I found Keratherm to be very good. But, I wet sand off the heatsink surface and transistor surface to be completely smooth and use enough but not too much torque to squeeze the Keratherm into micro-imperfections... if any are left... after the polishing process. A wide flat washer (to cover the whole transistor faceplate) and an elastic lock split washer on top, are also required here. Works exceptionally well.... better than mica and paste.
The problem with Keratherm is the price...
The problem with Keratherm is the price...
Ah. Yes. I was comparing mica vs Keratherm vs Silpad and neglected the various ceramics, AIN, whatnot.
The 0.25 mm Keratherm 86/82 has a dielectric breakdown strength of 1 kV (4 kV/mm).
Tom
The 0.25 mm Keratherm 86/82 has a dielectric breakdown strength of 1 kV (4 kV/mm).
Tom
Last edited:
Foremost I am aware of BeO material and even have to-3 leftovers from good old times. Good luck to get one.
Let's say we use phase change grease, basically paraffines. Use of that solves the "thickness problem", answering your second question. The PC material referred to is aawid ultrastik which does not contain any particulate matter having thermal impedance is 0.03 C-in sq/W @ 20 psi.
Flatness (the desirable level) is 50 microns per 100 mm for a ceramic based modules & 0.3mm ceramic baseplate. Fisher Elektonik standard products have 20 micron per 20 mm flatness. So-so ceramic substrates are polished to 0.2 microns, good ones to 0.01. Check Infineon/Cree/Semikron app notes. Exemptions prove the rule.
Sure thing one can't access heatsinks with mating surfaces as stated without an effort, leaving the one and only way to go, namely Keraterm which can deal with "not so good" roughness and, to some degree, flatness, unless you need real HV insulation.
Sincerely,
Alex
Just an afterthought: some distributors sell heatsink profiles "as drawn" and some - "milled". The latter is usually quite good.
I have had great results with these insulators:
https://www.mouser.com/ProductDetail/532-4180
And this thermal paste:
https://www.amazon.com/gp/product/B07LDK4F5R/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&psc=1
Also, the ceramic insulators are reuseable. I don't even clean them off when moving them to a new project. Just add more thermal paste.
The thermal delta between the mosfet case and the heatsink directly next to the pad is low when compared to mica. Also, I had a couple mica pads fail which meant the voltage supply went from a mosfet right to the heatsink. Always a good idea to ohm check the center pin to the heatsink.
https://www.mouser.com/ProductDetail/532-4180
And this thermal paste:
https://www.amazon.com/gp/product/B07LDK4F5R/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&psc=1
Also, the ceramic insulators are reuseable. I don't even clean them off when moving them to a new project. Just add more thermal paste.
The thermal delta between the mosfet case and the heatsink directly next to the pad is low when compared to mica. Also, I had a couple mica pads fail which meant the voltage supply went from a mosfet right to the heatsink. Always a good idea to ohm check the center pin to the heatsink.
After reading all I could find including this thread, links on this thread and other sources this was pretty much my
exact conclusion. I already have Arctic Silver MX-6 I used on the copper heat spreader on my amateur radio amp.
The ceramic insulators are on order. The amp is an Aleph J and the chassis is the big 5U from the modushop so
I don't foresee any heating issues. I'd like to thank everyone for their input and expertise!
Best regards, Mark
exact conclusion. I already have Arctic Silver MX-6 I used on the copper heat spreader on my amateur radio amp.
The ceramic insulators are on order. The amp is an Aleph J and the chassis is the big 5U from the modushop so
I don't foresee any heating issues. I'd like to thank everyone for their input and expertise!
Best regards, Mark