NI 9211 to read 20/40 Land-Jewell thermocouple

[nsmit106]

I am trying to get my NI 9211 to read a 20/40 Land-Jewell thermocouple. Not sure how to do this. I only have options for selecting a type J, S, R, B, etc. thermocouple. But this is a different type of thermocouple. Anyone have a similar issue? or suggestions? Is there a different module that I could use?

Thank you!

[EamonnFPMcKernan]

You could apply a polynomial custom scale as described here:

https://knowledge.ni.com/KnowledgeArticleDetails?id=kA00Z0000019KyDSAU&l=en-CA

I'm not aware of an agreed upon set of coefficients. Here is one set of calibration coefficients:

E = a0 + a1T + a2T^2 + a3T^3 + a4T^4 + a5T^5

where T is the temperature in units of °C, E is the emf in units of µV, and the best-fit coefficients a are given in table 3.

Table 3. Polynomial coefficients of the best-fit reference function (current work).

a0	0
a1	0.486 309 856 857 2046
a2	−2.922 674 377 066 488  ×  10−4
a3	1.924 154 232 934 0246  ×  10−6
a4	−8.503 478 468 499 287  ×  10−10
a5	1.256 271 081 535 4167  ×  10−13

Taken from

A pan-European investigation of the Pt-40%Rh/Pt-20%Rh (Land–Jewell) thermocouple reference function

J V Pearce¹, C J Elliott¹, A Greenen¹, D del Campo², M J Martin², C Garcia Izquierdo², P Pavlasek³, P Nemecek³, G Failleau⁴, T Deuzé⁴Show full author list

Published 9 December 2014 • © 2015 IOP Publishing Ltd
Measurement Science and Technology, Volume 26, Number 1

[EamonnFPMcKernan]

You probably need the inverse equation. I wrote a quick Matlab script. and got this for T as a function of emf:

a0 = 6.074251151e+01
a1 = 1.429021109e+00
a2 = -8.825260649e-04
a3 = 3.178675751e-07
a4 = -5.408772402e-11
a5 = 3.453389924e-15

This is not a well conditioned fit. Let me know your relevant temperature range and I can re-run it for you on Monday. Here's the script BTW:

fitParams = fliplr([0 0.4863098568572046 -2.922674377066488e-4 ... 1.9241542329340246e-6 -8.503478468499287E-10 1.2562710815354167E-13]);
temperatures = 0:2000;
emfs = polyval(fitParams,temperatures);
inverseParams = polyfit(emfs,temperatures,5);
inverseParams = fliplr(inverseParams);
for iParam = 1:6
fprintf('a%d = %.9e\n',iParam-1,inverseParams(iParam));
end

[nsmit106]

Thank you so much! My temperature range is from room temperature (~20 C) to 1200 C. This is for a furnace thermocouple where I am running my experiment from room temperature up to a glass melting temperature. I am going to go to the lab today to work on the LabView and see if I can take your advice and get it running. Thank you so much for your input!

[EamonnFPMcKernan · Accepted solution]

With a 5th order polynomial fitting over the range of 20 to 1200 C I got a residual of up to 12 C. With an 8th order polynomial, the maximum residual was 1.5C, and this dropped below 0.5C above 200 C.

Three caveats:

1) I am fitting to a fit to the data. One could get the original data and fit to it directly with more accurate results, but I would expect that you'd want to calibrate the measurements yourself to achieve best accuracy, and

2) the accuracy of the voltage measurement comes into play too.

3) As stated in previously, this is just one researcher's calibration curve. I have not tried to look deeply into any other attempts to calibrate these thermocouples. It was just the first reference I found.

Here are the 8th order parameters to put into your polynomial custom scale. (Sorry it's faster to write with than LabVIEW for this kind of calculation).

a0 = -8.125596472e+00
a1 = 2.540075538e+00
a2 = -5.308654744e-03
a3 = 8.631324635e-06
a4 = -9.153316445e-09
a5 = 6.156035023e-12
a6 = -2.524423070e-15
a7 = 5.746611195e-19
a8 = -5.559416761e-23

You could also bypass the polynomial and create a table, which is also an option for the DAQmx Scale. Then you'd eliminate one error source (the residual of the fit) completely.

Hopefully this shows you where to go from here if you need to refine things.

Good luck,

Eamonn

[AnalogKid2DigitalMan]

Does the T/C need a purge gas environment at the upper range to prevent oxidation or burning if exposed to ambient air?

You may want to check. I pulled a unit out of the purge gas zone at 1800C, it turned into a glowing sparkler once it hit ambient air.

-AK2DM

[nsmit106]

Hello,

I am trying to create a custom scale. I followed the steps on https://www.ni.com/en-us/support/documentation/supplemental/18/ni-daqmx-custom-scales-and-usage-expl... . I am stuck on one part. I need to make the units From Custom Scale. I can't find anywhere online how to implement this. Where on the function palette can I find the "From Custom Scale" box?

Hope this makes sense.

Thank you!

[GovBob]

I'm trying to follow you here...are you talking about the DAQmx Channel property node>>AI.Voltage.Units? If so and you have changed that property to 'write' just  right click the input and create a constant. Select from custom scale. 

[nsmit106]

Hey,

Sorry I am so very new at this. I will try to explain my goal better. I am trying to read a 20/40 thermocouple (my furnace thermocouple) with a NI 9211. Unfortunatly, the NI 9211 only recognizes type S, R, J, etc thermocouples. Therefore, in my software, I am trying to apply a scale. I believe I need to specify in the units as From Custom Scale for the DAQmx Create Virtual Channel VI. How do I do this?

See images attached. Capture is my image. Capture 2 is from https://www.ni.com/en-us/support/documentation/supplemental/18/ni-daqmx-custom-scales-and-usage-expl...

[nsmit106]

Hey Eamonn,

I was able to create a scale. I am attaching an image of what I have (Capture.png). I need to specify the units as From Custom Scale (see attached image Capture2.png). I got this from the website you originally sent. Do you know how to specify the units as shown in Capture2.png? I really hope what I said makes sense. I can restate it if I need to.

Thanks so much!

Natalie