Will a damaged input on a SLC analog card result in a fault?

[defcon.klaxon]

Hey guys,

Well, the troubleshooting continues for the client who got hit by lightning. New CPU has been installed, and is working! That's the good news. I've got a few things to tinker with to get working properly but the CPU was definitely damaged, so replacing it was the right call. Thanks for the help in determining that.

Current issue is some of the analog input signals aren't coming back with the new CPU, however. There are a couple 1747-NI4 modules, and I'm noticing two odd things:

1. Card 1, Input 0 is flickering between 0mA and 4mA. This signal runs several hundred yards to the headworks, so in theory lightning could have damaged it. However, Input 1 on the same card is showing normal operation so it's not like the whole card is fried.
2. Card 2 has a few inputs that are reading way high values; the scale blocks show max raw value of 16834, but one of the inputs is pegged at 32767. That worries me, since max value according to the official literature is 16384. Another input is reading a bit over 20000, again beyond the scaling max value.

My question is this: if an input on one of those cards was damaged, would it result in a fault? Currently the fault light on the CPU is off. I could inject a signal into the input and see what happens, but client is over an hour away and whenever possible I try to stay remote to keep the cost down, so I don't want to drive two hours to do a five minute test if it can be avoided.

 

[Mickey]

#2 Data format (values) depends on how you have the card configured. See pdf.

 

[defcon.klaxon]

Thanks Mickey, yeah that was the documentation I was referring to. All inputs are 4-20mA. That is, they *should be*. While trying to verify that, I can't seem to figure out how to open the "Configure" button when looking at Advanced Configuration. It's greyed out whether I'm in program or run, online or offline. Any help would be great.

 

[defcon.klaxon]

Ah, did some research and apparently the input type is configure by DIP switch, not software.

 

[Mickey]

My question is this: if an input on one of those cards was damaged, would it result in a fault? Currently the fault light on the CPU is off. I could inject a signal into the input and see what happens, but client is over an hour away and whenever possible I try to stay remote to keep the cost down, so I don't want to drive two hours to do a five minute test if it can be avoided.

I can't see how anyone could definitely answer your question ( to many variables)

If you have an input connected now and the PLC does not fault I can't see how a test signal will. On the other hand I am not the one who will have to respond to your client when it does fault.

 

[parky]

Many cards have isolation between the digital & analogue interface so it is possible that the CPU will not show a fault, it is also possible that not only the card or cards have suffered but also some of the instruments have failed. I have seen this happen, one instance we had was suspected lightning strike, however, this was not clear, one analogue card had failed on two channels but also three of the sensors had also failed, the other card was ok but one sensor had failed, 3 weeks later that card also failed, maybe related but there was no way of telling.

 

[ValeoBill]

You swapped out the PLC...what about the program? Were you able to get tthe program from it or did you download a copy? It wouldn't be the first time a backup didn't have the latest copy of the running program. I'd guess your module configurration is correct or you'd experience processor faults. So far, so good. As far as the AI go, the problem is in the wiring or the module OR both - don't assume only one problem exists. As the cards are the same, flip them. If the problem moves with the card, you've got a bad card. If it doesn't, you're input or its wiring is a problem...best of luck....let us know how you fare...

 

[OkiePC]

I have seen just about every mode of failure with SLC analog inputs. Data intermittent, Data frozen at some low or middle or zero value, data pegged at max value with any signal connected, data that seems to vary through the normal range from 4 to about 11 mA rather than 4 to 20. I even had one 16 channel card that would cause the CPU to fault and erase itself when I powered up a rack with that card installed. If there is lightning damage and a SLC analog card anywhere nearby, you have about a 60% chance that damage was done to that card in my experience.

Sometimes, unused channels can be used successfully, other times, the whole card suffers damage.

Those parts are getting rather expensive now.

Use a loop calibrator to check (measure in series) the raw signals and to validate (source mode) all the inputs...even those that appear to be working right.

 

[I_Automation]

Along with OkiePC's 4-11 value for 4-20 mA I had an analog input that read the proper values for 4-14 mA, then dropped drastically to about a 6 mA value, then at 18 mA go back to where it should. This was whether raising the current from 4 to 20 or dropping it from 20 to 4.

The SLC analog cards can take down the rack, wipe the CPU or cause weird things.

 

[defcon.klaxon]

The SLC analog cards can take down the rack, wipe the CPU or cause weird things.

I'm unfortunately seeing similar odd behavior. The CPU seems to get wiped, seemingly out of nowhere. And even though I reset the IP address with BOOTP, it's not showing up in RSLinx. Good lord, this might be a much bigger issue than just swapping a CPU.

 

[defcon.klaxon]

I can't see how anyone could definitely answer your question ( to many variables)

Sorry if I was unclear as to my actual question. My question, specifically, is if at least one channel in a 1746-NI4 is damaged, is there a way the system will tell you in Logix 500?

I don't want to tell the client to buy several new I/O cards "on a hunch". I want to verify whether or not I'd get a fault if the card was damaged, so I could tell the client "hey there is no fault, the card should be fine" or "hey there's no good way to know, they're somewhat cheap, let's try swapping one to see if that fixes the problem."

 

[I_Automation]

if at least one channel in a 1746-NI4 is damaged, is there a way the system will tell you in Logix 500?

Not for an analog input card. If it is a total failure the reading hopefully would be 0, way under the 4 mA value, but a failure my go full value or get 'stuck' on a specific value.

RTD and thermocouple cards do have a status register that can be read and the bits will show if the channel is enabled, open-circuit, under-range, over-range or a configuration error (taken from the -NT4 thermocouple book)

 

[defcon.klaxon]

Not for an analog input card. If it is a total failure the reading hopefully would be 0, way under the 4 mA value, but a failure my go full value or get 'stuck' on a specific value.

RTD and thermocouple cards do have a status register that can be read and the bits will show if the channel is enabled, open-circuit, under-range, over-range or a configuration error (taken from the -NT4 thermocouple book)

Thanks for verifying. Looks like I'll be making a trip up tomorrow to see what I can figure out.