RF ground... large heatsink?

[Al (astral highway)]

Still on the drawing board...

So (on this drawing board) I have a large IGBT switching a very large current around 70-100Khz -- part of a self-resonant circuit after a few RF cycles.

The aluminium heatsink is relatively large and is obviously capacitatively coupled to the IGBT. Currents will also be induced into the heatsink inductively from other circuitry and a very large static charge may also build on it.

My instinct is that if I leave this heatsink floating (it's not external and won't be able to be touched), it will stress the IGBT, possibly to destruction.

From which I conclude that I should provide a path for RF currents to ground; ie, a capacitor of suitable voltage rating and low impedance to RF, between the heatsink and ground. This isn't, I'm thinking, an item that should be directly returned to earth. Have I got this right?

Thank you

[lesmw0sec]

I can't see how differing stress is likely to be caused by earthing or not earthing. I would have thought that a direct earth would be fine provided that your device has suitable insulation between it and the sink.
What is it for - causing RFI to nasty neighbours?

Les.

[Al (astral highway)]

Originally Posted by lesmw0sec

What is it for - causing RFI to nasty neighbours?

N. It will have the maximum possible anti-RFI measures, no effort spared

[jimmc101]

I would earth the heatsink directly, then if the insulation to the IGBT breaks down there will be no nasty voltages on the heatsink.

If you really want belt and braces add a foil screen between the IGBT and heatsink, insulated from both, and connect this to the IGBT's local earth.
This keeps the capacitive currents localised and keeps HF away from a safety earth on the heatsink.
The penalty is of course increased thermal resistance between the IGBT and heatsink.

Jim

[Ed_Dinning]

Hi Al, at one time there were heatsink washers available of sandwich construction with an extra isolated inner layer. On transistors the isolated layer was connected to the emitter and did an excellent job of removing the RFI and circulating currents that arose in the heatsink when it was earthed directly.

Ed

[Al (astral highway)]

Originally Posted by Ed_Dinning

...on transistors the isolated layer was connected to the emitter and did an excellent job of removing the RFI and circulating currents that arose in the heatsink when it was earthed directly.

Ah, that's exactly what I was worried about, Ed! Thanks for confirming my concern. I'll investigate further...Initial search suggests these things don't appear to be around any more however...