PID Controller - MEW 269 - wrong connector

It would be a clever controller that could compensate for errors it did not know about

If the error is created by the use of incorrect connector/s then the error will vary with the temperature gradient across those connectors, which are probably going to be at different temperature to the one doing the measuring.

Not particularly relevant to this particular thread, but an important point to bear in mind when controlling of temperature (with whatever sensor) is that the only place in the whole system where the temperature is actually accurately known, is at the actual sensor itself.

Ian P

.

In the real world; I think it would be safer to turn that statement around, and say that the instrument needs to be connected in the way it was designed to be.

MichaelG.

It has been too many decades since I last used a thermocouple for me to remember how we wired them up. So I had to go back to basics (Wikipedia) to find out. What seems to have been missed in much of the above is that a thermocouple gives the temperature difference between its hot and cold junctions, not the hot temperature. The cold junction is typically within the instrumentation: the instrument measures the cold junction temperature by a different means (cold junction compensation) and uses this, with the thermocouple voltage, to calculate temperature. Hence the need to use “thermocouple material” all the way to the cold junction. After this, copper is fine.

Room for one more angel on this thermocouple?

Good to have your input, Dr Dave ... But I like to think I made that reasonably clear in my reply to XD 351, on p1.

MichaelG.

True, but I didn’t recall anyone pointing out the importance of measuring the cold junction temperature. Too many post to remember the details?

Not so much as measuring cj temp per se, but I did mention cjc when calibrating, some more advanced tc simulators have this incorporated saves having to take it into account when inputting the relevant millivolt input to the 'kit' being calibrated......Although we used to have a seperate thermometer measuring ambient temp when we calibrated transmitters/trip amps etc....

Hi, I would say that all temperatures in the article are wrong . 12600 Celcius, seriously ?

Actually the 'cold junction' is almost a myth.

usually it's drawn as the inner part of the special connector.

If you connect the two wires at both ends, you won't have any voltage to measure...

The 'cold junction' is always going to include the measuring circuit which is probably going to include copper, tin and lead if old style solder), doped silicon and who knows what else.

This is from Maxim Integrated:

Somehow the makeup of the special connector is critical, but you can use a copper wires to take it to your carefully crafted differential amp!

What's magical about calling some point INSIDE the special connector the cold junction, instead of OUTSIDE it?

Lot's of smoke and mirrors around these precious connectors! As long as your connections are symmetrical in terms of materials and their temperatures are close (I'd love to see a 10 degree gradient sustained across an XLR connector that's carrying less than a milliamp of current...)

Neil

There is nothing magic about the cold junction or where it is physically located: it is at the transition from “thermocouple material” to copper. So long as you measure the temperature of the cold junction, all is well. The cold junction is inside something like a handheld instrument so that cold junction temperature measurement can be integrated in the instrument. I suspect for Model Engineering projects, changing to copper in a Molex connector and feeding into the instrumentation via XLR plug will give acceptable results, even using ambient temperature as the cold junction temperature. The Wikipedia entry on thermocouples (or that referenced byMichael Gillian on p1) gives a good overview.

Ducks and awaits incoming criticism!

.

Perhaps this is a more helpful diagram:

**LINK**

http://enacademic.com/pictures/enwiki/84/Thermocouple_circuit.svg

MichaelG.

.

We are, I'm sure, in agreement !!

MichaelG.

Edited By Michael Gilligan on 18/06/2018 23:00:55

Indeed, it proves my point exactly.

Neil

.

Then could you please explain your point [either here or privately], because we appear to be having different conversations.

MichaelG.

.

Edit: I am particularly keen to understand why you made this obvious but irrelevant statement:

"If you connect the two wires at both ends, you won't have any voltage to measure..."

Edited By Michael Gilligan on 19/06/2018 09:08:17

As we are now well into p3 and still struggling ... I would just like to re-state the issue:

• The Schematic Fig.2 [not Fig.1] shows the Thermocouple correctly connected to terminals 9 & 10 of the PID Controller.
• The controller will treat those terminals as the reference junction[s], and will presumably have a local correction device to normalise it to a 'cold junction'. .... Please see the Wikipedia page for a picture of how this is done on a Fluke meter.
• Chris Gabel has, for convenience, added an extension lead of 'thin gauge stranded wire' and an XLR connector, which moves the actual cold reference junction 'outside' his ABS box.
• The temperature at the controller's 'cold junction' correction point is therefore not tied to that of the actual cold junction.
• If the extension lead had simply been made of thermocouple extension wire and a matched plug and socket, this invitation to error would not exist.

MichaelG.

.

https://en.wikipedia.org/wiki/Thermocouple#/media/File:Cold_Junction_Compensation_with_Thermistor_to_measure_the_junction_temperature..jpg

Edited By Michael Gilligan on 19/06/2018 09:47:53

Apologies if people are bored with this discussion:

This thread has just dropped off the first page, so I'm bumping it in the hope that Neil will see it again, and respond.

[ I simply don't understand your point, Neil; and that leaves me uncomfortable ]

MichaelG.

Reductio ad absurdum refutation of the idea that you have to use the thermocouple materials the whole way. The volatge has to be measured across the inputs of a high impedance amplifier which means the 'loop' has to involve non-thermocouple materials which will introduce errors in even the most perfect of setups.

The diagram you linked to makes it clear that what is usually lazily called the 'cold junction' is actually TWO cold junctions, at the points where the thermocouple materials transition to normal materials at the cold temperature.

In that case they are shown as being held at zero degrees by melting ice, but for ordinary purposes simply having the two junctions close to the (known) ambient temperature will be perfectly adequate.

Which is then point I have been making - the thermocouple wires only need to be consistent materials until they reach the 'cold junction' temperature and in the case of a furnace controller which is unlikely to use an ice bath the interposition of an XLR connector in unlikely to make any meaningful difference.

Sorry to labour the point, but to me it is very clear.

Neil

True you dont have to use thermocouple materials 'the whole way' but then the question is, what is the whole way, or in other words, where does the whole way end?

As I see it the thermocouple materials matter all the way to the terminal block because after that point the signal can be treated as pure voltage. Its a very low voltage so good circuit design has to be applied to ensure the accurate measurement signals of only a few millivolts. One significant thing about this final connector is that both its terminals are at identical temperature so any error introduced by the transition to copper is cancelled out.

I imagine that in the instrument engineering discipline the 'loop' only covers everything that connected to the input terminals of the controller or indicator, the internal circuitry is not part of the loop.

Ian P

.

Thanks for replying, Neil

I am now happy that the problem is as I suspected : We were having two different conversations!

The real objection to the XLR connector and its lead is as I [hopefully] clarified in my post of 09:45:02 yesterday. ... It inevitably moves the cold junction to a point which is not monitored/normalised by the instrument.

MichaelG.