The Official Cooling Thread!

im1992,

Thanks! Now I am spoiled! LOL!!

Russ

 

russ,
LOL, definitely, 5 minutes must be a world record!
I almost got a mirror finish which isn't good as we discussed earlier.
I think I will sand it with a lower grit just to get some scratch marks....
-im1992

 

why wouldnt you want a mirror finish?

 

why wouldnt you want a mirror finish?

 

im1992,

It sounds about right to me! I buff them until I get a dull shine, not a mirror finish. Although the Enzotech NB cooler has a mirror finish and it seems to do a great job of keeping the heat away from the cpu. Maybe it's different because it's one piece forged Copper I used a small acrylic artist's brush and painted a light coat of MX-2 on each surface and it seems to work real well. The theory being that there has to be total contact when you stick the two together as an air pocket will just be forced out but the surfaces will still have MX-2 on them. Seems to work!

Best Regards,
Russ

 

I think if you get a perfect finish the grooves are so tiny they can't hold the drops of the thermal paste so it sits on top.

 

thermal paste is not an adhesive. thats why there are pushpins and back plates with screws.


the mirror finish gives ALOT less imperfections on the Heatspreader.

ofcourse, there are still bumps, as the atoms are arranged in a lattice in a solid, and atoms are rounded, NOT square, so yes, there will be bumps, but this strips most of those atoms that have been OXIDIZED. oxidization, causes the Cu/Al to become LESS effective at transferring heat, therefore a mirror finish is ideal. the more Cu/ Al Surface area as possible to touch the better heat dissipation.

and with a smaller bump, less TP is needed anyway. the TP will eb there for the Bigger bumps, the smaller ones will eb so small, there is no need to try and fill them, as it will be only a few mols space.

 

I know thermal paste isn't adhesive (well not normally, you can get thermal adhesive), but why should that make any difference?

 

Russ,

what did you use to get it flat? using a round ragwheel and jewelers rouge will only polish. it will not make it flat and flatening is the reason behind lapping. just because it is shiny by no means means it is flat.

 

yeah, i realised, after i posted it i realised you hasdent said, it would be harder for the TP to stick to the HS. (i dunno where i got that from (slaps head))

and mort, if ti is shiney, there is less almost no oxidization on the surface, so haveing it shiney is as important.

obviously if its not flat, one is screwed, but if they can shape the IHS to that shape, all shoudl be fine

 

thats exactly the reason i want to get rid of the mirror finish
thermal paste wont be able to stay in there...just for testing sake, i will try it with both the mirror ginish as well as the non mirror finish.
-im1992

 

m8, TP doesnt stick in the first place. it will stay in there. dont worry about that

why do MANY CPU/GPU coolers come with mirror finishes now?

 

Are they really that mirror? I don't think I've seen one that impresses me that much yet.

 

the older ones didnt, but from this year they have.

like the new sunbeam, and new noctuas.

the duoorb. etc.

 

Sunbeam made a CPU cooler? lol I was genuinely unaware of that! Still, apart from the Rheobus I don't remember being wowed by the quality of Sunbeam stuff.

 

Sunbeam is who makes/made the Tuniq Tower. I think it just goes under Tuniq now. Sunbeam does make other cpu hsf's too.

 

Ah right, fair enough. They're obviously pretty good at it then, the Tuniq's a respectable cooler.

 

dont forget the core-contact freezer :
http://www.tweaktown.com/reviews/1429/sunbeamtech_core_contact_freezer_cooler/index.html

check out the new king (sort of)

http://www.tweaktown.com/reviews/1439/4/noctua_nh_c12p_cpu_cooler/index.html

 

Those reviews are awful. Not only have they missed three of the big players in theheatsink market (The TRUE, Scythe Ninja and Tuniq Tower) but their acoustic readings are propostrous, and where's the testing methodology? The results are meaningless unless they tell you at the very least what CPU they're using.

 

Mort81,

I've never really had a problem with flatness except once. That was a Scythe Katana or something like that and that was a while ago. It had a high spot which I sanded first. Usually despite using a round wheel, you can work any minor imperfection in a given area out, like a high spot. I have a machined steel block that I use to check overall flatness along with Articulating paper which will show a contact pattern like you get with the thermal paste, only without the mess! The Arctic coolers I use are very flat to begin with. I also use a precision straightedge. It's much easier when you are used to the very fine tolerances that I work with every day. If I ever needed to, I have a precision milling machine at the shop that can handle up to about a 2"x2" wide surface. Actually I don't even know if it's a milling machine or not, but it works somewhat like a router or flycutter. The work is mounted in a jig and the rotating head makes passes over the surface in a pattern! All computer and laser controlled. Not your average machine shop, that's for sure! Then again, not many machine shops need the kind of precision it takes to build miniature air turbines that have to stay together at 400,000+ rpms!

Best Regards,
Russ