How can I calibrate a thermocouple?

Question

I'm trying to calibrate a K-type thermocouple, but I need something with a reference temperature that I can trust to within .1 C. Is ice water trustworthy?

Answer 1

The triple-point-of-water is one of the fixed points in temperature calibration. The typical thermocouple has a specification of about ± 2.0 °C over the entire operating span and calibration uncertainties of about 0.25 °C. These uncertainties are usually only achieved in a calibration laboratory. The TPW cell is a laboratory standard and properly set up produces 0.01 °C to within 0.0002 °C and sometimes a bit better.

If you are trying to calibrate a probe/meter combination a good ice bath should get you in the ball park (± 0.5 °C) and boiling water accounting for your elevation, should give you a second point to validate the linearity of your meter/probe combination. The dominant uncertainty in your calibration will likely be the resolution of your meter. To get uncertainties of near 0.1 °C at the time of use, you really need to go with a PRT.

If you are trying to calibrate the thermocouple only, you need to connect the leads to copper wires and submerge this junction in an ice bath and apply known temperatures to the thermocouple bead or probe. The output will be in microvolts and can be checked for compliance to the NBS/NIST tables using a good nanovoltmeter or a differential measurement scheme and a calibrator.

There's lots of stuff about temperature calibration on the Omega Engineering website. Yellow Springs Instruments also used to have information. You may also check out Hart Scientific (Fluke Corporation).

Good Luck! Oh... by the way a TPW cell goes for about $750 (US).

Answer 2

I think you can make what's called a triple point cell. This will cause water to exist in its solid, liquid and vapor forms all at once. That is what the NBS used to use, and it is probably good enough for whatever purposes you want. I don't recall the exact temperature at which it maintains... I think it was something like 4 degrees Centigrade. And you really need two temperatures between which to draw a scale. Boiling water at seal level would do fine for the other.

Answer 3

I'm not sure ice water would not be a good reference as it may not be exactly zero degrees Celsius.

Boiling distilled water, and knowing your altitude above sea level may be a better standard. I'm not sure of the exact formula.

Answer 4

Ice water is about as good as you can get without calling some cal company.

Answer 5

Yes. Ice water is fine.

Answer 6

I would use a fever thermometer as a control.