Arctic silver thermal resistance

[Oznog]

Any way to guesstimate the pad-to-sink thermal resistance of a Luxeon Rebel
glued with arctic silver adhesive to a piece of brush-finish aluminum?

I tried to do a dV/dTemp measurement awhile back, but wasn't able to get a meaningful number I think. The likely problem was in catching the voltage at the "cold" die temp at the specified current, the die itself has little thermal mass and warms up to its new temp pretty fast making it difficult to measure.

 

[frenzee]

Much as I like Arctic Silver, I have my doubts about how well it conducts heat in real life situations becasue of imperfect interfacing. AS's specs are given here, but that's in a theoretical perfect-world situation and doesn't address how microscopic gaps would affect the K value.

I propose an experiment using identical emitters and heatsinks one attached with AS, one soldered and plotting the delta T over time. (a little like Newbie's experiment with polished vs. black aluminum heatsinks). If I had two micro thermocouples, I would do the test with Rebels. I have everything else.

 

[Oznog]

Yeah Arctic gives thermal resistance for the grease in terms of deg C/W per sq in, but not for the epoxy. I wonder why?

It might make more sense to use a standardized interface test... like just a bolt-on plate with resistor and thermistor. This may seem less applicable to the specific question but the difficulty is in getting an accurate die temp measurement on the Rebel.

On my latest project I tried throwing the piece with epoxy on it into the vacuum chamber and watched the air bubble out. That kinda makes sense, doesn't it? Unfortunately even though I'd mixed the epoxy just out the fridge and expected it to be slow I was pretty lazy on working this out, took a few extra minutes and it started to gel up on me. That's stuff's quick! The vacuumed material bubbled and left craters that didn't self-level before I added the Rebels. I'm uncertain how good the bond came out.

 

[HumanLumen]

With respect to gas bubbles, is it a case that the epoxy needs careful; application, or does the product liberate gas when it is mixed? I was about to AS a q5 cree onto aluminium and this thread has now cought my interest.

HL

 

[Oznog]

Epoxy does not create gas.

When 2-part epoxy is mixed, small bubbles of air are whipped into the mix. Small bubbles do not float to the surface and leave, the liquid is too viscous. And when applying it to a brushed aluminum surface like mine, there may be tiny air bubbles at the interface where the AS doesn't flow into the pores of the metal. These air voids would increase the thermal resistance.

Honestly I have no data to say they're all that significant. You've probably noticed bubbles in say the clear 5-min epoxy and there are not a lot of big ones or anything. It was just an idea. Yes I did get plenty of bubbles as it boiled out in the vac chamber but that chamber can suck bubbles out of just about anything.

 

[VanIsleDSM]

I haven't used the arctic silver epoxy, I bought MG chemicals thermal epoxy and potting compound.. It's aerospace quality and NASA uses it.. It's not viscous at all, and the instructions call to leave it sit once mixed for 30min to let any bubbles come out.. it has the consistency of warm cooking oil or so maybe less.. it flows very easily.. takes a full 24 hours to cure with no heat.. or about 1 hour at 65C.. awesome stuff so far.. damn expensive though.