Narrow-Bandwidth Television Association

I borrowed a neat "heat camera" from work, and did some measurements of my circuitry...





The above is one of my AC-DC converter boards, which is supplied with 9.5V AC.
It's getting up to about 52C. This one is supplying everyting but the LED matrix.

The Arduino itself has several "hot spots". It's getting up to about 53C.

The motor control board has a negligible heat profile.

The motor is doing OK - note that the intensity is relative not absolute, so use the scale on the side

An overview of the sub-boards. Left to right; LED board, IR board, motor board, Arduino. The resistors are the main heat source on the first two; even so we're just talking 20C or so. The motor board has barely measurable heat. The Arduino is blazing away at low 50s.

Very interesting! If the reading is relative how do you work out the actual temperature?

I've seen some of these cameras here and as usual with all technology they're getting cheaper by the day, though still a bit too expensive for me to justify the cost.

Steve A.

Steve Anderson wrote:Very interesting! If the reading is relative how do you work out the actual temperature?

I've seen some of these cameras here and as usual with all technology they're getting cheaper by the day, though still a bit too expensive for me to justify the cost.

Steve A.

I should have said the *colouring* is relative. Red in one image is not the same temperature as red in another image. You have to refer the colours to the scale to get the absolute temperature. Something "red" could be very hot or very cold, depending on the temperature range of all other pixels. So, refer to the scale

ALSO, something to be aware of - the heat sensor is VERY sensitive. Often what looks like a hot spot is actually a reflection of heat coming from another location. For example, I point the camera at the fridge, I see very clearly... ME! Reflected heat off the door of the fridge.

Currently for checking temperatures of semiconductors I use a Maxim DS18B20 direct-to-digital device. Packaged in a standard transistor TO92 case it can measure from -55 Celcius to +125 Celcius, the output is a 2's compliment serial data stream. The resolution is 1/16 of a degree (0.0625 degrees) and the accuracy is +/-0.5 degrees. The resolution can be adjusted between 9 and 12 bits, the 12-bit measurement takes 750ms, lower resolutions are proportionately quicker.

This data is fed into an 8-pin micro and simple RS232 signal comes out in a CSV format for direct import into Excel, or if you only want a current reading, something like Hyperterminal is all you need. It could easily feed a 7-segment display at the same time, but I haven't bothered with that, I rarely use it anyway. In the past I did need it for one particular project.

This only measures the temperature on the outside of the device being monitored, you need to calculate the actual chip temperature from the thermal data in the datasheet of the test subject and the power measured going into the device. You would have done these calculations anyway before even building the unit, this just confirms all is hopefully well.

Datasheet attached.

Steve A.

Same brand, different model. Infrared Flir Thermal Camera on AliExpress. Not too pricey if you need one.

You should know a few things more about the teperature of semoconductor devices.

- Silicon semiconductors are made by diffusion on a temperature of about 800 degrees C.
- This diffusion process never stops,
- but every drop in teperature by 10 degeees slows it down by a factor of 2, so half as fast.
- So at room teperature the speed is very slow, almost halted.
- On the other hand then every raise in temperature speeds up the diffusion by the same factor of 2.
- Continuing the diffusion will ruin transistors on the long run.
- In most cases at a temperature of 150 degrees, of the silicon, the speed is such that the life time is still 10 years.
- The heat finds its way to the environment via the encasing,
- but the outside of that encasing is already colder than the inner side of the transistors.
- The documentation of the devices gives information about the difference in temperature inside-outside.
- An outside temperature of 50 degrees outside is rather common for semiconductors.