[Jeff Pahl]

Originally Posted by Al Skierkiewicz

Fredi,
At high current, I would be concerned with expansion and contraction coefficients of the material and the electrical components. I think the epoxy can do a great job as a heatsink when exposed to water. What I might suggest is a double dip such that the tabs of the transistors only are embedded in the epoxy while the circuitry remains open. Then encase the assembly as you have shown. This would allow the components on the board to move around with temperature without being broken off as material expands and contracts.

If you find the expansion / contraction to be an issue, another option is to pot the power devices with a layer of a thermally conductive RTV to provide a "cushion" layer between the packages and the epoxy. We use a material from NuSil at work that has a thermal conductivity of 0.81 W/mK to bond high power RF terminations to mounting surfaces. This is for a vacuum application where there is no other way to get the heat out besides conduction. The part number we use is CV-2960. I doubt it is cheap, but you could probably talk to them and get them to sample you enough to coat several units.