I wish I could solder thermocouple wire and extension wires on RTDs as easily as I can other electronic leads and copper wire. It’d be especially nice if the joints were good to, say, 420 C. But to only be able to use these things at 200 or 300 would still be worthwhile.
Could I do these things with some kind of silver-bearing solder and ideally a special high-temperature iron? Or a tiny torch if need be? Or is the issue having the right flux?
I’m interested in thermocouple types K, E and N. Those alloys are various combinations of nickel, copper, chromium, manganese, silicon, and magnesium.
As for the RTDs, while the active element is a fine platinum wire, I don’t know what the leads are (and there are a variety of element designs so maybe the leads are of various alloys). The RTD extension wires I know of are nickel plated copper (yes, nickel - not tin).
I do know that ordinary rosin core lead-tin electronics solder doesn’t wet the RTDs, RTD extension wires, and thermocouples I have tried.
I’m suspect that I’m just repeating what you already know, but since thermocouples are based on dissimilar metals, the metals used for any extensions have to match the thermocouple wires. I think that Omega specs mechanical connections to extension wire.
No, not really. For “quick and dirty” setups I’ve been known to use plain ‘ol copper for extension leads. It’s certainly not ideal. But it will work as long as the junctions are at (more-or-less) thermal equilibrium.
Thermocouple extension cable is made from “extension grade” wire as opposed to “thermocouple grade” wire. The Seebeck coefficient for extension grade wire matches the published value at room temperature. Which is all you need for extension cable.
You could silver-solder the connections. But keep in mind that flux must be used, and the flux might (will?) contaminate the junction.
The following is one description of how (I believe) a professional manufacturer typically makes a general-purpose PRT probe. I understand you’re not a PRT manufacturer, but perhaps you can glean some useful knowledge from this info anyway:
The PRT element is a 2-conductor device. Four nickel wires are connected to the PRT element; two nickel wires are welded to one PRT lead, and two nickel wires are welded to the other PRT lead. (Brazing could also be used in lieu of welding, but you run into the same problem as soldering – flux contamination.) Each nickel wire is insulated with fiberglass.
The four nickel wires can directly terminate into a 4-pin connector, or they can be directly hardwired into a signal conditioner. But nickel wire is rather expensive, so if you have a long run (more than a foot or so) it makes sense to transition from nickel to silver-plated copper. The latter is insulated with TFE.
Another method is to use nickel-clad copper wire. Because it’s fairly cheap, you can connect four nickel-clad copper wires directly to the PRT element and make the wires as long as you want. The neat thing is that there’s only one transition. But I believe there are drawbacks to this approach in terms of temperature rating and stability over time. The “Cadillac” approach is the nickel -> silver-plated copper scheme described in the previous two paragraphs. (Well, almost. The “ultimate” approach is to run platinum wires all the way!)
As far as thermocouple wire goes – same thing. You have no idea how much the flux from the soldering will contaminate the junctions. I’m not sure what you’re doing, but I assume you’re aware Omega sells pins, ring connectors, spade connectors, etc. made from thermocouple alloys?
Good point about the flux contamination, Crafter_Man.
I have the Omega catalog - correction, I have a string of various age Omega catalogs tucked away here and there for reference - an Omega-made custom thermocouple with reference junction built in, on the table behind me - a drawing for a custom RTD on my desk - a cut-open Omega sheathed RTD on top of my monitor - I have their stuff all over, and a smattering of Hart and Minco and Tempco and Tinsley and Burns stuff thrown in for fun.
The struggle I hoped silver solder might fix is when I want to build tiny TC assemblies into other things I’m making, where there’s no room for connectors and I can’t subcontract it.
Sounds like there’s a good reason they use welding when they have to cram these things into a tight space (it’s amazing they can weld a wirewound RTD onto 4 wires and then cram the mess into a sheath that’s only 1/16" outside diameter). I’ve seen a few folks hawking capacitive-discharge welders for TC wire and other small wire - maybe I should try one of those?
Now again, I don’t know what you’re doing, but is there a reason you can’t run the thermocouple wire all the way back to the signal conditioner? Must you have a discontinuity?
Keep in mind that, for an R&D project, thermocouple wire is cheap, especially when compared to engineering labor time…
I want to measure a small differential temperature across a thin wall in a machine. So I want to drill holes through the wall, thread type E thermocouple wires through the holes (alternating positive and negative types), weld the tips together, and run, say, copper from the ends of this system. This will produce a thermopile for measuring the difference. Since the voltages are actually generated in the lengths of wire between the junctions, where the gradients are, I don’t want to run the wire outside of the machine through an enormous gradient and back in again. I want all the wires to be at high temperature, and the only gradients are the small ones (except of course that I have to run the leads out through a big gradient - and I can use the same roll of wire for both leads and can study it beforehand by playing a heat source along its length looking for spurious signals).
I guess I still don’t understand what you’re doing.
As you’re apparently aware, in a traditional thermopile arrangement the junctions are wired in an alternative fashion back and forth between the two measurement points. In other words, all the “inner” wiring can be done between the two temperature measurement points. You then “bring out” just two wires, one from each end of the thermopile. These two wires go from the “hot area” around the measurement points to the “outside world,” wherein they connect to a nanovoltmeter.
The fact that these two wires are subjected to a large temperature change along their length does not cause a problem.
One thing… you said you’re using type E thermocouple wire. There’s nothing wrong with using type E, but keep in mind that the two copper-Chromel (or copper-Constantan) junctions must be made using a common isothermal block. (You don’t have this concern with type T, which is why thermopiles often use type T thermocouples.)