Best Thermal paste

[TheMawn]

I've heard somewhere of people really rubbing the thermal paste into the heatsink and heat spreader and taking a super straight edge of some sort to scrape off excess paste from both surfaces and then just sticking the heatsink and chip together after that.

If your edge is straight enough, you SHOULD have just an ever so slightly soft and perfectly flat surface. A decent substitution would be just the tiniest little drop of paste to compensate for any non-flatness.

If I get really crazy, I'll try applying the paste in a hot environment (we actually own a small electric sauna, believe it or not) so it gets softer and easier to spread thinly...

[xilman]

Originally Posted by lavalamp

No, for best thermal conductivity, what you need is liquid Helium cooled to 1.8 K or lower.

That's not at all obvious to me. Perhaps smoothing the two surfaces to atomic precision would remove the need for an interface altogether. If not, a single layer of graphene might suffice.

[xilman]

Originally Posted by TheMawn

(we actually own a small electric sauna, believe it or not)

What's so incredible about that? I also have an IR sauna, referred to in our household as a heated walk-in double wardrobe. Seats two in comfort and has a decent music system installed. Unfortunately, it's not quite big enough to lie down in properly.

[TheMawn]

A sauna is just not something I would expect many people to have. Not my idea, mind.

EDIT: Also, in the lattice direction perpendicular to the sheets, thermal conductivity is borderline bad. The conductivities (electrical does this too) are much better along the sheets than across them. I'm hoping for carbon-based microchips. I think it would be cool to use diamond as the semi-conductor and graphene as the conductor.

"Running my i7-8770k at stock settings. 0.560V, 150GHz and about 275C under full load"

"Dood you haz so much OC headroom diamond only turns to graphite at 700C OC dat shiz"

Could be fun, no?

[kracker]

I just got some AS5. We'll see how it goes. First though is replacing my GPU default paste...

[kladner]

Really clean is important for all TIMs. Arctic Silver instructions online make a point of the size of gap a human hair, or even a fingerprint causes. Isopropyl alcohol, the higher percent the better, is a minimum. I used to start with acetone, and end with isopropyl 91%. At the moment I have some Arctic Silver cleaner and surface prep solutions. Some people really push tinting the surfaces: rubbing some of the compound into the surfaces with a glove, or plastic wrap on your finger, then wiping most of it off with lintless cloth or paper. This is supposed to shorten the curing period by pre-filling the irregularities.

[kracker]

Originally Posted by kladner

Really clean is important for all TIMs. Arctic Silver instructions online make a point of the size of gap a human hair, or even a fingerprint causes. Isopropyl alcohol, the higher percent the better, is a minimum. I used to start with acetone, and end with isopropyl 91%. At the moment I have some Arctic Silver cleaner and surface prep solutions. Some people really push tinting the surfaces: rubbing some of the compound into the surfaces with a glove, or plastic wrap on your finger, then wiping most of it off with lintless cloth or paper. This is supposed to shorten the curing period by pre-filling the irregularities.

If you mean this, I got one. Not sure how well it works though, alcohol never did "completely" remove the TIM, what I mean is it was always very slightly foggy.

[Batalov]

...behold the religious rituals of Paste-farians.

[kracker]

Or maybe I should have gotten TX-4! it is 2C lower average temperatures!!!

(ref. above post)

[Uncwilly]

No one is suggesting lapping the CPU and heatsink?

[kladner]

Originally Posted by Uncwilly

No one is suggesting lapping the CPU and heatsink?

Been there long ago. Done right it can give good results, but doing it right is a challenge. I've seen people here with plates designed for lapping. I was using a sheet of glass and progressively finer Wet-or-Dry sandpaper. It was an interesting experience, but I haven't been motivated enough lately to seriously consider it.

@kracker- I'm sure you know that kind of "what if" game turns into endless tail chasing. I'm obviously susceptible to it. I return to the idea that any of the better products, well applied, will do a good job. I've seen too many roundups of TIMs which disagreed with each other to think otherwise. It does seem that the phase change metal is a different beast from pastes, if you have the guts to get it right. I think it might be similar in behavior to the solder under the heat spreader on Sandy Bridge chips.