I am doing some temperature measurements for reliability testing, over an expected range of 16C to 45C. I have an NIST traceable thermocouple thermometer (S220-T8), however it does not come with thermocouple probes.

Without a calibrated thermal reference, how can I be sure that K-type thermocouples purchased from Amazon are accurate? Shouldn't thermocouples be subject to unit-to-unit variability and thus require some sort of voltage-to-temperature characteristic that is reported? I am finding that sometimes even "name-brand" companies are not providing such a datasheet. I understand that they work based on something called the Seebeck effect, and maybe the Voltage/Temp relation is intrinsic to the metal alloy used. As long as the little pellet of metal is "K-type" and is in thermal equilibrium with the thing being measured then maybe that is good enough?

I have found an NIST traceable thermocouple probe for $100. I assume this probe will be within some spec or that they will provide the voltage to temperature relation. Is it advisable to purchase one of these NIST traceable probes and then use it to check the error of the remaining probes?

  • $\begingroup$ Is that 100$ really a dealbreaker? $\endgroup$ – joojaa Feb 17 at 21:35
  • 2
    $\begingroup$ Yes but how much does the lost time cost you if they dont work? $\endgroup$ – joojaa Feb 17 at 21:37
  • 4
    $\begingroup$ you can check it with an ice bath and boiling water (adjusting for barometric pressure and make sure water is pure). Hardly NIST traceable, but if you are on a shoestring budget . . . $\endgroup$ – Pete W Feb 18 at 0:15
  • 3
    $\begingroup$ Buy 1 traceable one with a curve and compare the cheap ones against that. $\endgroup$ – Solar Mike Feb 18 at 2:56
  • 1
    $\begingroup$ Also, read here about thermocouple measurements are read and compensated via cold junction maximintegrated.com/en/design/technical-documents/app-notes/4/… $\endgroup$ – GisMofx Feb 21 at 19:51

Go to Omega.com to learn all you want to know about thermocouples (TC). I'll give you here a few basics, directed at your questions.

TC science is well understood and standardardized. Every type TC is made with it's own precise metal alloy and all TC of the same type will output a voltage uniquely related to the temperature of the TC junction, within accepted accuracy. Review "The Basic Laws of Thermocouples." The insulators on each TC type are color coded and each Type have their appropriate accuracy within appropriate temperature ranges. Omega used to provide a table of Temperature/Voltage data increments in their hardcopy catalogues, so my guess is that they have this data on their website. You can use such a chart to calibrate your TC, but you'd need an accurate Temperature measuring device and an accurate voltmeter, and it's really not worth the bother.

The TC industry is extremely well controlled, and so it's highly unlikely that any TC wire you buy will be "inaccurate." So if you buy only the wire itself, no need to worry. TC wire is very cheap, coming in all Types and diameters. It's very easy to make a TC junction from the wire itself. All you have to do is fuse the two separate wires (one + the other -) together by a hot acetylene torch, or by an electronic "TC maker" that uses a discharge from a capacitor. This is essentially fool proof. If the wires are fused, the TC is a TC. You can also imbed (drill a hole and "peen" it in) one wire in say the metal piece you want to measure the temperature of and the other wire a short distance away, and the reading will give you the average temperature of the space between. The TC Law that assures this is the Law of a Common Junction.

You mention "probes," but you didn't say what kind of probe geometry you're using. In general, TC wires are often mounted inside probes, and there, the manufacturing quality is important. There are "surface probes" to measure the temperature of flat surfaces, "point probes" that are used to stick inside things, and others with variously different probe materials suitable for different environments.

TC don't age, unless the probes become corroded or broken, but the basic TC itself never wears out or loses calibration, as long as the TC welded junction is intact.

That's about all I can do for you here, but please go to Omega.com and learn more.

  • 1
    $\begingroup$ I would just buy my thermocouples from Omega too. Great company. $\endgroup$ – Eric S Feb 19 at 2:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.