Bad thermocouple measurement's on 50 Hz

[GDthomas]

We are using  9213 thermocouple modules on a Cdaq chassis to measure various temperatures including winding temps on an AC motor in a medical device. We are using J type  unshielded thermocouples. All  the readings are steady when testing the device on 60Hz but when tested at 50Hz the reading rapidly fluctuates as much as 12º C. This affects all the measurements not just the motor TC's. In both cases the IUT is on a seperate power source isolated from the Daq units source. The IUT is double insulated and has no real "ground" being mostly plastic. Any suggestions??

[SteveD123]

How are you generating the 50 Hz? IF you are using any kind of frequency inverter there could be all sorts of nasty HF noise in there being aliased down to the signal you see. Is your noise perodic in any way?

[GDthomas]

Yes we are using a frequency inverter. I know they can be noisey but I've seen this kind of problem with systems running on a genset with a clean signal. Even so I wouldn't think a change of 10 Hz would make such a big difference in the amount of noise the supply makes. I didn't mention that the 60Hz is from the same inverter supply as 50 not the line. The only thing I can think of is that the DAQ is on a different frequency than the supply but why would it cause this? Doesn't the DAQ really run on DC? Why would it care what the frequency line frequency is?

Now I'm really confusing myself.

[Henrik_Volkers]

Your TC reading device provides two COM connections. (RTFM 😉 ).  These migth help you to reduce the EMC issues.

External filtering is another option. But a propper grounding (and shielding) technique of the test setup should be the way to go.

I would use a scope or the mic input of a soundcard (with two 100nF to couple and two antiparallel diodes for protection) to have a look at the AC content of my TC EMF. 

(usually one side of these devices is closer to ground, so also swap the inputs)