lapped my ultra-120 extreme (pics and temp results)

[luvclub]

Ahhh, but what if the heatsink and processor were equally rough and that both surfaces were able to connect like interlocking teeth?

Heatsink
/\/\/\/\/\/\/\/\
/\/\/\/\/\/\/\/\
Processor

Wouldn't this provide the greatest surface area?

I suppose this is impossible to do in microscopic terms but what if they were like instead matching grooves/teeth on both surfaces, say 1mm width.

[Andy3536]

Ahhh, but what if the heatsink and processor were equally rough and that both surfaces were able to connect like interlocking teeth?

Heatsink
/\/\/\/\/\/\/\/\
/\/\/\/\/\/\/\/\
Processor

Wouldn't this provide the greatest surface area?

I suppose this is impossible to do in microscopic terms but what if they were like instead matching grooves/teeth on both surfaces, say 1mm width.

Only if they were a perfact match
Mabee make them out of gold aswell!

[Phil_P]

Ahhh, but what if the heatsink and processor were equally rough and that both surfaces were able to connect like interlocking teeth?

Heatsink
/\/\/\/\/\/\/\/\
/\/\/\/\/\/\/\/\
Processor

Wouldn't this provide the greatest surface area?

I suppose this is impossible to do in microscopic terms but what if they were like instead matching grooves/teeth on both surfaces, say 1mm width.

Yes, that would massively increase the contact area. Difficult to engineer though and if they didn't mate properly, who would you blame - the cooler manufacturer or the CPU manufacturer. I'm sure each would blame the other and refuse an RMA

Maybe this is the way forward if CPUs continue the current trend of getting hotter and hotter. LOL, I remember building an i486-25SX that didn't even have a heatsink!

[Kumagoro]

The thermal paste would have a higher surface area in contact with the rough heatsink, but a lapped surface would have a higher metal to metal contact which would transfer heat far more efficiently. Hence why the lapping shows an obvious improvement.

You want to use as little thermal paste as possible by massaging a tiny bit into the heat sink and die shield. cling film
and finger I find work best.

[Phil_P]

Here's a thought - how about a very thin sheet of soft copper foil between the two - the idea being that it compresses into the grooves or imperfections of the two surfaces under the pressure of the clamping system.

[Clunk]

But then, wouldnt you have 4 surfaces that need thermal paste?

[Flewis]

How about a blob of mercury between the cpu and heatsink. Surely its metallic properties would make it a good thermal conductor and being a liquid would get in all the gaps. Shame its toxic tho.

[Andy3536]

Here's a thought - how about a very thin sheet of soft copper foil between the two - the idea being that it compresses into the grooves or imperfections of the two surfaces under the pressure of the clamping system.

copper is not a soft metal, if you wanted to go that way you would need a soft metal like magnesium, but that would probobly still leave air pockets, the best thing you can do is to rough up the 2 surfaces with a piece of wet & dry fine grade and use the best paste (normally a silver one)

[Kumagoro]

Can the Die shield (what the hell is the proper name for it) be popped of reasonably easy.

graysky how about seeing what temps you get after doing that.

[graysky]

@ Kumagoro - you're talking about removing the IHS (internal heat spreader) which I have read about, but aren't interested in trying myself. I think lapping is pretty safe, but when you start poking sensitive CPUs with screwdrivers... that's a whole different story.

[[GSV]Myocardial]

There was a discussion paper written (that handily enough I can't now find) but the gist of it was that if you had good TIM and a micro-grooved heatsink then you could increase the available area for heat transmission by a significant amount if the TIM could be made to be forced into the grooves. Given that Thermalright are pretty good at these CPU coolers, I find it hard to believe they would leave the base of $65 cooler unfinished like that unintentionally. When mine arrives I'm going to give it a month to let the TIM set properly and then see how it compares.

Interesting post though.

I think I've read the same thing, they were using lasers to cut 2 mating surfaces and in doing so they would actually reduce the thermal transmission of heat when the surfaces where both dry mounted and mounted with a compound. They found that a slightly rough surface would conduct the heat far better. As Andy3536 sais it's to do with surface area.

[Phil_P]

Getting sort of back to the original topic, for us it's not so much if the surface(s) are smooth or rough, but rather if they are flat or concave/convex. So long as the two surfaces are relatively flat, I don't think it overly matters if they are rough or smooth in their finish.

[tigerboyce]

just use a fridge as your case

But, anyway yes it is definetly worth lapping to get the few degrees less

[Andy3536]

Getting sort of back to the original topic, for us it's not so much if the surface(s) are smooth or rough, but rather if they are flat or concave/convex. So long as the two surfaces are relatively flat, I don't think it overly matters if they are rough or smooth in their finish.

If they are smooth then you would loose 2deg easy, the genral idea with lapping is to use fine wet and dry paper to create lots of tiny scratches (i.e. rough) it is doing that that will keep the paste in place and create a larger surface area.