countersinking a TO247 mosfet

mandag den 19. juni 2023 kl. 17.50.16 UTC+2 skrev John Larkin:

how low do you need?

The heights are good, the prices aren't!

A regular screw and a lockwasher cost me over 0.1", and I need less. With locktite, I can eliminate the lockwasher.

A really thin custom AlN insulator will help a bit. Isolated TO-247s are expensive and hard to get.

I could just epoxy the fets down, but that has downsides.

You are really good at parts research. Please email me.

mandag den 19. juni 2023 kl. 18.13.01 UTC+2 skrev John Larkin:

yeh, but now you konw they exist and they can surely be found much cheaper, I guess you want something like #4-40, but M3 seems to be more common

mandag den 19. juni 2023 kl. 18.13.01 UTC+2 skrev John Larkin:

maybe something like this?

I've done that

but it adds more height. I'll have boards in a card cage on 1.6" centers, and I'm stacking a CPU cooler/fan thing with giant mosfets. The screw heads could hit adjacent boards.

I could go to TO220s, which need less screw height.

Be careful about overtorqueing it, because stretching the metal might conceivably cause the die to crack or delaminate.

I'd definitely want to torture-test for that possibility.

Cheers

Phil Hobbs

That's the idea, as long as it doesn't destroy the silicon.

We'd need carbide countersinks. That mosfet epoxy is hell on tools.

mandag den 19. juni 2023 kl. 20.11.58 UTC+2 skrev John Larkin:

might be better with a counter bore instead of a counter sink

Yes, we could drop in a lockwasher.

I'm working from home today and I have a hand drill and a very rusty old countersink, a family heirloom.

This looks OK. The 4-40 screw should apply plenty of force and it centers the fet nicely. It would need locktite so thermal cycling doesn't loosen it up.

Stainless steel flat head hex socket screws with a 100-degree cone total angle are common, originally developed for use in airplanes.

.

Suitable 100-degree carbide countersinks are also available.

Joe Gwinn

Rivets?

A copper rivet could work, but there are issues with installing: the nose of the rivet setting tool may want to mash something else nearby...

The hole is 3.5mm, so an M3 flat head cap screw with a countersinked washer would hold it down; here are 1.8mm thick offerings for M3 screws. Is 1.8mm too high?

I'm not seeing any measure of the recess dimensions around the clear hole, though. The screw would be capable of touching the tab, which might not work for you.

Sounds risky, the wedge shape of the screw head will be working against the glass-filled resin of the body. I don't think that stuff is very strong in tension.

blind rivets tend to cause mechanical shock when the stem snaps leading to early failures. Swage rivets - maybe that could work if you can find a way to set them without damaging the MOSFET.

Rework would be difficult, and a rivet adds height too. The reason to not just epoxy the fets down would be the rework issue. And the mess, although there are some instant-set thermally conductive super-glue things.

That might work, depending on how deep I dare to countersink the mosfet. The ideal height of the screw is zero.

TO247s have insulated holes, so that's not a problem. No metal shows up on the one I countersunk.

I could use TO-220s, but they would need an insulating shoulder washer, and they are not as good thermally, having a smaller metal footprint than a TO247.

I'm trying to get maximum power dissipation with a bunch of mosfets on a K199 copper CPU cooler. 200 watts would be nice.