2

I have a thermocouple with me which is custom made and its leads are very thin. It has a diameter of 0.19 mm and this has to be soldered to a chip (Max6675). This setup will be placed on a harsh environment. I am stumped on how to deal with the soldering of such thin leads which will withstand the environment.

It would be great if you guys could help me out here. I have added the picture of the thermocouple for reference. One thing is that the length of the leads are 10mm now and not 5mm. Thermocouple

Ther thermocouple will be embedded in an engine valve and the custom PCB will sit on top of the retainer.

enter image description here

So I am looking to connect the thermocouple leads to this PCB for measuring the temperature of the valve. Ofcourse this is the first prototype of this experiment and I will be trying this out in a customized rig and not in the engine itself.

I appreciate all the comments!

Vish_evo
  • 209
  • 2
  • 7
  • 1
    What do you mean with 'harsh'? I think thermocouples are usually connected via terminal screws, but you need to provide a stiff mechanical mounting to avoid bending of the leads. – Vladimir Cravero May 09 '17 at 07:33
  • mechanically couple strain relief onto the probe and into the first few cm of the cable – dandavis May 09 '17 at 07:48
  • 1
    Harsh chemicals? Harsh temperature? Usually the tip of the TC extends through a hole into the harsh location, and the bare wires are kept outside in cold dry desktop environment. If not, then your TC probe is far too short! :) On the other hand, you can get 2-pin thermocouple connectors made from ceramic, Omega SHX type. IIRC the plastic type are only good to 150C – wbeaty May 09 '17 at 08:51
  • Some thermocouple wires don't solder very well. Mechanical clamping is a common solution. Are you planning to have a small PCB hanging on the end of the probe or a larger board that is rigidly connected to the probe? – KalleMP May 09 '17 at 11:25
  • First, and most important - have you verified that you CAN solder your TC leads? Both of them? – WhatRoughBeast May 09 '17 at 12:56
  • I should have been more clear when I had typed the question. I apologise for that. The engine head's goes upto 80 deg C and has oil (not a lot) splashing around for valve's lubrication.There is no room for thermocouple connectors so that is ruled out, since the retainer is about 18mm in diameter. The leads will not have any weight on it as PCB will be fixed on top of the retainer. – Vish_evo May 10 '17 at 03:40

2 Answers2

5

What is the thermocouple type? J? K?

Sounds like you need an extension cable to go between your TC probe and your circuit board. Omega.com has wire going down to #40ga size, search on their site for "TT" wire, it's teflon type (e.g. their TT-K-24 is K-type 24-gauge thermocouple wire, both wires inside a teflon jacket.)

The usual technique is to screw the tiny leads down inside a larger thermocouple male-pins connector, like these from Omega, they're about $3 each. Then plug this connector into a female of the same TC type (yellow plastic is the code for K-type.)

If the temp-gradient at this connector is minimal, then just use copper wire to connect between the female side and your PCB terminals. They do have female panel-mount, search "RMJ," so your project enclosure box can have a thermocouple jack rather than permanent wiring.

Note that some thermocouple alloys are not solderable with electronics solder, non-wetting. (That's why we use screw-clamp terminals.) Also, if there's significant temperature gradient at the location of the connections (for example, on a PCB next to a warm processor chip,) then the TC wire and copper will become a new thermocouple, and the non-uniformly warm PCB will throw your measurements off. The screw-clamps inside the plastic TC connectors are made from thermocouple metals, K-type chromel-alumel etc., so it's possible to make extension-cord connections in large-gradient locations without disturbing the main thermocouple.

Finally, one trick is to make your own connectors: get some tiny male and female gold crimp-pins, like those used inside older DB connectors. Crimp each one onto the unsolderable thermocouple wires. Crimp the opposite sex onto copper wires. Then plug them together, and perhaps shrink-tube the individual sides to keep them from touching. I get those gold crimps from Digikey: A1008-nd and A1009-nd

IIRC, with K-type TCs, the Chromel is solderable, but the Alumel is not. Both alloys are mostly nickle. With K-type and T-type thermocouples both sides are solderable.

And finally, when making TCs from wires, if you can't spotweld the alloys, and they won't wet with solder, instead use one of those same gold crimps above. Just crimp your chromel-alumel wires into the same crimp terminal, and snip off the rest of the gold terminal. Seems to work better than just twisting them, but not as well as actual spotwelding. Also, the proper TC crimps are little nickel sleeves sold by Leeds & Northup "Quiktip" TC pressure connectors # 128543. Procinst was still selling these a few years back, along with the crimp-tool http://procinst.com/legacy-product-support/leeds-northrup/quiktip/

wbeaty
  • 10,761
  • 23
  • 39
3

UPDATE

You need to, if possible, stop the PCB fab house from proceeding. You need to examine the PBC design to determine if you can retrofit the current board or if another revision is required. Ask them to put it on hold immediately to give yourself time to work out the mechanicals.

End of Update

You neglected to mention a couple of things. Usually when a thermal couple is attached to a chip, thermal epoxy is used. Unless the chip is a thermal couple to digital converter.

A photo of the PCB would help.

Not sure why you are asking this question. I cannot fathom someone having a custom thermal couple made without considering how to fasten it to the PCB.

The answer is very simple, solder it in the through holes on the PCB connected to pins 2 and 3.

The PCB should also have been laid out to accommodate the necessary mechanical fasteners to secure the thermal couple to the board to avoid damage from vibration.

The only reason I can think you'd ask this question is because you do not have a PCB for the chip. In that case the answer is to fabricate a PCB.

The MAX6675 requires a PCB with with a a large ground plane for good thermal management.

From the datasheet:

Self-heating degrades the temperature measurement accuracy of the MAX6675 in some applications. The magnitude of the temperature errors depends on the thermal conductivity of the MAX6675 package, the mounting technique, and the effects of airflow. Use a large ground plane to improve the temperature measurement accuracy of the MAX6675

Misunderstood
  • 7,287
  • 1
  • 11
  • 24
  • You are right. The PCB is being made for us by a supplier as I type this. Let me rephrase the question then. What would be the right way to mount the leads onto the custom PCB? – Vish_evo May 09 '17 at 09:03
  • That's not good. Usually the PCB has to have the mounting holes for the fasteners. And it appears @wbeaty has this reworded question covered very well. – Misunderstood May 09 '17 at 09:09
  • 2
    The PCB is being made before you have worked out what form of connectors it needs. Do you really need someone else to tell you what's wrong with that situation? – Andrew May 09 '17 at 10:09
  • Please tell me that the PCBs which are being made are a small test batch for prototype evaluation, and not a production run. – WhatRoughBeast May 09 '17 at 12:58
  • Ofcourse it is not a production run. In fact only a couple of PCBs are being made as first prototypes to see the feasibility. I have updated the question with more details. – Vish_evo May 10 '17 at 03:32