Author Topic: Microcontroller temperature rating  (Read 1042 times)

0 Members and 1 Guest are viewing this topic.

Offline rbeTopic starter

  • Newbie
  • Posts: 7
  • Country: ie
Microcontroller temperature rating
« on: May 29, 2023, 12:30:39 pm »
I am considering a microcontroller for PFC design in a universal 1500W PSU.
STM32G071xx microcontrollers seem exactly what I would need for this application. However I am indecisive when it comes down to temperature rating of it.
The cost of uC is important and thus I have my eyes on the lowest temperature rating of T_a = 85 °C and a corresponding T_j = 105 °C part, but I wonder if that would be enough for this type of application or should I consider a more expensive higher temperature variant instead? Ambient temperature exceeding 85 °C seems very high even with a relatively high room temperature say - 30 °C.

Let me know what are your thoughts on this? Perhaps, somebody may know what is the usual accepted temperature rating in these type of applications and are there any requirements that would suggest to use higher temperature rated packages?
 

Offline Scrts

  • Frequent Contributor
  • **
  • Posts: 797
  • Country: lt
Re: Microcontroller temperature rating
« Reply #1 on: May 30, 2023, 08:06:01 pm »
Is your power supply going to be air cooled? Is there going to be a fan? You can use a cheaper part, build the power supply, test it and if it doesn't work - upgrade to a higher rating part.
On the other hand - are all other components going to be rated for higher temperature as well?
 

Offline PCB.Wiz

  • Super Contributor
  • ***
  • Posts: 1458
  • Country: au
Re: Microcontroller temperature rating
« Reply #2 on: May 30, 2023, 08:26:05 pm »
I am considering a microcontroller for PFC design in a universal 1500W PSU.
I guess that has a fan ?
Is that going to operate where people are, ie in a lab ?
Or does it need to work up a pole, in a small enclosure, in outback Australia ?
 

Offline uer166

  • Frequent Contributor
  • **
  • Posts: 872
  • Country: us
Re: Microcontroller temperature rating
« Reply #3 on: May 31, 2023, 03:26:50 am »
One thing to remember is the T_a of 85C is completely irrelevant, what matters is the T_die spec of 105C. If you're running the micro full blast, then the temp rise of die is ~10C, but by reducing clock, peripherals used, VCC, and other such stuff, that delta reduces drastically.

I don't see why you can't use this micro for a PFC from a temp rating perspective, if the ambient conditions are too much for it, then they are also way too much for the electrolytic caps, switching devices, etc etc.
 

Offline rbeTopic starter

  • Newbie
  • Posts: 7
  • Country: ie
Re: Microcontroller temperature rating
« Reply #4 on: May 31, 2023, 07:19:19 am »
Thanks for reaching out to everyone. This PFC stage will be deployed inside a household, so no harsh environment per se.
The problem is that for this particular microcontroller there is no higher temperature rated variant (so far).
Yes - there will be forced convection cooling, but devices do get hot at these power levels. 
 

Offline Siwastaja

  • Super Contributor
  • ***
  • Posts: 8090
  • Country: fi
Re: Microcontroller temperature rating
« Reply #5 on: May 31, 2023, 07:32:01 am »
You omitted the important footnote (3):

Quote
The T A (max) applies to P D(max). At P D < P D (max) the ambient temperature is allowed to go higher than T A (max) provided
that the junction temperature T J does not exceed T J(max). Refer to Section 6.9: Thermal characteristic

Thus Ta_max is a convenience number for those who plan to enable all features, use the MCU at maximum clock frequency and maximum Vcc etc - probably; check out how they define PD(max).

You should do a proper Tj calculation, first calculating Pd based on the tabulated peripheral, core and FLASH current consumptions, then Tj = Ta + Pd*RthJ-A.

Running at slower clock both decreases power dissipation and hence increases the Ta while keeping Tj the same, but also increases the chances of proper opertion even if Tj is slightly exceeded. Don't count on the latter, though.
« Last Edit: May 31, 2023, 07:34:26 am by Siwastaja »
 

Online Miyuki

  • Frequent Contributor
  • **
  • Posts: 896
  • Country: cz
    • Me on youtube
Re: Microcontroller temperature rating
« Reply #6 on: June 03, 2023, 07:19:24 pm »
Especially with forced cooling, you will get a hot side and a cool side
So you can find a cold spot to place it
And that Tj limit tends to be to guarantee a flash life of 20 years (need to check it in the datasheet)
 

Offline barshatriplee

  • Regular Contributor
  • *
  • !
  • Posts: 130
  • Country: bd
Re: Microcontroller temperature rating
« Reply #7 on: June 05, 2023, 07:28:07 am »
For a universal 1500W PSU, the temperature inside the power supply can increase due to factors such as power dissipation, environmental conditions, and thermal design. However, reaching an ambient temperature of 85 °C seems unlikely in typical operating conditions, especially if the room temperature is around -30 °C.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf