Thermocouple aging

[marekpasek]

Hi,
Let´s say, I want to buy some obsolete multimeter from ebay. One of the functions of the multimeter is temperature measurement. Thermocouple is included. And there my question comes.
Does the thermocouple age (or rather may its voltage potencial ,caused by temperature, change by the time)?
Since the thermocouple itself consists of two metal touching each other, shouldn´t those metals be fusing to each other by the time then creating some alloy? If so shouldn´t this newly created bimetal alloy have different response to temperature and then make the reading unprecise or even invalid?
So eventually wouldn´t be the best to buy a new thermocouple probe rather than using some decade or so old one?
If so how big the drift is? (I am just curious and I was not capable of finding any datasheet which could provide these informations)

Thanks

[DTJ]

A good thermocouple junction is usually created by arc fusing the dissimilar metals together to form the junction.
I'm not sure a protected junction would age at all, one that was exposed to a chemical process or high temperature exhaust stream may end up with the junction contaminated which could perhaps cause an error.


I'm not sure about the aging but unless end to end calibrated the thermocouple error is likely to outweigh any aging error.

[wraper]

Metals are not touching each other, they are molten together, it's not the junction itself which is producing voltage. It does not even matter if they are molten together, basically you could even attach 2 wires with a screw terminal. Voltage is produced along the wire because of temperature difference along it. If 2 wires are made from the same metal, returned voltage will be 0 as voltages will compensate each other. If dissimilar metals are used, voltages produced along the wires will be different and the difference between them can be measured on the terminals of thermocouple.

[Kleinstein]

The typical type K thermo-elements don't show significant aging at low temperatures. It is only at higher temperatures like above 400 C or 500 C that aging can become an issue.

If you are really after mKs, bending (beyond the elastic limit) of the wire can be more of a problem. So one should handle with wires with some care. 

The junction itself is no problem. As long as the junction part is at a constant temperature it does not change the EMF even if you add a 3rd metal in between, like using solder instead of welded junctions.

[jmelson]

Hi,
Let´s say, I want to buy some obsolete multimeter from ebay. One of the functions of the multimeter is temperature measurement. Thermocouple is included. And there my question comes.
Does the thermocouple age (or rather may its voltage potencial ,caused by temperature, change by the time)?

Thermocouples in high temperature ovens DO age, but it usually takes thousands of hours, at 1000 C or so.  But, just sitting in a box?  No way!

Jon

[jmelson]

Metals are not touching each other, they are molten together, it's not the junction itself which is producing voltage. It does not even matter if they are molten together, basically you could even attach 2 wires with a screw terminal. Voltage is produced along the wire because of temperature difference along it. If 2 wires are made from the same metal, returned voltage will be 0 as voltages will compensate each other. If dissimilar metals are used, voltages produced along the wires will be different and the difference between them can be measured on the terminals of thermocouple.

WOW!  VERY FEW people actually understand how thermocouples work!  There are actually two effects at work, both a voltage at the junction and a voltage along the temperature gradient in the wires.  But, the above is basically correct - with most thermocouple junctions, the gradient along the wire is dominant.

Jon

[nfmax]

Labfacility Ltd publish a useful 'Handbook of Temperature Measurement' with information about thermocouples (and much more). You can order a physical copy, or download it from https://www.labfacility.com/downloads.

[macboy]

I'll add that if your thermocouple is looking particularly nasty or suspect, you can just snip the end off, strip a bit of the wires and twist them tightly together. Optionally spot weld (do not solder).

[Jeff1946]

Another low tech way to make a thermocouple junction is to twist the ends together then smash them with a hammer against a hard surface.  Not a rugged as welding but good enough.  At work we would weld thermocouple ends together with a rich hydrogen-oxygen torch.  The extra hydrogen gives a reducing atmosphere to prevent oxidation.

[marekpasek]

Well that is really helping.
You know I wonder that I want to find datasheet of thermocouple so I typed "thermocouple datasheet" to google and the only thing I have found were some thermocouple IC decoder datasheets.
So I sincerely thank you for suggestion

[retrolefty]

By the time I retired in 2007 our refinery had decided that RTD sensors were to replace all TC sensors in new construction (assuming maximum operating temp is supported by RTD). The higher labor cost Vs sensor costs had made the higher cost RTD sensors less a factor with the much greater accuracy benefits of RTD usage, up to about 1,200 degrees F.  TC sensors are cheaper and more rugged but we found the +/- 5 F accuracy in real practice rather dated in this age.   

 

[retrolefty]

Well that is really helping.
You know I wonder that I want to find datasheet of thermocouple so I typed "thermocouple datasheet" to google and the only thing I have found were some thermocouple IC decoder datasheets.
So I sincerely thank you for suggestion

 Omega engineering is a good source of technical information on thermocouples as well as anything else
involved in process control engineering.

Here is a good TC tutorial: https://www.omega.com/prodinfo/thermocouples.html