I have a device with 2 mechanical parts. One side is stationary and can be connected to an electronic circuit with power. The second side is an enclosure that can have high ambient temperature (100-120°C). I cannot bring a wire from one side to the other, but I would like to read the internal temperature of the second side. The mechanical parts have facing plates of ~50mmx50mm with a distance of 10mm. Sensor must be able to read from -40°C to 220°C. One of the facing plates will rotate, as it is connected to a motor.

diagram of the configuration

Reliability and low cost are important factors. What type of sensing could be done?

I thought of magnetic coupling with energy harvesting, just like RFID. As the environment will be very noisy and reliability is important, I'd like something with a modulation robust to amplitude variation (FM or better). I thought maybe an oscillator connected with a RTD that would change the impedance of a transformer with a variable frequency could do the job, but I've never done this type of circuit.

Can somebody point me out to a temperature sensing technique that could fit? Or some literature that would help me design such circuit.

  • 2
    \$\begingroup\$ Would thermopiles work? IR is better than RF if there is lots of noise \$\endgroup\$
    – DKNguyen
    Nov 1, 2019 at 15:25
  • \$\begingroup\$ I hardly make the link between thermopile and IR sensing. Also, I have forgotten to mention, but the facing plate orientation may change over time, so IR coupling may be difficult \$\endgroup\$ Nov 1, 2019 at 15:30
  • 1
    \$\begingroup\$ Sorry, they were not linked. Two separate and distinct approaches. Either thermopile temperature (long-IR) sensing at a distance, or contact sensing with near-IR for a comm link instead of RF. You can always put multiple LEDs on different facets and drive them at the same time. \$\endgroup\$
    – DKNguyen
    Nov 1, 2019 at 15:32
  • 1
    \$\begingroup\$ Provide details (in your answer and not in comments) of the facing plate axial misalignments. Don’t forget anything and be clear. A diagram would really help. \$\endgroup\$
    – Andy aka
    Nov 1, 2019 at 16:04
  • \$\begingroup\$ Question updated \$\endgroup\$ Nov 1, 2019 at 16:15

2 Answers 2


I am working on something similar at work as a side project. It's to measure currents on a rotating winding. I've decided to go with IR LEDs and photodiodes as a communication link with multiple LEDs being driven in parallel so one of them is always in view by the photodiode.

It isn't clear whether a rotating connection is allowed. If it is, you could also go with slip rings, of either mercury or fiber optic kind. That's the classical method.

A more novel idea that only works because you're measuring temperature is to have a device that reproduces the temperature measured at the temperature sensor somewhere on the rotating wheel where it is exposed and could be measured that with a thermopile or IR thermometer. A third temperature sensor would be required on the shaft itself so it could monitor the temperature of the exposed area and adjust it as necessary to match the internal temperature sensor.

Given the high ambient temperture and noisy environment, this actually might be simpler and more straightforward than any of the other methods which which require processing to occur within the chamber or on the shaft. The main hiccup is how to provide enough power for the exposed temperature element and how to provide lower than ambient temperatures if they are possible. A peltier element might be able to be used to provide both above and below ambient.

  • \$\begingroup\$ What do you mean by reproducing the temperature? Do you mean having a thermal conductor connected to the rotating disc? I fear that convection will cool down the disc \$\endgroup\$ Nov 1, 2019 at 20:50
  • \$\begingroup\$ @Pier-YvesLessard No. I specifically said "reproduce", not "transfer" or "transmit". I mean a system that actively measures the temperature at the sensor against another temperature sensor at the exposed element, and applies energy to heat it up (or cool it down) so that they match. A temperature repeater, if you will. \$\endgroup\$
    – DKNguyen
    Nov 1, 2019 at 20:54
  • \$\begingroup\$ Can you give an example on how this is done? How do you bring energy on the rotating part? \$\endgroup\$ Nov 1, 2019 at 20:56
  • \$\begingroup\$ @Pier-YvesLessard Well that's the tricky bit. Batteries, induction coils, lasers, slip rings. It is very dependent on the exact mechanics of your setup and how much power is required to run the heating/cooling element so we don't have enough information. You didn't even make it clear if the faces are sealed inside the enclosure or how harsh an environment is harsh. I think that's asking for trouble if it's sealed because then you're stuck with RF in a noisy environment or IR transparent plastic in a hot environment, not to mention everything has to run at 100C too. \$\endgroup\$
    – DKNguyen
    Nov 1, 2019 at 21:12
  • \$\begingroup\$ At some point it's just easier and cheaper to slap an expensive window that can pass 5.5um-14um on your enclosure and point a thermopile at it from the outside. But I don't think those can survive 100C. The one I saw can only survive 70C \$\endgroup\$
    – DKNguyen
    Nov 1, 2019 at 21:27

Adding a solution proposed by this paper


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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