Good day I have not used thermocouple before. I only know the basic principle. From what I understand, the voltage produced is due to the dissimilar metals with their joined ends being applied with some heat. (seeback effect).

My main concern before I do my test is that I want to make sure that it is ok to attach the thermocouple junction (joined end) to conducting trace (2.5A) without destroying the measuring circuit of the thermocouple. I wanted to measure temperature of a trace which I would flow 2.5A controlled.

In the figure, I wanted to also know if the 2.5A in the test circuit will only flow on the main circuit (and not some of it flowing into the measuring circuit). The ground of the DC power is different from the measuring device. My dc power supply is a table top DC power supply while the measuring device is a battery powered digital thermometer (with ports for the thermocouple).

My understanding is that, since my DC power supply (which would supply the 2.5A of the test circuit) has different ground compared on the ground of the measuring device, then the 2.5A would flow only on the main circuit.enter image description here

Edit: I should've made it clear earlier that the trace I want to attach the thermocouple into, is a trace that is exposed (has no solder resist).


2 Answers 2


This will work as long as you keep the resistance between circuits high, this is also called isolation. So if the measuring device was a DMM and it was separated by air, then there would be nothing to complete the circuit between grounds and no current would flow.

But there are caveats, high voltage can be a problem as it brakes down insulation so also make sure you are using a low voltage (maybe lower than 100V) and you should be fine.

The other caveat is, you need to really make sure that the measuring device really is insulated from the DC power supply circuit. You can do that with a DMM, and put the meter in continuity mode and put one terminal on one ground and the other terminal on the ground of the thermocouple. If it measures very high (more than 1Meg preferably in the 10Meg or greater range, you should be fine).

enter image description here

Another thing you could do is put a thin layer of kapton tape between the PCB trace and insulate above the kapton with a good insulator and you might be able to make a good temperature measurement that is slightly delayed but close to the trace temperature. Another option would be to use a thermistor which would yeild greater accuracy.


There will be a thermocouple voltage at every junction of dissimilar metals, not only at the weld that is indented to be the temperature sensor (the "hot junction")... but also at the other end, where the thermocouple wires enter the measuring device (the "cold junction"). But if the weld is also in contact with another dissimilar metal, that will create yet another thermocouple junction.

In this configuration, the thermocouple voltage is related to the temperature difference between the hot junction and the cold junction. It forms a loop, so Seebeck originally though of it in terms of current. But most modern thermocouple measurement systems measure the voltage difference that develops across the weld. (As long as the cold junction has a known reference temperature, and both the positive and negative cold junctions are at the same temperature, the two additional junction voltages cancel out.)

The voltages that are being measured are in the low millivolts range, and the relationship between temperature and thermocouple voltage is non-linear. The relationship is at least predictable when using known alloys (like type K thermocouple for example), but introducing additional junctions is probably going to not work as expected. Worse, it is not even guaranteed to fail in a predictable way.

A better approach is to stick with conventional, electrically isolated thermocouple sensing, making sure that there is electrical isolation but still good thermal contact. This could be as simple as using mica or maybe Kapton tape as an electrical insulator, or routing the PCB trace on an inner layer.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.