WTF? Is 3 wire RTD just an evil plot?

Factual Questions
1

Using RTD’s (resistance temperature detectors) and their upscale cousins, SPRTs (standard platinum resistance thermometers) has made me wonder if the “3 wire” method of connecting RTDs was just an evil plot devised by competitors to ruin one of the only perfect ideas in engineering.

All these devices allow measuring temperature by taking advantage of the fact that metals generally have electrical resistances roughly proportional to absolute temperature. More specifically, these devices contain a fine platinum wire or a thin platinum film, most often with a resistance of 100 ohms at the triple point temperature of water (0.01 degrees C or 273.16 K). Electrical resistance is very easy to measure accurately.

The trouble is that you only want to measure the resistance of the sensor, not including the wires that connect it to your meter. So the brilliant solution to this problem is to use one pair of wires to pass a current through the sensor, and another pair to measure the voltage across it. The sensor gets two wires connected to each end. This is almost perfect - in fact if you do filtering in the frequency domain and average measurements with both polarities, it’s about as close to perfect as you can get.

So they messed it up. By far the common standard is to only use 3 wires, and try to cancel out the unknown resistance of one of the three by being able to measure the resistance of another. This only works if all the wires are matched. To make matters worse, the common bridge design that is supposed to accomplish this is not even correct - using large values of lead wire resistance as examples demonstrates this if you calculate it out (as I did), and if you’re a dab hand at the calculus you can demonstrate the error without examples (as my friend here did).

Why the hell did they do this? I am now spending days trying to find ways around the problem, and there aren’t any pretty ones. And all it does is get rid of one of the 4 wires you really need - how much can that save? Hell, ethernet cables have all of 4 completely unused conductors in them! They don’t even get connected! If somebody wants to save some copper, let them fix that!

Does anybody have any idea who started this nonsense and why???

Better still, does anybody know where they live?

2

From what I can tell, the 3-wire RTD configuration is primarily done to reduce cost. But like you, I don’t like 3-wire RTDs; while it is true that instrumenting a 3-wire RTD to a Wheatstone Bridge (or constant current source) can compensate for lead resistance, it only “works” if the leads are matched.

Are you sure about this? I operate a temperature metrology lab where I work, and I’d say 90% of the PRTs I work with are 4-wire. I have also purchased new 4-wire PRTs for $70 (not including calibration). Why even use 3-wire PRTs?

3

Couple more things:

In a 3-wire RTD, it is only necessary for the two current-carrying wires to be matched assuming it is connected to a Weatstone Bridge or constant current source that is properly configured to condition a 3-wire RTD. (The resistance of the third wire has negligible current on it due to the meter’s high input impedance, and is thus not that important.)

Minor nitpick: The (nominal) resistance of an RTD is specified at 0 °C, not the TP of water (0.01 °C).

4

Crafter_Man,
Again, you have the goods!

Here’s why I think 3-wire RTDs are taking over the world: They are all over the place here, where I get to figure out how the machinery doesn’t work. In many of the catalogs 3-wire is the standard configuration and others are special, or at least get listed later and in fewer versions. And one of the electricians tells me he doesn’t even think you can buy 4-wire input types for the controller cards. Plus, some of the cards deal with 4-wire by tying two of the wires together at the terminal block and then treating it as a 3-wire (which of course is even worse, because it’s now incorrect).

Also, I buy some of my equipment from Omega, because buying temperature equipment from their catalog is like buying your food at the supermarket. It is never the best, and hardly any of it got made there, but nobody else has it all under one roof. Anyway, most of their meters and controllers don’t accept 4-wire, and many of the sensors aren’t available that way.

Don’t know what people have against it, but it sure doesn’t seem popular.

With laboratory PRTs, certainly, 4-wire seems to be the rule rather than the exception. And with good reason, what with it being correct and all.
>In a 3-wire RTD, it is only necessary for the two current-carrying wires to be matched…
True, but in practice there are two black leads and one red lead, and what’s necessary is that whichever black lead happens to be connected to carry the current matches the red lead. Since the black ones aren’t distinguished from each other in manufacturing, they had all better match for things to work.

>…assuming it is connected to a Weatstone Bridge or constant current source that is properly configured to condition a 3-wire RTD.
Now, this is interesting. I tried to work through the diagrams I could find of how this is done, and have convinced myself that the diagrams are incorrect. These bridges tend to reduce the importance of lead resistance but don’t exactly cancel it out. A couple of us got interested and tried several ways to confirm this and every calculation and reasoning we could come up with convinced us it doesn’t work. Maybe I’ll find an example on the web and work it out here.
>The (nominal) resistance of an RTD is specified at 0 °C, not the TP of water (0.01 °C).
Well, say, you’re right… I was thinking of RTPW, but things they call RTD’s seem to refer to 0C. Learn something new every day.

5

Forgot this one:

>Why even use 3-wire PRTs?

For the same reason I use Windows: because everybody else does in spite of the obvious disadvantages right down to the miserable rotten core. And I have to suspect it has something to do with marketing…