Sciencemadness Discussion Board

How to repair a thermocouple?

Pumukli - 6-1-2019 at 10:38

I have a thermocouple at the end of a meter long wire-pair, (thermometer sensor.) Unfortunately one of the wires broke at the small junction globule. I'd like to repair it but don't know how to proceed.
I thought of discharging a bigger electrolytic capacitor into this junction and make it remelt again. (But how much joules?)

I know the couple is cheap (USD 1, including shipping and handling, maybe), but I don't want to bin it yet.

Had anyone repaird a thing like this?

Sigmatropic - 6-1-2019 at 10:40

Sand the wires and wrap 'm together. Worked fine in my case.

happyfooddance - 6-1-2019 at 10:48

Yes, I did it similar to how you described, I welded it by discharging a capacitor from what I believe was a 12v power supply. I'm not sure the amount of energy discharged, but I varied it (by using a resistor), and it took a few (maybe 2 or 3) tries before I got a clean weld. I cleaned the pieces between tries and added a small amount of flux (probably a mix of dilute HCl, Zinc Chloride, and borax or tri-sodium phosphate).

There is a certain trick to getting the arc to form in the right spot, I'm not a welder or an expert, but for me the ends needed to be more parallel and less tip-to-tip, if that makes sense.

Fulmen - 6-1-2019 at 10:57

If it's used for low temperature you can solder it.

morganbw - 6-1-2019 at 11:11

Quote: Originally posted by Sigmatropic  
Sand the wires and wrap 'm together. Worked fine in my case.


Not the correct way but just twisting the wires at the junction often work remarkably well.
Just saying.

[Edited on 1/6/2019 by morganbw]

Twospoons - 6-1-2019 at 11:35

I've used the arc from a neon sign transformer to make thermocouples. Melts a nice little ball on the end of the wire.
As has been said, soldering works too, though you need an aggressive flux such as ZnCl / HCl.

Fulmen - 6-1-2019 at 13:02

In my experience twisting and soldering works fine without any special flux. At least for "single use" or non-critical applications. We routinely measured heat production in curing concrete, the TC was twisted&soldered with regular 50/50 flux core and embedded un-shielded into the wet concrete. This worked fine for at least a week.

Twospoons - 6-1-2019 at 15:16

Your 50/50 solder probably does have an aggressive flux, given the usual application for that alloy is soldering pipes etc. This is different to 63/37 Sn/Pb used in electronics, which usually has a much milder flux, and in my experience doesn't work as well with nickel alloys.

Sulaiman - 6-1-2019 at 15:37

For temporary use twisting wires together works ok,
type J used in a lot of older equipment can be soldered,
I've never succesfully soldered type K

Ubya - 6-1-2019 at 16:52

i don't think soldering would be a good idea, i mean a thermocouple works because when to different metals are in contact and are heated they generate a small electric potential, if you solder them with a random filler metal, the propriety of the probe would change right?

Fulmen - 6-1-2019 at 18:02

My bad, I meant the standard fine electronics solder. I suspect it acted more like an adhesive, and that the electrical contact was formed by twisting the pair. Either way it performed flawlessly, even when suspended in concrete.

As for the TC a solder joint shouldn't affect the performance. The chromel-solder / solder-alumel dual couple would produce the same voltage difference as the standard chromel-alumel couple.

DavidJR - 6-1-2019 at 19:00

Solder may work but it most certainly will change the temperature/voltage curve of the thermocouple. It may, however, still operate acceptably well, but no guarantees.

unionised - 7-1-2019 at 02:20

Quote: Originally posted by Ubya  
i don't think soldering would be a good idea, i mean a thermocouple works because when to different metals are in contact and are heated they generate a small electric potential, if you solder them with a random filler metal, the propriety of the probe would change right?

Odd as it seems, as long as all the soldered bits are at the same temperature, the solder does not affect the output voltage.
So, chromel, solder, alumel gives the same voltage as chromel, alumel. (provided that the alumel solder and solder chromel junctions are at the same temperature).

yobbo II - 7-1-2019 at 15:33


The soldering will work as long as you are measuring lowish temps. If you go near the melting temp. of the solder the junction will dissolve into the solder and things will not work. Brazing will work so long as you do not go near the melting temp. of the braze. If you go near the ....... (as with the solder above!)

Try to melt wires together. Not easy to do.
Yob