Making Thermocouples

The industry accepted method is spot welding.

That said, I've made passible K thermocouples by oxy/ace welding the wires together.

Iron/constantan works well in the 500 degf range with the wire ends twisted together with a pair of pliers when using a 10 meg DVM. An input with lower resistance would likely require a better connection.

Hul

M> How does one make a junction with TC wire? Silver solder? Braze? What

I agree with Jim, I just used oxy/acetylene until I had a tiny bead formed (it took a try or 3 though).

We used to weld them in mercury. Fill a quart jar about an inch deep in mercury. Add about an inch or two of transformer oil on top. Get a 10 amp or so 102 vac variac. Run a 12 gage copper wire from the variac "common" output down into the mercury. Twist the two thermocouple wires together about 4 turns for the junction end and loosely twist the other ends together a turn or two. Clamp onto their "other ends" with an appropriate alligator clip. Using appropriate insulted gloves, etc. Set the variac at about 60 volts for 16 ga. thermocouple wires. You will have to expirement on this. Dip the 4-twist end of the pair down through the oil barely into the mercury . A spark will occur as you do this. The weld only takes a second. Pull the junction up into the oil to cool it and to mimimize scaling. We did this on a regular basis at the vaccuum heating furnace factory where I worked back in the early 60's. If you are bothered by the mercury, don't write back to chew me out about its use.

Alternately, twist the wires together, folding a little bit over at the end of the twist like a lineman's butt splice. Hold the junction with the end DOWN and heat to fusion with a neitral flame from an acetylene torch. Done properly, a nice little sphere will form at the end of the junction.

You should not be brazing a thermocouple because you are adding additional metals to the mix. This will change the characteristic of the joint. You won't know what the "true" temp is.

Pete Stanaitis

--------------------

M> How does one make a junction with TC wire? Silver solder? Braze? What

When I was in school, they didn't have DVMs. We hooked a pair of bucking thermocouples to a bridge, and put one in an ice bath. When the bridge was balanced, there was no current flow, so the resistance of the junctions and wire did not matter. I have seen a lot of welded thermocouples, but silver solder is okay too. The introduction of a third metal into the circuit cancels out. For example, the net voltage across iron=silver=silver=constantan sums out the same as iron=constantan.

Weld is the standard and best way (no rod at all), however, depending on application (say, bean probe in a coffee roaster), silver solder can work just fine... for that matter depending on application (say, a boiler in an espresso machine), regular solder can work fine.

In article , Steve Ackman wrote: Thanks to all for your replies.

Pete don't worry. I am of the age where we distilled mercury from a compound in High School chemistry lab and the prceeded to coat all our silver coins and put them back into circulation. The distillation was part of the lab. What we did with the product was our bussiness.

Chuck Pilgrim

In the following thread (and probably others), Spehro Pefhany writes "Inert gas shielded welding is the "right" way, BTW."

-jiw

I use a thermocouple welder at work when I make them. Definitely a trick to making them work. In a nutshell the machine charges up a capacitor and then you hold the wires in a pair of pliers that are connected to the machine and press them gently against a special block of something (graphite? Not sure.) and then push the button. Get a spark and you have two wires that are either welded together if you did it right, or burnt away if you had the knob turned up too far. If you want more information on the process, I can tell you more, but the only thing I'd do differently is to do the welding in some sort of neutral environment, nitrogen would do better than air. Some of the come away with little bubbles in the bead. They all work fine, just look like shit. For the hassle, though, Omega at omega.com has them cheaper than you can make them, usually. I like the ones with the stick on tab. The last time I had to make them, we needed leads that were way longer than they sold, and it was a funky application, so otherwise we'd have bought them. Their cheapest tc's are here:

Any way you want. Welding is preferred because there is no danger of making a joint less tolerant of oven-accelerated corrosion than the wires.

Simple contamination of the couple at a point junction does NOT affect the accuracy. For thermal-contact reasons, spot welding both wires to an iron washer, then riveting the washer in place, is common practice; you need the rugged iron washer to tolerate the riveting force.

If it actually makes contact with food or potable water, be sure to use one of the cadmium-free brazing alloys.

Best regards, Spehro Pefhany

TIG is the easiest way by far!

Just twist the ends together, strike up an arc with the torch on a piece of scrap and feed the twisted pair into the arc. You can watch the ends fuse and form a little ball. All is good! If you are not happy with the result, nip it off and try again.

Cheers Trevor Jones

I've used 3 methods to weld them. One is as Trevor described with TIG. I also often use an O/A "Little Torch".

Decades ago I made lab thermocouples with drafting leads (graphite) and a 10-amp elex bench power supply. Twist the wires, nip off tag ends. Connect one lead to the TC-to-be and the other to the drafting lead. Wearing dark glasses, touch the lead to the end of the pigtail. A little arc forms, and then a nice little molten bead.

Note that this is not vision-safe procedure. The arc is quite bright, one should properly use eye protection suitable for low-current arcwelding because that's what's being done. That said, I made a lot of TC's this way and never got a flash burn, and my eyes are still very healthy 35 years later. YMMV.

A car battery would probably work just fine, and you could get the lead (graphite) out of an ordinary #2 wooden pencil.

Yet another method is to use a big capacitor charged up to not more than its rated voltage, along with a piece of graphite. You'd have to experiment a bit, but I know it works because I knew a tech that did it this way. Charge up cap, connect cap to graphite post or block and twisted-and-nipped pigtail as above, close eyes and scratch the pigtail on the graphite like striking a match. ZAP. When the cap size and voltage are right, you get a nice weld every time.

This is a great thread- mercury, PCBs, arc-welding, portentailly exploding caps and batteries, yet everyone involved seems to be alive. Ahhh, life before the nanny-state...

Dave

Like others here, I twist 'em and heat with an AO torch till the wires fuse. A light touch is needed, or you get a molten mess. Mostly (for quick use) I just twist the wires together and go, but, especially with J type, corrosion (rusting) may present a problem after a while.

While ready-made TCs are kinda cheap, if you have a bunch of TC wire, rolling your own is a *lot* cheaper. Beware, however, that "extension grade" wire does not make as high quality a thermocouple as the better wire; it's good enough for most uses, though. If you really need good accuracy, buy an RTD; the thin-film platinum RTDs are pretty reasonable & very rugged, and get you better accuracy than with a TC. Also, long wires are easily compensated for with RTDs. (The higher-priced wire-wound RTDs are even better, but are easily damaged by physical shock.)

I use dozens (OK, well over 100) of RTDs in my work, and only about 10 TCs.

Joe

Lots of great methods have already been posted. I have used a purpose-built thermocouple welder lots of times (>500). Its finikey until you get the settings right and then its perfect. Even better is what they call an "intrinsic thermocouple". This is where you spot-weld the two wires to the substrate to be measured. The substrate itself is the actual electrical connection between the TC wires. The setup measures the average temp of the substrate between the two wires (they are generally very close together). This is used most often when you are doing measurements where rediculous accuracy is required and you can't create a hole to insert a thermosouple into - it minimizes errors between what is actually being measured (keep your eye one the prize) and the junction itself. I like the TIG method as well - I'd use that when our TC welder would go on the fritz. As for all the other improvised methods, they really aren't as bad as some would have you believe. Just remember the fundamentals: The thermo-electric effect is the core of what we are using - that is, a given voltage will be developed between the hot end and the cold end of a conductor and that depends on a) the material and b) the delta T. Those wires are totally unaffected by the conatact material between them. Still worried? The the other metal (eg solder) might affect the result and it is supposed to be due to another junction, WHAT is the delta T that produces the error? what is said delta T? Nil. That solder is supper thin and all of it is at the same temp: thus no thermoelectric voltage and no error. Any way, that's my two cents. Lots of voodoo surrounds TC's. If you want to avoid it, find a spot welder and you'll be fine. If you are unaffraid or broke, solder away.

Cheers, Dave

jgandalf wrote:

All good information from DaveO and Joe, but I have one quibble with DaveO's statement, "The setup (two TC wires connected to the object being measured: awright) measures the average temp of the substrate between the two wires (they are generally very close together)."

The response of the two TC wires connected separately to the object under test will NOT be influenced by the average temperature of the substrate between the two wires, but rather by the temperature at the two junctions formed between the wires and the substrate. If the points of contact were far enough apart, you could apply a torch to the substrate between the two junctions and have zero change in TC reading as long as the temperature of the two junctions was not changed.

awright

Snip

A quick and dirty high temperature resistive sensor is a mains voltage quartz halogen lamp.

Jim

Jim, can you elaborate on use of the T-H lamp as a temperature sensor? Do you thermally couple the complete lamp to the OUT and measure the resistance? What kind of sensitivity do you find? Interesting.

awright