Question about thermocouples...?

Question

I have a system set up where I am using Type K thermocouples inside a vaccuum chamber. The thermocouples are soldered to a connector that goes through the wall of the chamber, and then a normal wire (aluminum I think) carries the signal from there to my DAQ system.

The problem is that the thermocouples no longer read the correct temperature. For room temperature they should be reading 25-26 C, but they are reading 34-36 C. I have a feeling that it may be because I am using the aluminum wires that are not the same as the chromel and alumel wires in the thermocouples, thus creating two more junctions. I'm not sure if this is important or not, but it's the only thing I can think of. Anybody know what my problem is?

Answer 1

Some pretty good answers above. Omega does have good info and TC hardware. But...

Usually type K is intended for high temps. Is this your case? If so, a 10 deg. C error is not considered to be significant in many setups.

A related question is: Does your DAQ system have built-in cold junction compensation? Its pretty common to just run the TC wires into a scanning voltmeter or A/D. If you've got the latter (no CJC), then you should either give up on your error or use a real ice/water bath cold junction. The Keithley TC DAQ has fabulous cold junction compensation hardware.

Let's say you've got got good CJC. Yes, it's true that you should ideally have 100% continuity of the chromel wire and the alumel wire from the measurement junction to the DAQ. But with care and knowledge you can do very well without that.

For example, a vacuum feedthru having short and thick metal conductors will usually be OK regardless of metal type or solder joints. The dissimilar metals don't matter if there is no temperature difference between ending point of the one chromel wire and the starting point of the chromel wire on the other side of the feedthru. E.g. a short & thick copper feedthru conductor conducts heat really well and will not support a temp difference.

Another question: Who uses bare aluminum wire? That's pretty weird. Are you sure it is aluminum? Try replacing it with the correct TC wire.

My best guess is that you do not have CJC and you are measuring the temperature variation of your DAQ. It is not OK to calibrate your system by subtracting off a temperature offset, if a variable DAQ temperature is the problem.

If this is a serious issue, you could hire me to consult.
831-643-9049

Answer 2

Yes, by adding a different kind of wire, you have created at least two additional thermocouples. When reading high temps (like 1000F), it doesn't make a lot of difference if the extra junctions are at room temperature. But still there will an additional error introduced.

But with the low temps you are reading, it is even more important to eliminate those errors. You need type K wire all the way back to your DAQ.

I would avoid solder. That introduces another metal. Use Type K screw terminal junctions.

To a first approximation, the error is a constant. So if you subtract 10C, you'll be about right. If the measured temp goes up 5C, the reading will go up about 5C.

http://www.omega.com/ has everything you'd need plus some background info on sensors and measurement techniques.

An aside: Sometimes a old mercury thermometer has its advantages!

Answer 3

When using thermocouples it is often important to control the reference junction temperature as well as the junction sensing the temperature. Sometimes three-wire lead systems (and even four-wire lead systems) are needed to avoid instrument errors caused by temperature changes of the leads (inside and outside enclosures). Instrumentation specialists know the specific techniques. With TC' s it is usually best to avoid using wires and junctions different from the TC's two correct metals. Extra junctions or even reversing the correct wire materials along the TC circuit may cause problems too. .

Answer 4

Sound's Like You Are Turning Gas Valve Off Instead Of On. If Pilot Stays On After You Light It. It's Not The Thermocouple.

Answer 5

Your feeling is correct. You should have thermocouple wire all the way back to your DAQ system. Otherwise your reading will never be right.

HTH

Answer 6

I cannot speak to the connection or if it causing a misread. If you are getting incorrect results it could be from the data acquisition device. Are you calibrated correctly. If you get consistent results you should be able to perform a simple calculation to correct the reading. If you are using something like LabView this should be easy to do.

Answer 7

The previous answer summarizes what I was going to write, including the web site of Omega.