thermocouple noise

[jdsnyder]

These four graphs might help. 

I don't know enough about how labview operatess to properly interpret their meaning.

Setup:  channel 0 positive and negative sides shunted together with a copper wire 

Graph 1, maximum and minimum voltages declared to be +/- 10V

Graph 2, maximum and minimum voltages declared to be +/- 5V

Graph 3, maximum and minimum voltages declared to be +/- 1V

Graph 4, maximum and minimum voltages declared to be +/- 0.2V

notes
graph 1) Theoretically possible voltages = 0.0, +/- .000305, +/- .000610, etc.

              Actual voltages .000600, .000280, -.000050, -.000370

graph 1)  Average voltage of plot = roughly .00012.  Also, plot seems relatively symmetrical about this value.

               => a) noise on average has a positive voltage (?)

               => b) something wrong with labview's (or board's) division, enunciation of voltages (?)

graphs 2-4) plots lose their symmetric nature.

               => something wrong with labview's (or board's) division, enunciation of voltages (?)

graphs 1-4)  As noted previously, noise is significantly greater than the noise stated by the board's spec sheet.

Thanks,

Jon Snyder

[Brandon_B]

Hey Jon,

Have you tried using the LabVIEW example Cont Acq Thermocouple Samples-Int Clk.vi?  I noticed that you were measuring the voltage coming from the thermocouple, but this example will allow you to specify the thermocouple type and then convert it to degrees Celsius.  The example can be found in LabVIEW by going to Help>>Find Examples and then clicking on  Hardware Input and Output>>DAQmx>>Analog Measurements>>Temperature>>Cont Acq Thermocouple Samples-Int Clk.vi.  Could you post a screenshot of what your graph on the front panel looks like.

[jdsnyder]

Hello Brandon,

Yes I have tried the sample VI.

I may be close to the answer now.  I originally purchased one board, and as noted at the start of the thread, I tried using it to read temperatures on three different computers - to no avail.

I just received two additional boards. 

new board 1 - on one of the three computers noted above) I get one temperature that is a little bad, three temperatures that are very bad.

new board 2 - on a computer I hadn't tried before)  All temperatures good!

Next step - I have to wait for the computer czar to unlock the computers - then I will switch the other boards into the computer that is working.

A) if temperatures are then good, it is a computer related problem

B) if temperatures don't work, it is a board problem

Stay tuned - same bat time, same bat channel

[ASTDan]

You might try shielded Thermocouple cable

http://www.omega.com/pptst/EXPP-K-TWSH-UL.html

Connect the shield to the AI gound of you DAQ system.

[jdsnyder]

I already use shielded cables, but thanks for the suggestion.

Jon

[jdsnyder · Accepted solution]

The noise originated from a graphics card.  After removing the graphics cards, all machines are showing 50 ma of noise.

This is liveable.

Jon

[hjz]

Hi, Jon

     Can you tell me what's the graphics card you removed? I had the same problem. My voltage results also had a big noise.   Thank you very much!

Junzhan

[MosaicPaul]

It looks to me from the aliasing of the noise that you describe that the noise could be a much greater frequency than the slow thermocouple signals you're trying to measure. There's often no substitute for a little filtering in the hardware. A few small ceramic capacitors can do wonders, as shown in this schematic for thermocouple noise filtering.  You might want to try something similar.