Upcoming Thermal Compound Shootout

[rwaritsdario]

Well folks, the poll is done. DaClan has debated and these are the pastes you'll be seeing in the review:
Intel stock TIM
Arctic Silver 5
Zalman ZM-STG1
Coollaboraty Liquid Pro
FrozenCPU Copper Thermal Compound
Arctic Silver Ceramique
Shin-Etsu X23 (we like to surprise you like that 8) )

Ninja has been recording everything, from humidity to airflow, from weight to how many times he craps a day. So now you'll be sure it was well done and done by the best. He has already tested about half the goops to give it some two weeks for the review to be finally published.
Thanks again to everyone who posted and gave us ideas and to all of our 99 voters.

 

[Doughbuy]

Where the fudge did you get the Shin-Etsu... this I have to check out...

Shoot, I was willing to say either the AS5 or the Liquid Pro will take first, but the Shin-Etsu might be the black horse...

 

[dasickninja]

Where the fudge did you get the Shin-Etsu... this I have to check out...

Shoot, I was willing to say either the AS5 or the Liquid Pro will take first, but the Shin-Etsu might be the black horse...

We operate like that. The Shin-Etsu, in my opinion, will take the lead in every test.

 

[Doughbuy]

Where did you find it!?!?!?!?!?!?

If it does take first, then I'm willing to pay a little extra for it.. however ya'll made it... unless it's as expensive per gram as crack...

 

[dasickninja]

Its a secret that will be kept till the review comes out. So give me time to finish testing.

 

[Doughbuy]

Meanie :x

 

[dasickninja]

Been called worse before.

 

[samir_nayanajaad]

well now that you got the stuff picked out when might we expect some results? Don't need an exact date just a general idea as to how long it might take will do fine.

<edit> ahh Im dumb found it in with the list of pastes you all picked </edit>

 

[little_scrapper]

I know you know that the flatness and finish of the coolers will affect the results. for example. I think Ceramiq will work better on a mirror finish surface. I bet Ceramiq could get 3-4 degrees better than AS5 if the surface is polished to a 600-800 grit finish because the particle size. AS5 would do better on a rougher surface because its particle size is larger. Some pastes are better for a gap fill. The problem it that using a single test case is not scientific.

As someone who has to keep extensive notes in a lab book on every experiment I conduct. It would bother me if you simply used one test sample (stock cooler) and then stamped your report as the new gold standard. Especially since it just makes sence that certian meterials will work best on certain finishes. And since you are doing this from the OC'er perspective, we definately will want to know where we can get every degree possible out of the CPU.

I say start with a cooler that has a nasty finish, run the tests, and progressively resurface it closer to mirror in 100-200grit finishes running the tests at each interval. Now that would be scientific. And even I would sign off on results like that. Assuming I had witnessed the finishing of the furface, and the application of material to cooler then cooler to CPU. Then I would sign off on it.

But then again sounds like you have already chosen what to do and are in process already. My $0.02

 

[dasickninja]

Errr, yeah. A bit late. Just a bit.


Edit: And just a note.... its not my test. Everyone that PM's me about it, aim your questions at Dario. It's his review.

 

[rwaritsdario]

Everyone that PM's me about it, aim your questions at Dario. It's his review.

Yeah, I dont bite ...soft :twisted:

 

[little_scrapper]

Shin-Etsu: Never heard of it, Can I get link?


is this the stuff or are there different varieties? I will keep looking.

http://www.crazypc.com/products/50118.html

 

[clue69less]

I say start with a cooler that has a nasty finish, run the tests, and progressively resurface it closer to mirror in 100-200grit finishes running the tests at each interval. Now that would be scientific.

Not quite. If you want to bring the surface prep aspect into the equation, then you have to measure the surface after each stage of prep. Many people that hand lap HS or IHS surfaces get those surfaces out of flat - which can be beneficial in small doses. So for example, one could use profilometry to get a measure of flatness and surface roughness. You'd want to lap the CPU IHS and the HSF and carefully look at the surface characteristics before each stage of testing. Adding these procedures would probably quadruple the amount of work and maybe increase the cost by 10X, but would add much insight. I have access to profilometry and am a metallography expert. Maybe we should see the results of this study, then pick three or four thermal compounds for a lapping study?

 

[SuperFly03]

I am looking forward to the results. Any ETA?

 

[Guest]

Now that would be time consuming but kick-4ss. I never saw this kind of review...

That would be really cool for the whole argument of: Lapping or /not and to what grit? You could see the diminishing return at some point.

Another little problem is the fact that the IHS gets more and more thin, which probably accounts for some of the thermal advantages. Actually not a problem but something to take into consideration.

I could see DaClan's shot at being famous 8).

This review could be up by 2009 when we will have carbon nanotube goop :lol:

 

[tool_462]

Buckminster-Fuller balls are the future of thermal compound! Maybe?

 

[clue69less]

Now that would be time consuming but kick-4ss. I never saw this kind of review...

That would be really cool for the whole argument of: Lapping or /not and to what grit? You could see the diminishing return at some point.

Another little problem is the fact that the IHS gets more and more thin, which probably accounts for some of the thermal advantages. Actually not a problem but something to take into consideration.

For sure. I'd also consider picking a CPU with high heat output and vary the load to get a plot of dissipation vs. load. For sure there will be diminishing returns, especially for people just learning to hand-lap. It's as much art as it is science. Automated machine lapping is the real way to go but then the cost would skyrocket again. I'm guessing that this kind of stuff has already been done by the likes of Thermalright, Zalman, etc.

 

[clue69less]

Buckminster-Fuller balls are the future of thermal compound! Maybe?

It would be an interesting material to test. I have not seen any thermal cycling studies when fullerines are in contact with air. But please, call them either Buckminsterfullerines or Bucky Balls. You can now actually purchase sputter guns that shoot Bucky Balls! That could be adapted to a video game weapon with panache. Blow Duke Nukem's nuts off with a Bucky Ball cannon!

 

[eckre]

Be sure to read this and review these as well:

http://www.dansdata.com/goop.htm

 

[cb62fcni]

I seem to recall reading that buckyballs are quite hazardous, they tend to cluster in the brain if inhaled/ingested. So I'd probably pass on thermal interface compound made from them, seems a little too risky...

 

[rwaritsdario]

Shin-Etsu: Never heard of it, Can I get link?


is this the stuff or are there different varieties? I will keep looking.

http://www.crazypc.com/products/50118.html

Yes that is the one. There are other variants but that is the newest one and the one that features the most thermal conductivity, also the runniest.

 

[rwaritsdario]

I am looking forward to the results. Any ETA?

Not sure how Ninja's doing, give it two weeks tops tough.

 

[rwaritsdario]

Be sure to read this and review these as well:

http://www.dansdata.com/goop.htm

The goops have already been picked and most of them have already been tested.
Is it actually because of that kind of outdated reviews that we are doing our own, with all the top contestants 8)

 

[little_scrapper]

Yea I never said anything about hand lapping, Again that would not be scientific. I am talking about taking the HS to a shop and have them progressively finish the surface.

 

[cb62fcni]

Oh crap, somebody already tested the thermal transfer properties of toothpaste? There goes my whole reason for living.