estimating junction temperature of a power mosfet

Thanks for those links. They are quite interesting.

Thanks to all those who responded. I decided initially to limit the power of the electronic load to 300W (two fets in //), which seems safe. Later I will make an experimental setup to measure the parasitic diode drop at various temperatures, and see if I can safely dissipate 400 to 450W. For now, I am fighting against parasitic thermo-electric effect of sense resistors, which seriously compromises the precision of intensity measurement and regulation athigh power dissipation.

-- Alain Coste

m9oq8a$sfu$ snipped-for-privacy@dont-email.me...

--- L'absence de virus dans ce courrier electronique a ete verifiee par le logiciel antivirus Avast.

Manganin has a low tc and a low thermoelectric potential against copper. Zeranin is even better.

We make our own manganin shunts, which we epoxy to an anodized aluminum block which is itself thermally regulated to sit at the flat spot of the manganin's parabolic resistance/temperature curve. That's extreme, but heat sinking a strip of manganin can make an awfully good shunt resistor.

What's important is thermal symmetry of the two shunt-to-copper junctions. With modern chopper opamps, you don't need a lot of voltage to work with, which helps keep the shunt heating down. Interesting geometries are possible.

You can buy thru-hole and surface mount low TC shunt resistors, which are fine if the power dissipation is low.