Damaged thermocouple wires: Can they simply be soldered back together?

[Jeff7]

The bead at the end is just fine, but the wires themselves have been damaged.

A search match on Google says that thermocouple materials may not be solderable, that a metallurgical bond will not be created. That would leave twisting the wires + heatshrink, or else some kind of butt-splice connector.
Concerns:
Twist + Heatshrink = no "outsider" metals, but the connection isn't too secure.

Splice: The crimp-on splice will likely be made out of plated copper, so there would be two dissimilar metals touching, which may or may not be a problem for a thermocouple.


Halp!

 

[dullard]

Having two dissimilar metals touching is not usually a problem for a thermocouple. Imagine a solder on a type K thermocouple (on the chromel leg). You'll get this at the solder location:

chromel--solder--chromel

At the first chromel--solder connection, you'll create a voltage that depends on the temperature of that first junction. At the second solder--chromel connection, you'll create an opposite voltage that dpends on the temperature of that second junction. In almost all cases that I could imagine, the solder joint is very small and made of thermally conductive materials. Thus, the chromel--solder connection is at the same temperature as the solder--chromel connection. If that happens, then the voltage created by the first connection will be the same magnitude but opposite sign as the second solder--chromel connection. The two voltages will cancel perfectly.

The only real issue is if you somehow make the solder connection be large and put it right in a location with a large thermal gradient. Then it'll cause problems. But, putting your wire legs in a temperature gradient is bad practice to begin with - solder or not.

That said, why not just use a thermocouple connector? Link (I linked the expensive large ones for high temp there are other options). They are a few dollars each (get a male and a female), and easy to use. I've used those hundreds of times with no problems at all.

I do wonder about your damaged thermocouple wire though. Thermocouples are so finicky that every small bend in the lead wire will successively degrade the calibration. If you put enough bends in it (or damage it through another method sufficiently harmful enough that the wire breaks), it may still work, but your calibration may now be far off.

 

[bobdole369]

That depends on the thermocouple! High temp ones might need special wiring (we use platinum wires for exhaust temp sensing thermocouples). If its the one on your meter, or a simple K-type I think it would be OK.

 

[Jeff7]

These are going in an environmental test chamber. Two thermocouples measure the temp of some lines in the compressor system. The other three measure various locations within the chamber.

We don't need extreme accuracy here either - they probably just need "cold, cool, warm, and hot."

I'll look into the thermocouple connector, but they want it working soon - and cheap. This is yet another thing that was bought semi-functional at an auction. Most such purchases are under $500.

 

[LordSnailz]

I'm curious say you cold wire them back together, don't you have to somehow calibrate them again?

 

[Rubycon]

You can try it but be prepared to replace the probe completely as if the readings cannot be trusted why measure in the first place?

 

[Jeff7]

Originally posted by: LordSnailz
I'm curious say you cold wire them back together, don't you have to somehow calibrate them again?

I'd imagine that they "calibrate" themselves by nature of their chemistry.

Twist + heatshrink seems to be working; and what Dullard suggested with soldering makes sense too.


These TCs appear to use copper and constantan wires.



More about these thermocouples: The ends are not bonded together, at least not by a weld. it's just a crimp-on connector of some sort which provides the electrical connection. Well one of these crimp-on connectors fell off - the wires had been stressed too much evidently, and it just fell off when it jiggled during removal. Incidentally, just twisting the wires together does generate a reasonably accurate signal.

Originally posted by: Rubycon
You can try it but be prepared to replace the probe completely as if the readings cannot be trusted why measure in the first place?

In this case, ±10°C would be quite acceptable.
Hell, at this point, just getting the chamber to power on and stay on would be nice. It clicks on, a transformer powers up, and then it reports some kind of thermal overload error, indicating that the chamber is somewhere over 250°C. Well, it's just about room temperature, and this isn't the surface of Mercury. So something is amiss somewhere.
As far as I can tell, this specific overload trigger is controlled by a thermal switch/fuse, which is marked:
Microtemp
KGAAST
G4A01
TF 240C

It measures as a short circuit, which makes sense - you'd want a sensor to fail open, as you wouldn't want a loss of connection to say to the controller, "Everything's ok! Let's cook some shit!!!"

We've got documentation for the controller itself, but not the chamber. Go figure.

The supporting software comes on a 5.25" floppy.

 

[Rubycon]

Originally posted by: Jeff7

As far as I can tell, this specific overload trigger is controlled by a thermal switch/fuse, which is marked:
Microtemp
KGAAST
G4A01
TF 240C

It measures as a short circuit, which makes sense - you'd want a sensor to fail open, as you wouldn't want a loss of connection to say to the controller,

Microtemps are one shot high limit thermal fuses. A healthy one is supposed to measure 0 ohms.