Thermal resistance of small SMDs versus Copper Pours

Den onsdag den 4. december 2013 12.02.50 UTC+1 skrev Klaus Kragelund:

maybe this is the one?

-Lasse

Yes, it certainly has very good information on the subject, thanks :-)

It's kind of strange that apparently no information (including thermography plots) of dumping power into 0805 and 1206 resistors exists.

Cheers

Klaus

We've done a bunch of work on this, keeping resistors cool with copper pours. You can push an 0603 to at least half a watt if you keep the end caps cool. The central hot-spot temperature of 1206 through 0603 resistors are about the same at equal dissipation if they are on big pours.

(note the delta-v cursor values on the first and last. I flubbed the middle one.)

To really cool a resistor, the pours should be on multiple layers, stitched by close-in thermal vias. Use as thick copper as you can, 1 or 2 oz.

There are AlN substrate resistors with astounding power ratings.

This is pretty good even though the resolution of figure 7 is poor:

If you have to push performance within a given real estate consider side-contact resistors for increased heat transfer:

That's a nice trick, using the RTD for both the power dissipation and measurement, will reveal the correct thermal resistance delay

Cheers

Klaus

Good stuff, thanks for the link :-)

Cheers

Klaus

You can figure out theta from those pics. I've done it a couple of times and never remember to save the numbers.

You may be looking for Nat Semi AN1025 for SOT23. AN1028 covers patterns for SOT-223 There was one for SOIC parts, but I seem to have mislocated it in electronic format.

Strangely enough, there are app notes with the same numbers from Fairchild.

RL

Also - check out IR AN994.

RL

AN1029 is SO8

RL

For a good time, google "Package and Thermal Information" to see some thermal images for different package patterns from Infineon.....

RL