See attached graph. I have a PXI-6229. There are two temperature measurements from identical RTD amplifiers plugged into adjacent channels in Referenced Single Ended (RSE) mode. Note that one of the channels is considerably noiser than the other. I have swapped everything upstream of the A/D card, yet the noise stays with this channel. Why?
Thanks for posting you question in our forums. I noticed you have also opened a phone support service request for this issue with Joe S. I'll make sure to work with him on this issue proceeding from here. My understanding is that each of the RTD's are connected to channels 5, 6, and 7. Are all of these RTD's in a very similar environment?
Yes, all of the RTD's are in the same environment. We have swap RTD's, cables, amplifiers, to try to make the problem move in order to isolate the problem. We have determined that the problem is NOT upstream of the SCB68 which is connected to the PXI-6229. Channels 0-4, 11, 14 & 15 are open. Channels 8-10, 12 & 13 are connected to voltage inputs.
I am waiting for test results from one of the machines in the fields to quantify the signal going into channels 5 & 6 with an o-scope.
Assuming you're using the SHC68-68-EPM cable, the wires are configured in shielded twisted pair bundles. The AI+ and AI- lines (if you were measuring differentially) are each twisted together. You can see the full wire pairings here. For a visual idea of the cross-section of the cable, you can refer here.
Anyway, due to this architecture AI5 ends up being paired up with AI13 (since AI13 is also AI5- when measuring differentially). It seems quite possible that the noise could be coupling onto AI5 from the signal on AI13. I'd suggest trying to disconnect AI13 (or even ground it if you are still seeing noise) to verify whether or not this is the problem. AI6 and AI7 would be paired with AI14 and AI15, which it sounds like are unused in your application so this would be consistent with only seeing noise on AI5.
I'd suggest configuring and connecting the hardware in differential mode since you only have 8 signals. However, to avoid too much rewiring it might be easier to simply use AI11 (which is paired with the unused AI3) for the power supply measurement rather than AI 13.
If this is not the problem then I'm not sure what the issue could be, let me know if this doesn't help and I can look into it a bit more thoroughly. You should also confirm what cable and terminal block you are using.
I am pretty much stuck with RSE and can not move to Differential. I have tried disconnecting AI13 and connecting to ground. No improvement. I have tried grounding AI13 and AI15. No improvement.
Joe mentioned that you have potentially discovered the problem--tieing the ground from your power supply to the ground of the DAQ device through pin 59. Pin 59 is tied to the shield of the twisted pair that includes AI5 and AI13, so any ground potential difference would be dropped across this shield carrying your analog signals and would be tightly coupled to the actual signal on the line.
If Differential isn't an option perhaps this would be a good time to point out NRSE as an alternative. This involves connecting your external ground reference to the AI Sense terminal. Doing this would prevent the ground loop you are seeing with the two dissimilar grounds tied together through the shield on the cable. As long as the ground reference of your power supply is not drifting relative to the DAQ device (presumably they are both tied to earth ground in some way) then NRSE should be a reasonable approach.
You'll probably still want to avoid having two different voltage signals on the same twisted pair if you can help it, but it sounds like in this case this wasn't the cause of the problem.