus rendering channel 1 inoperable while channel 2 was functional.

[paulengr 44]

PLC 5/40E's self-check their own internal hardware. If you can manage to download ANYTHING at all and put it into run, the processor is fine. That being said, I've only had one bad processor in 10 years from AB (a Micrologix 1100...outputs wouldn't work from factory stock). As mentioned before about the only thing that the self-test does not check is the physical inputs/outputs. If it will successfully communicate in/out all the communication ports and to at least one card in a back plane, then you've verified all those functions. The communication ports are actually on little daughter boards sandwiched in between the main boards so it is possible for things to go wrong on them independently.