how do I measure temperature from a thermocouple accurately in the presence of a strong electromagnetic field

[AEI_JR]

My company routinely heats parts instrumented with strain gages to their full operation temperature and measure the reaction of the strain bridges to temperature change.  For this we attach thermocouples and generate a curve to relate 'apparent strain' to temperature.  For most applications, we use high-temperature convection ovens, which pose few issues.  For the highest temperature applications, though, we heat the parts in a kiln.  The kiln invariably produces such a strong electromagnetic field that the signal from the minerally-insulated thermocouples are supressed as read by a SCXI-1102.  However, a hand-held meter was able to read the correct temperature without problems.  Is there another NI conditioner I should be using for this measurement?

[Ian_Ren]

Have you tried

1) an "ungrounded" TC? Omega.com has quite a few TCs to choose from.

2) software average/Low-pass filtering?

Ian

[P.Sanders]

There's a good list of recommendations here for diminishing noise/electromagnetic intereference for the SCXI 1102.

[AEI_JR]

@ Ian,

Ungrounded TCs are not an option - we need to measure metal temperature with intrinsic TCs.  I am averaging and low-pass filtering, but neither of these are the issue - it's not noise, but an apparent signal suppression.

[AEI_JR]

@ psanders:

Thanks for the link.  Most of this I knew already, but the discussion at the end may have sparked a solution - the kiln receives 3-phase power which I believe may be grounded at a different location than my data system.  I'll experiment with grounding when I try this measurement again.

[AnalogKid2DigitalMan]

If it is a grounding issue, you could insert a thermocouple/millivolt isolator between the T/C and the SCXI-1102 to break the ground loop.

-AK2DM

[AEI_JR]

Like this?

http://www.mod-tronic.com/Minco_millivolt_thermocouple.html

or do you have another one you'd recommend?