Measurement - Thermocouples


Thread Starter

George Robertson

Theory states that a dissimilar metal terminal introduced to a thermocouple circuit shouldn't pose a problem, so long as the temperature going into and out of the terminal block are the same. What practical experiences have list members had with non-T/C terminal blocks in thermocouple circuits?

[email protected]
George G. Robertson, P.E.
Manager of Engineering
Saulsbury Engineering and Construction

Leighton Toews

My personal experience with using non-tc type terminals in a TC circuit has been... problematic shall we say? I was involved with a rather in depth data acquisition project which required sampling temperatures along pultrusion dies at 6 points, a total of 5 dies, equalling 32 type J TC
channels. I was responsible for the software development in this project, not the electrical design, nevertheless I ended up having to deal with the other individuals "idea" that at the lower end of the type J thermocouples' temperature range, introducing another junction by using standard terminals SHOULD not be a problem (personally I hate that word, should, I want it removed from the english lang. "it should work" means i really have no clue). Well, as it turns out, it was a problem, a big one, oh not for him of course, to a hand held temperature monitor plugged into one of the channels the signal seemed as stable as a rock. HOWEVER, to a high speed data acquisition channel, the signal was about as messy as it could get, nearly unusable and required some serious programming magic in order to come up with something that resembled the actual temperature. Thermocouple
terminals are quite expensive, several times the cost of their standard counterparts, but let me tell you, if you DON'T get them, you will wish like hell that you did.

Leighton Toews
Process Control
Del-Tech Industries

You can't have just one "dissimilar metal" terminal - there have to be 2.

For instance, if you have a chromel-alumel Type K thermocouple, and connect to an instrument using copper, there are two possibilities:

1. There is a connection between chromel and copper, and another between alumel and copper, in which case the temperature of these connections is effectively your cold junction reference temperature.

2. You break into either the chromel or alumel leads, and have two chromel-copper or two alumel-copper junctions, in which case the original cold junction reference will stand.

In case 1, which is probably the most likely, there will be no ill-effects due to the extra
two junctions provided the two junctions are at the same temperature. This is where you need a ceramic type terminal block with good thermal conductivity between the two circuits. However, if you are connecting the other ends of the copper pair into an electronic module which is carrying out cold junction compensation based on the temperature at its terminals, there will be an error introduced. Even without thermally
connected terminals, this is likely to be the source of the biggest error.

In case 2, you will have few problems with cold junction compensation. However, there is a small chance of an error being introduced if the two connections are not at the same temperature - in a typical installation, not a very big error.

Hope this helps,