Electronics > Projects, Designs, and Technical Stuff

Proper thermal derating of parts

(1/1)

Pack34:
I have a part that has a max operating temperature of 70C (158F). The maximum board temp is about 45C (114F) with the part itself ringing in at a max of 40C (104F). I don't see the part getting any hotter in a standard use-case. I had to really smother the PCA and enclosure to get it that hot.

Is 20C typically a good minimum buffer for operating temp? In the past I've always overkilled the rating to a preferred 125C and a minimum 85C.

bob91343:
Theoretically, you don't need a buffer.  If the hot spot temperature is below the maximum in the data sheet, you are okay.

Practicaally speaking, one tends to leave some room for error or abuse.  That margin is up to the designer and there is no hard and fast rule.  And in fact, the maximum temperature isn't hard and fast either, and occasional excursions above it will only result in a statistical reduction in life span.

If you want long life, leave a bigger buffer.  At the molecular level, there is migration which will eventually cause a significant shift in characteristics.  If there is a buffer there as well, life will not be compromised.

Even a 120 Volt incandescent light bulb when operated at 100 Volts will eventually fail.  Minute impurities in manufacture may progress into major faults.  These failures have been documented very well and there is no escaping the eventual catastrophe.

So the answer to your question is, make your design as conservative as you feel is necessary to insure a reasonable life expectancy.  I know that isn't what you are asking but there is no definitive answer.

I have equipment that is nearly a century old and still works within specifications.  And I have fairly new stuff that has failed.  Looking at the design of these units one cannot pinpoint a reason for this.

Navigation

[0] Message Index

There was an error while thanking
Thanking...
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod