Author Topic: monitoring the package temperature  (Read 854 times)

0 Members and 1 Guest are viewing this topic.

Offline metebalciTopic starter

  • Frequent Contributor
  • **
  • Posts: 451
  • Country: ch
monitoring the package temperature
« on: January 23, 2022, 12:41:16 pm »

I would like to monitor the temperature of ICs/transistors etc. when prototyping, to be sure they are in safe operating area. Is the easiest/best/good enough method to use a PT1000 and attach it to the IC package with a thermal compound ? or something else ? or should I not bother with these and find a thermal imaging solution ?
 

Online magic

  • Super Contributor
  • ***
  • Posts: 6733
  • Country: pl
Re: monitoring the package temperature
« Reply #1 on: January 23, 2022, 01:36:58 pm »
Thermal imaging is a bit exotic/expensive stuff.

Thermocouples are cheap and low thermal mass sensors that can be glued to stuff with sticky tape. At a moderate expense I have bought a very fine thermocouple (0.1mm wire diameter) - it works fast and fairly accurately even without gluing (thin wires → low thermal conduction through the sensor itself), which is to say, probably to a ± few degrees but that's good enough in such application.

This is an interesting device, but the lack of distributors, the shipping costs from them to Europe and the EU VAT have so far successfully discouraged me from buying it. And I can't find alternative products with similar functionality. There was a Chinese clone of unknown quality years ago, but it's long gone (lawsuits?).
https://www.thermoworks.com/close-focus-ir/
 

Offline rpiloverbd

  • Regular Contributor
  • *
  • Posts: 157
  • Country: bd
Re: monitoring the package temperature
« Reply #2 on: January 23, 2022, 01:59:23 pm »
You can also use an LM35 or DS18B20  and Arduino/any other microcontroller of your choice. If you use arduino, you'll be able to see the temperature in your serial monitor. That will save the cost of any extra display.
 

Online magic

  • Super Contributor
  • ***
  • Posts: 6733
  • Country: pl
Re: monitoring the package temperature
« Reply #3 on: January 23, 2022, 02:04:17 pm »
TM-902C costs a few bucks, probe included.

And anything TO92 has huge thermal mass, have fun waiting for it to catch up to the temperature of whatever you placed it on. And they top out at 125°C, although if your project runs that hot then maybe you are doing it wrong ;)
« Last Edit: January 23, 2022, 02:06:11 pm by magic »
 

Offline metebalciTopic starter

  • Frequent Contributor
  • **
  • Posts: 451
  • Country: ch
Re: monitoring the package temperature
« Reply #4 on: January 23, 2022, 02:36:19 pm »

Thermocouples are cheap and low thermal mass sensors that can be glued to stuff with sticky tape. At a moderate expense I have bought a very fine thermocouple (0.1mm wire diameter) - it works fast and fairly accurately even without gluing (thin wires → low thermal conduction through the sensor itself), which is to say, probably to a ± few degrees but that's good enough in such application.

I said PT actually because I can directly connect it to the DMM without any other device and PT temperature range is more than enough. I see PT sensors usually with capsule and I was looking to get an individual thin film PT100/1000 component rather than a probe/like the one sold by adafruit etc., but it is still not a bare wire. I guess then the response time of thermocouple is better than PT100/1000 ?

 

Offline Jeroen3

  • Super Contributor
  • ***
  • Posts: 4067
  • Country: nl
  • Embedded Engineer
    • jeroen3.nl
Re: monitoring the package temperature
« Reply #5 on: January 23, 2022, 02:54:18 pm »
A thermal camera.
Get yourself a Seek Thermal to attach to your phone. It's worth it when brining up boards and finding defects!

Second choice would be a thermocouple meter. The RTDs are too big, too slow.
You can attach them with a bit of polyamide tape.
 

Offline metebalciTopic starter

  • Frequent Contributor
  • **
  • Posts: 451
  • Country: ch
Re: monitoring the package temperature
« Reply #6 on: January 23, 2022, 03:12:24 pm »
A thermal camera.
Get yourself a Seek Thermal to attach to your phone. It's worth it when brining up boards and finding defects!

Second choice would be a thermocouple meter. The RTDs are too big, too slow.
You can attach them with a bit of polyamide tape.

OK thanks. I used a thermal imaging accessory before but not for electronics. I guess it does not matter the accuracy or the effects of the environment for this application, correct ?
 

Offline Jeroen3

  • Super Contributor
  • ***
  • Posts: 4067
  • Country: nl
  • Embedded Engineer
    • jeroen3.nl
Re: monitoring the package temperature
« Reply #7 on: January 23, 2022, 05:01:52 pm »
You can't really use them in direct sunlight or with reflective parts (tops of aluminum caps) but other than that, they're great to view heating ic or even illuminate traces to shorts on a pcb.
They are within a few degrees accurate.
 

Offline metebalciTopic starter

  • Frequent Contributor
  • **
  • Posts: 451
  • Country: ch
Re: monitoring the package temperature
« Reply #8 on: January 23, 2022, 06:04:03 pm »
OK thanks. I will try these.
 

Offline metebalciTopic starter

  • Frequent Contributor
  • **
  • Posts: 451
  • Country: ch
Re: monitoring the package temperature
« Reply #9 on: January 23, 2022, 07:10:04 pm »

I just saw a TI paper (ref below), and my summary from that (similar to the points above):

- order of accuracy (paper date 2016): IR camera, fluor-optic probe, thermocouple, IR gun (max field of view 4mm diameter)
- thermocouple: 36 to 40 gauge J or K wire. It is important for wire to be thin, 36+ gauge.
- it is better to attach the thermocouple with a thermally conductive epoxy, taping is not recommended.
- when using an IR camera or IR gun, reading should be corrected for emissivity of the surface.

More info on section 3.1 Case Temperature Measurement of : https://www.ti.com/lit/an/spra953c/spra953c.pdf
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf