thermal paste

[Mitternacht]

I got a new mod in store, and I gotta know; what's the best thermal paste out there?

[J-Roc]

i was recently recomended arctic silver MX-2. Apparently its in the "best of the best" range. Im still waiting for mine in the mail so i cant give a %100 on it yet.

[Helix666]

There's a thermal paste out there that (from the reviews) beats the Arctic Silver MX-2 range...
OCZ Freeze review (http://www.tweaktown.com/reviews/1265/ocz_freeze_thermal_interface_material/index.html)

[Spawn-Inc]

Arctic Silver 5 is the best right now. who ever did that review for the ocz stuff should have included Arctic Silver 5 in the tests...

[mtekk]

There is some of that crazy Liquid Silver/Metal stuff out now that is supposed to be better than Arctic Silver 5 (the high-end standard for some time now), however it is not compatible with aluminum, so you better have a copper base to your heatsink if you are going to use it.

As a side question of speculation, now that the competitors are beating Arctic Silver now, when will we see something new from Arctic Silver (e.g. Arctic Silver 6)?

[Spawn-Inc]

oh i forgot to say there is new stuff that uses powdered diamonds!

the product is called IC Diamond 7 Carat, article here (http://www.innovationcooling.com/article1.htm) and here (http://www.ocforums.com/showthread.php?t=538919&page=1) are a bunch of home done tests on it.

[mtekk]

I remember Inventgeek (http://www.inventgeek.com/Projects/diamondgrease/Page1.aspx) had a how-to to make the diamond thermal paste, I didn't think it was available on the market yet.

[crenn]

I've been reading about that diamond thermal paste. It seems it's actually hit and miss. Some are getting worse temps with it and others are getting 1-3c change. There were a few cases of massive changes however it may have been due to bad application of the TIM. When I can, I'll be doing a special piece to applying AS5.... and how temps can vary.

[.Maleficus.]

Arctic Silver 5 is the best right now. who ever did that review for the ocz stuff should have included Arctic Silver 5 in the tests...
They didn't actually test it in THAT test, but they do say that all of those tested do in fact perform better than AS5.


It has a maximum load temperature of 51C and averages just a bit higher than the other products, all of which still manage to beat out the Arctic Silver. If you are still using the AS goop, it really is time to change.


AS5 looks like "ASS". I just realized that.

[Ap1thy]

ive read some positive reviews of tuniq tx-2, picked up a tube for stuff like chipsets where i dont want to worry about something shorting but apparently it works as well as as5

[mtekk]

I've been reading about that diamond thermal paste. It seems it's actually hit and miss. Some are getting worse temps with it and others are getting 1-3c change. There were a few cases of massive changes however it may have been due to bad application of the TIM. When I can, I'll be doing a special piece to applying AS5.... and how temps can vary.

I have a feeling people are getting different results due to their methods. As you said it is all in the application of the paste (Look at the picture in this review for Tuniq TX-2 (http://www.extremeoverclocking.com/reviews/cooling/Tuniq_TX-2_Thermal_Compound_3.html), I'm pretty sure they used way too much compound), and how they are measuring temperature is important as well. Every time I see an infrared heat gun pointing on a heatsink I shudder. The reason is, if your thermal paste is insulating rather than conducting the heat then the heatsink will be cooler, so a good conducting thermal paste should cause your heatsink temperatures to increase. What really matters is the temperature that die is at, I think that is what the CPU reports to the bios, and we want that to be low. The thing is I don't think that is too accurate either for absolute temperature (at least some motherboards are known to have issues with reporting inaccurate temperatures).

I still use AS5 as my little syringe of it is still half full. I also have the ArcticSilver epoxy which works wonders when you need to permanently attach a heatsink to something such as a electronic speed control for an RC car.

[crenn]

I have a 12g tube of AS5, and a couple of degrees hotter is something I can handle.

Depends how you define too much compound, too little affects performance, too much affects performance.... but it could be done right. I'm going to be doing something soon about it. Different methods of applying, different results.

Most BIOSes can't handle come processor's T-Junctions of 100c instead of the old 85c.

[J-Roc]

MTEKK, on that note, how much paste should one use? When i installed my termaltake block, i used about as much paste as they did for that tuniq review. Allthough, while browsing, i commonly see a miniscule amount of paste.

[Spawn-Inc]

everyone does it different, here (http://www.arcticsilver.com/pdf/appinstruct/as5/ins_as5_intel_quad_wcap.pdf) is what arctic silver says for quads, and here (http://www.arcticsilver.com/pdf/appinstruct/as5/ins_as5_intel_dual_wcap.pdf) for duals.

the way i do all my stuff is put maybe a grain of rice then push it all around until it is completely covered. it's to much paste but i still get good temps.

[J-Roc]

Good to know, im going to install my d-tek blocks soon and since i paid top dollar i'd like to get top performance. I also can tell that i have an insane amount of thermal paste on my current block. :) live and learn...

[mtekk]

Sometimes I get systems (I do computer upgrades for some people that I know) that have so much thermal grease on them that the CPU has to be slid off of the heatsink, and that is way too much. ;)

As long as that doesn't happen to you you are within in an acceptable range. I like to go almost as thin as possible, just enough that the entire core or heat spreader can not be seen through the AS5. I usually follow the guides by Arctic Silver, Spawn-Inc linked to two of their newer ones, and use a plastic sandwich bag and my finger to apply AC5 (I would use latex gloves but the dust they have would probably have adverse effects).

[crenn]

People say spreading the AS5 over the HSF first before installing the HSF can have a good effect on temps. I've seen it happen, 10c cooler from just applying it differently

[Scotty]

I use MX-2. Great stuff, id say its one of the better ones out there.

From reviews i have seen it's better than AS5, but by 2*C at the max.

I got it over AS5 because it was cheaper and as i was layering it onto my graphics card i didn't want any electricity to conduct to places it shouldn't.

[crenn]

AS5 doesn't conduct. It's designed not to conduct.

[mtekk]

True, AC5 doesn't conduct. However, it has capacitance properties, which can be almost as bad, probably causing instability verses frying components.