Thermocouple readings NG on scxi chassis but ok using FLUKE calibrator.

[jacbenimble]

PCI-6052E
SCXI Chassis
1102 TC boards
1303 Connectors
Scanning at 1khrz, 100 samples per channel, on 96 channels.
The system will run at twice this rate with the same problem.


I seem to be having a problem with thermocouples sensitivity. After setting up a test stand that includes thermocouples, the data on a few channels will become erratic.
During trouble shooting I check the problem channels with a FLUKE calibrator and they read fine. I have also checked them with another data acquisition system with the same result, good data readings. When I plug them back into my SCXI chassis the data on the same channels is erratic again. I have removed the thermocouples and replaced them with new ones and this seems to correct the problem. My concern is, why does the SCXI chassis have a problem with some thermocouples while other systems can read them just fine? Is this erratic behavior a common problem? Could you tell me why the chassis has such a hard time with marginal sensors?

Thanks for your time, and any information would be greatly appreciated.

[AlanArmstead]

There shouldn't be any reason that an SCXI system would have any more difficulty reading a thermocouple than any other DAQ system. Problems of this nature usually occur due to bad connections somewhere (whether it be the thermocouple junction or the screw terminals that attach the thermocouple to the SCXI-1303). Do your 'marginal' sensors produce the same bad readings in other channels? Can you describe what you mean by erratic signals? Are there any differences in how the thermocouples are positioned when measuring with the SCXI system compared to the other systems? What I'm getting at is, when the thermocouples are connected to the SCXI system, is there something outside the SCXI that cause the thermocouples to work incorrectly (some are pretty sensitive). Just a few thoughts, hope it is helpful.
-Alan A.

[jacbenimble]

It seems that the proper configuration of the 1303 connector fixed the problem.

I was measuring temperatures from various sources and had a combination of both grounded and ungrounded thermocouples.

The proper configuration was to use the 10 Mega ohm resister packs for all of the pull up and bias resistor banks .

Thanks for the help.

[UCLA-engineer]

What do you mean "use the 10 Mega ohm resistor packs for all of the pull up and bias resistor banks?"  Does this mean put a 10M ohm resistor from ground to TC - and another from ground to TC +?

Please clarify as I'm a bit confused as to what you mean.

Thanks for the help.

[UCLA-engineer]

What do you mean "use the 10 Mega ohm resistor packs for all of the pull up and bias resistor banks?"  Does this mean put a 10M ohm resistor from ground to TC - and another from ground to TC +?

Please clarify as I'm a bit confused as to what you mean.

Thanks for the help.

[jacbenimble]

On page 9 of the 1303 installation manual , you will find a drawing of the internals of the connector. This shows the location of the sockets that hold the pull up and bias resister banks.

Here is a link to the manual.

http://www.ni.com/pdf/manuals/321923a.pdf

[jacbenimble]

In short, Yes.

 

The 1303 block has locations for these resisters, but could be done with other termination options.

The 1303 also has a cold junction compensation port making it the ideal connector for thermocouples. (in An SCXI platform).