Thermocouple question

[Broken_Arrow]

If ten (10) J-type thermocouples (same type/manufacturer) are hooked to a SCXI-1102B, can they be expected to read the same within 0.1 degree C? I'm not talking accuracy, I'm talking about the ability to track each other within a tenth. Thanks.

[Henrik_Volkers]

Based on the principle of thermocouples your cold junction board then also needs to be stable and of even temperature within 0.1 (or less) degrees.

And you need to look into the spec of your card, if you have that resolution with your ADC. Thermocouple EMFs are quite small 😉  0.1K for a Type J is about 5µV change

The usual way to track small changes is to bring your sensors to an equal temperature and record and null out the offsets.

Depending on your TCs lengh and their (constant or changing) thermogradients along their way, you may be able to measure 0.1K  . 

In a lab I have seen setups to measure mK changes(!) with TCs but thats involved a lot more than just a ADC card and some TC wire......

[AEI_JR]

Richard,

What connector block are you using?  I generally use the TC-2095 block for TCs into the 1102.  I've found that when accuracy is needed, using the box and shipping foam the 2095 is shipped in, modified to allow the SH9696 cable in through the back, works very well.  Also, ensure that you have the correct resistor networks and DIP switch settings.  If using the SCXI-1303, insulating is more difficult, but generally less necessary.  You still should be sure to use the correct resistor networks with the 1303.

[johnsold]

From an OLD Omega catalog the limits of error for a Type J thermocouple are +/-2.2 degrees C or 0.75%, whichever is greater, for standard grade and about half of that for Special grade.  That is based on a 0 degree C reference junction temperature and no other errors, such as installation errors or voltage measurement errors.

 Lynn 

[Broken_Arrow]

Thanks for the replies! That is a very good point about the connector block needing to be stable. The ADC has plenty good specs, but if anything alse is drifting, then all bets are off.

I think zeroing the readings at startup, every day, with TC's on a known stable thermal mass (as in something heavy) is the way to go. I've had excellent results in the past with RTD's, but from my experience, TC's aren't as impressive.

Thanks again!

Message Edited by Broken Arrow on 09-17-2008 11:05 AM

[AEI_JR]

You're right about the instability of TCs vs. RTDs.  However, TCs are much less expensive, more rugged, easier to install, etc.

Types E, S, R, and T thermocouples can be obtained with better accuracy than Type J (though T is limited to 350 C).

[Henrik_Volkers]

A good source for TC knowledge:
Manual on the use of thermocouples in temperature measurement,
ASTM PCN: 28-012093-40,
ISBN 0-8031-1466-4

(Page1): 'Regardless of how many facts are presented herein and regardless of the percentage retained,

                all will be for naught unless one simple important fact is kept firmly in mind.

                The thermocouple reports only what it "feels." This may or may not the temperature of interest'

[Broken_Arrow]

Amazing quote. With that in mind, I doubt you could ever "calibrate out" the errors in multiple TC's and have them track dead-on for any length of time.

It has been my observation that TC's are much more "dynamic" than RTD's. Whether these dynamics are reflecting reality is up to the user to decide. A bi-metalic sensor is just more complex, in its elemental form, than a resistor (RTD).

Message Edited by Broken Arrow on 09-18-2008 08:19 AM

[AEI_JR]

The same could be said of pressure probes, or accelerometers, or strain gages . . .  The installer, or the one who specifies the install method, must intrinsically understand the underlying principles of the sensor's abilities and limitations, otherwise the data retrieved may turn out to be useless for the requirement.