I am preparing a stress report on an old board which includes a transistor in a metal TO220 size package which is bolted directly to the heatplane of the board with no thermal pad or grease. I have a figure for the junction to case thermal impedance from the device datasheet, but I am unable to find anything for the thermal impedance of the case - heatplane interface. Can anyone suggest what I should use for this please, based on a pressure of about 50psi as used by Bergquist in their comparison of different Sil-pad products?

Hi Tiptonian

It is hard to put a figure on this as there are many things which can affect heat transfer in this kind of situation.
At least it is not riveted! - ala exports of a bygone era!

If a TO220 is overtightened enough to distort the plated copper lug (surprisingly easily) then the thermal resistance is increased since there is less contact with the heatplane (unless it distorts in exactly the same way which it usually doesn't).

Also any kind of corrosion can significantly reduce heat flow - especially if it is an old board.

As for a value - hmmm !?!?

I guess there is no easy way of remounting the TO220 with either: thermal pad, transfer compound, thermal adhesive, or thermally conductive phase change material? You could then use datasheet values for these mounts e.g. thermally conductive phase change material of 0.001" thickness at 50psi for TO220 = 0.92C/W.

Sorry not to be of more help.
Regards
E.

-------------------------
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The meaning of life is a blank sheet; write on it wisely. ~ M.Cutler.