Temperature Measurements

Both methods work although some type of bonding for durability is a plus. Also, type J T/C's in particular are corrosion prone which can lead to an open circuit condition if just twisted together.

Use some good high temperature solder for the best result. As stated before, you can get away with simply twisting them together, but you are more susceptible to corrosion and loss of contact.

The best lasting results would be obtained by welding. Not sure how - maybe zap them with a spot welder? pressure (a whack with a BIG hammer)? flame? - but it would eliminate any possibility of a poor contact between the two metals.

A search on "thermocouple welder" got many hits. From very casual looking around, prices seemed around US$ 1000-2000; maybe less if used. It looked like an argon-shielded mini spot welder using capacitive discharge.

Most preferred if twisted & soldered together at end point. Just like the other guys are saying, mere twisting is prone to corrosion. As the cable insulation for two wires will be corroded, mere twisted wires contact with each other at somewhere farther up than end point(remember the reading will be at contact point). You might have an offset temperature point there(instead at end point). Thus you will be getting a lower reading than the point of you concern.

Think of this, what if, i put the two wire in your thermocouple in contact somewhere at the middle of the wire length, what will effect to the reading? Does it read still the end point of the wire or the reading of the middle.

Electrical current travels the shortest and the most accessible points, so when the wire comes in contact, the readings from you monitor will be right at the contact point.

Hi,

We built some long thermocouples for sensing skin temperature of tubes in one of our heaters.We twisted the wires then we "melted" the wires together with an oxygen /acetylene torch. They have being is service for about three years without any problems.

Thank you

Thermocouple materials whether Type J or any other should be "fused" together for greatest accuracy. Although if +/- 5 degrees is an ok tolerance then soldering or twisting will suffice but as has been mentioned, if only twisted, then corrosion will quickly turn +/- 5 into upscale burnout!! twist them together and fuse with an oxy set!!

We used to use the arc of a TIG torch to weld them together. This eliminated any worry of changing characteristics due to foreign material contamination of the dis-similar metal joint.

There is a method where a small ring is placed over the wire and squeezed. It is much like terminating a wire with a crimp connector.

All these answers are helpful. Thank you all very much.

Soldered is better than twisted, but the best junction is to solder the 2 leads, untwisted as parallel lines seperated by a small gap on a metal plate. This is referred to in the aerospace industry as the "Moffat Junction Technique" & there is also the "Moffat Thermocouple (4 wires instead of 2)", that also uses this technique. Professor Robert J. Moffat (Standford University) is considered the Grandfather of Modern Day Thermometry.

To start your understanding, acquire a copy of his paper "The Gradient Approach to Thermocouple Circuitry". Bob has developed many practical techniques throughout his research. Bob was a guest lecturer for Peter K. Stein's "Measurement Enginerring" Short Courses, "Taking Valid Data On Purpose & Being Able to Prove It!" Bob opened my eyes to thermcouples 30+ years ago & his research & conclusions have stood the test of time. Paul Morgan, Engineered Measurement Systems.