I see where there are 50 Ohm high power (500 W and more) that look like thick film hybrid microcircuits. has anybody had direct experience with these and can cast light on heat sink requirements
Yes.
The heatsink requirements will be exactly the same as heatsinks for semiconductors.
If you get the downloaded catalog from a heatsink mfr, like Wakefield, they will have
a complete section with the equations and examples, graphs, of the thermal considerations
for heatsinks. Probably on some websites as calculators and pdfs these days.
I can say from experience, IF you have a small surface area that is supposed to dissipate
a
continuous amount of heat - in this case 500 watts - you will need a very low
thermal resistance pathway. Probably natural convection will be insufficient, even with
a very large heatsink - although one would have to run the numbers, and maybe try
a prototype test.
In general, you can take away from the solid state linear guys that you will want a
"copper spreader" between the device and the aluminum heatsink. Flatness counts too.
The copper spreader does just that, with higher thermal conductivity than aluminum, it
serves as a way to make the effective surface area of the device being cooled larger,
so the heatsink works closer to ideal.
The good news is that a reasonably small flow of air is surprisingly more effective than
natural convection. Also natural convection works only if the heatsink fins are properly
oriented and of proper height and spacing. Some heatsinks are actually designed to be
"blown" and do not work as well as one might imagine just looking at them.
_-_-bear
