Cryogenic application and measurement offset with SCXI-1125 and SCXI-1327

[coalflame]

Dear all,

I am trying to measure temperature such as that of liquid nitrogen. In the application, a fine wire T-type thermocouple is soaked in liquid nitrogen and connected to SCXI-1327 (in which a thermistor serves as CJC) and then to SCXI-1125, and finally linked to a PCIe-6351 card.

The temperature of liquid nitrogen should be about -196 deg C. However, if I set the CJC source as "build in", it reads room temperature accurately but for liquid nitrogen it reads -179 deg C, if I change the CJC to "constant" with value obtained reliably from another sensor, it reads room temperature 3 K lower but more sound for liquid nitrogen as -195 deg C.

And the relationship between the constant value set for CJC and the measurement result for liquid nitrogen seems to be nonlinear: 25 deg C vs. -163 deg C; 15 deg C vs. -183 deg C; 10 deg C vs. -195 deg C.

Can anyone give some suggestions? Thank you in advance.

Su

[AnalogKid2DigitalMan]

What is the accuracy of your thermocouple?

Standard T units are typically around 1 deg C or 1.5%

-AK2DM

[johnsold]

I seem to recall that temperature gradients can be an issue in some thermocouple measurements.  You may have a rather large gradient where the wires leave the dewar.

See page Z-29 of the Omega Engineering document on thermocouples. It describes the gradient effect.

Lynn

[coalflame]

AK2DM,

The thermocouple was a standard T type one. So it seems the offset is too remarkable.

Su

[coalflame]

Lynn,

I read the document about gradient effect. But maybe that's not the problem. Just above the level of the liquid nitrogen there is a big temperature rise, but my previous experience with a high temperture measurement where the gradient can be expected to be much larger yielded good result (the error indicated by melting salt experiment was within 5K).

Supposing the large temperature gradient could account for the offset, how can we explain the different outputs given by "constant" and "build-in" setting for CJC? Value assigned to "constant" is certain to be a precise one and the SCXI-1327 module was placed distant from temperature variation region.

I changed the task to sampling voltage (avaraged to). Of course the reading didn't consider CJC. At room temperature it read -0.17mV and in the liquid nitrogen it showed -6.07mV.

Can you give some further suggestions?

Su

[johnsold]

Su,

The room temperature reading seems reasonable.  It is just a few degrees from the reference temperature.   According to Omega's thermocouple charts -6.07 mV on a Type T thermocouple indicates about -236 degrees.  As that is below the reported freezing point of nitrogen (-210), it seems unrealistic.

Is it possible that you are picking up some other current in the thermocouple wires, such as power line frequency interference?

Lynn

[coalflame]

Lynn,

Thank you for your prompt reply. I am sure no external current can be picked up.

The two values measured by "voltage task", I think, should be the combination of voltages generated by the thermocouple junction and the terminal in SCXI-1327. The actual room temperature was around 13-16 deg C.

Su