Wrong Sil pad?

I imagine the "really" good stuff tends to perform worse, for that reason.

Tim

You can screw it up like that, sure. But you don't have to.

Cheers

Phil Hobbs

Realistic. This is an engineering forum, not materials science.

The one time I tried to measure thermal conductivity, (3/8" brass 6-32 threaded standoffs. maybe 1/2" long. (I'd have to check notes... or do it again.)) I found that the thermal resistivity was ~x2 what I expected, from the brass piece. I assume surface resistance was to blame. More screw torque reduced it. I never tried any goop. I guess I should do that.

George H.

Maybe you could use thermal equivalent of a Kelvin connection to measure bulk thermal conductivity more accurately. Eliminate convective cooling (vacuum) and minimize radiative cooling.

Or put a known material thermally in series and do two differential temperature measurements.

--sp

The classic Dow thermal grease will squish down to below 100 microinches, the limit of my resolution.

If "thermally conductive" filler particles increase the gap, that may make things worse.

To get good thermal conduction, both surfaces have to be very flat. A lot of heat sinks aren't. Grease helps fill the air gaps that result from non-flat contacts.

I haven't. Don't they conduct heat best in the plane of the sheet?

My favorite thermal insulator is AlN, and after that, hard anodize. Both with grease.

Do you have a source for AlN sheets? I found this,

George H.

No, these are oriented in the 'Z' direction. They were here last week and dropped off dead trees. Interesting stuff but i haven't tried them, either.

If you don't need insulation, the best spacer is no spacer. Anything that you put into a gap adds theta, unless it is filling voids between rough surfaces.

We machine flat extruded (ie, wavy) heat sinks, and then hard anodize them, when theta really matters. And use more or bigger transistors to spread the heat around, avoid local hot spots.

That's the issue. Often it's not possible to put enough pressure on the part to make good contact, either.

Pricey! We use both cast and extruded heatsinks. They're pretty porous/grainy, too (cast, more so than extruded).

More pricey: Bolt big slabs of nickel-plated copper to the aluminum heat sink.

Yes, "tends" in the sense of, if the user merely goops some on, and doesn't take the time to fill and screed both surfaces carefully, so that the application is thick and nonuniform.

Tim

Hah, I drew this "stack-up" in my notebook, all to-220-pacs, I've drawn some as soldered together on the tab, I've never done that in practice. (Tin, clamp, heat and press. should work.)

Assci art's too hard for my crippled brain, from top to bottom,

heater(FET and resistor, fet on top) solder diode temp sens.

gap for stuff.

diode temp sens, solder brass standoff, 4-40 solder diode temp sens,

(another gap.. who cares?)

Heat sink/ thermal bath/ etc.

Two 4-40 screws clamping top and bottom bits together.

I've got three sensor's. I could add another and have another space for different stuff.

George H.

I think John's point is that the particles may be larger than the gap that would otherwise be there without them, making them counterproductive.

You still have to dump the heat to the environment eventually.

At what thickness does the 2.5 W/mK pale in significance to the 205 W/mK of the thickness of the heat sink? Somewhere around a few mils thickness of grease the resistance would be equal to the thermal resistance of the heat sink. The proper thickness of heat sink grease is far thinner than a few mils and not hard to apply. So it just doesn't matter much which heat sink grease you use unless you have a seriously kick-ass heat sink.

At 2.5 W/m/K, they'd have to be 100x larger than the gap to be a net negative. Don't think that's likely except in the case of extremely fat fingers. ;)

Cheers

Phil Hobbs