Heating up NI9212 results in a temperature drop

[Epicurus]

Hello everyone,

I am trying to assess how a temperature change in the Acquisition Card can affect the temperature reading of my thermocouple. 

I am currently using a NI9212 acquisition card in high-frequency mode (95 Hz) to acquire the temperature with an Omega thermocouple Model 5TC-TT-KI-40-1M.

What I did was to keep my TC in an oil bath at a constant temperature and heat the acquisition card with a heat gun so that only the junction thermocouple-acquisition card can feel the temperature step. You can see the result of such a procedure in the image in the attachment.

I don't understand why the temperature reading is affected so much by the heat gun. Furthermore, while the acquisition card has been heated with hot air, the temperature reading results in a drop of tens of degrees. 

Anyone has already had experience with similar behaviour? Any help will be more than appreciated.

Feel free to ask me any details you think are relevant to understand my problem and help me. Thank you very much.

[mj0541]

Did the NI 9212 return to reporting the correct temperature after you removed the heat gun and allowed it to cool down?

Also, how hot were you heating the NI 9212?

[Epicurus]

Yes, it returned to give the same temperature but it took really a lot of time (more than half an hour).

I don't know the exact temperature of the acquisition card nor of the heat gun, but I am pretty sure I stayed within the safety limits of the card.

[mj0541]

It occurred to me that thermocouples measure a differential temperature so by heating the NI 9212 module, the module is sensing that the other end of the thermocouple is colder than the terminal at the NI 9212. This is assuming that the oil bath is at roughly room temperature.

[Henrik_Volkers]

What you see is the loose thermal coupling between your CJ temp sensor and the actual CJ.

Did you use the TB-9212 ? 

I haven't used it myself, but did a lot of TC calibrations in a cal lab in a former live 😉

We used a dedicated ice point....

If you treat your device with an heat gun, you shouldn't be afraid to open it and see if you can improve the CJ thermal block.

Better thermal coupling between the CJ temp sensor and the actual TC CJ and

build a thermal low pass around your CJ.  all cables will conduct heat 😉

[Pater_Mater_Filius]

Hello Epicurus,

I presume that the source of the problem is in the junction of the Thermocouple wire and Analog Input of the NI 9212. As you heat the Acquisition Card the temperature of the input metal increases in tens (maybe hundreds) of degrees, but the TC wire is cooler than the Acquisition Card input. So this procedure causes the drop of temperature during the acquisition (the junction is similar to a junction of TC so it acts like a second TC in this case). I hope this was useful to you, but for understanding the situation further, could you let me know which CJC source have you chosen for the acquisition?

[Epicurus]

Hi Henrik,

I was indeed using the TB-9212.

Thanks for your nice suggestions!

[Epicurus]

Indeed I was thinking that the problem could be related to that. 

When I put the CJC = constant (25°C) the temperature drop was indeed very much accentuated.

What is a bit surprising to me is that even with the CJC = Built-in (I am using the TB-9212) I still get a similar phenomenon. If my TC is in ice it registers 0°C. When I start to heat the Acquisition card with my heat gun, the temperature rises to 1-2 °C. Do you think that this is a real temperature felt by the TC or is it again related to a CJC problem?