Sign in with
Sign up | Sign in
Your question

Help me pick new thermal paste

Last response: in Overclocking
Share
September 13, 2011 10:22:19 PM

Currently running an i7 920 with Noctua D14 and been using AC5 for a couple years on it. I would like to lap both my HS and CPU and use a new paste for some lower temps. I was thinking about Coollaborotory Liquid Ultra, ICD, and indigo extreme. Are these in the top runnings for lowest temps? Can i use the Liquid Ultra on the noctua heatsink? im thinking no since my heatsink is alluminum correct?

What would you guys say is absolutely the best for the lowest temps in September? granted everyone has their own opinion on what they have tried but just looking for your thoughts.

thanks

More about : pick thermal paste

a b K Overclocking
September 13, 2011 10:35:02 PM

There's really only a few degrees difference when comparing the best thermal compound. Currently I like the MX pastes. They do a good job, and aren't electrically conductive. Not a big deal if you're neat and careful, but an added plus.
September 13, 2011 10:50:56 PM

Seems kinda risky to lap a processor you've had for years. You will need at least a granite surface or a flat stone to do it, I don't think you'll get the noctura any flatter, but some processors are sometimes not quite flat. The easiest way to tell is if one core has temps a few degrees above the others.
Related resources
September 14, 2011 3:54:21 AM

Ive lapped before and had good results, Real temp reports 3 cores around 53 oc to 4.5 while core 4 is at 48. I would like to drop my temps at least -5 degrees. the D14 is a good heatsink all around but huge..anyone recommend a better or even HS?
a c 197 K Overclocking
September 14, 2011 6:34:28 AM

vigilante212 said:
Seems kinda risky to lap a processor you've had for years. You will need at least a granite surface or a flat stone to do it, ...

Piece of plate glass.
a c 206 K Overclocking
September 14, 2011 11:04:33 AM

Both contact surfaces lapped require very little thermal interface, so I recommend Cool Laboratory Liquid Ultra, it's a warranty void product but that's a non issue under these circumstances. It is electrically conductive so be careful with it, it spreads very smooth comes with a brush applicator.

If you don't want to go that route and would rather use a non electrically conductive TIM, I recommend Artic Cooling MX4, it's a thinner consistency than the other top line TIMs and will clamp down to a very thin layer, perfect for lapped surfaces.
a c 206 K Overclocking
September 14, 2011 6:44:30 PM

DelroyMonjo said:
http://www.hardwaresecrets.com/article/Thermal-Compound...
Prolimitech seems to be in the lead, but most are within 3C.


Those tests are totally useless information, they allowed room temperature variances of 12c to 26c, clearly stated in their first chart, unless the room temperature is the same for all tests across the board the results are seriously inaccurate.

If you want solid factual numbers check my Thermal Roundup out in my sig, all the tests were performed exactly the same, same room temperature for all tests.
a b K Overclocking
September 14, 2011 8:03:18 PM

32C to 39C is a 7 degree difference between room temp and core temp regardless of the ambient temp in the hardware secrets test. Yes I noticed the ambient temp differences. In your tests, 4ryan, we see a 9C difference from different heat sink compounds even thought the ambient temp is the same for all your tests. Which IMHO should be the ONLY way anything should be tested.
I looked for a thermal compound chart on Toms but couldn't find it.
Sorry, I didn't mean to'diss you or Tom's Hardware.
September 14, 2011 8:08:16 PM

The liquid ulta cant be used with aluminum though. From what I have been looking at seems like the IC Diamond works the best, has only an hour or 2 of break in. I am a fan of AS5, but that's old tech they haven't changed in almost 6 years..


JSC glass may or may not be flat no guarantee. were talking about tolerances of less than .001 inches or a third the thickness of a human hair.
a b K Overclocking
September 14, 2011 9:17:46 PM

4Ryan6 said:
Those tests are totally useless information, they allowed room temperature variances of 12c to 26c, clearly stated in their first chart, unless the room temperature is the same for all tests across the board the results are seriously inaccurate.

If you want solid factual numbers check my Thermal Roundup out in my sig, all the tests were performed exactly the same, same room temperature for all tests.


It's always preferred to test with the ambient stable but the results will be 99% the same if you use the delta's so the data is still useful. I've tested both ways so I know this for a fact and you can test it yourself if in doubt.

If you had a change in ambient temps. of 30-50C yes then the data would be skewed and the higher ambient temps. would lower the HSF efficiency and make the TIM test data ~1-2C higher than if tested at the same ambient. That was not the case however with the Hardware Secrets testing. They had a max of 14C diff in ambient.

The diff between the top TIMs is 1-3C and that's all that really matters, not the absolute temps. Even your own test data supports this reality. ;)  The diff between AS5 and Liquid Ultra is ~2C.

BTW - I have bookmarked your TIM test so that I can refer folks to it as it shows pretty much the same 1-3C spread as everyone else's test data which is good confirmation that reviewers are getting similar results unlike one site where I've seen a claim of 10C from two TIMs from the same company that everyone else measures at 1C diff.
a b K Overclocking
September 14, 2011 9:26:31 PM

DelroyMonjo said:
32C to 39C is a 7 degree difference between room temp and core temp regardless of the ambient temp in the hardware secrets test. Yes I noticed the ambient temp differences. In your tests, 4ryan, we see a 9C difference from different heat sink compounds even thought the ambient temp is the same for all your tests. Which IMHO should be the ONLY way anything should be tested.
I looked for a thermal compound chart on Toms but couldn't find it.
Sorry, I didn't mean to'diss you or Tom's Hardware.



The bottom three TIMs would not be used by any serious PC builder. The spread is typically 1-3C for all the top TIMs and that's what 4ryan6's data and also Hardware Secrets and numerous other reviewers have measured.
September 16, 2011 12:05:58 AM

thanks for all the info guys, has anyone here tried indigo extreme?
a b K Overclocking
September 16, 2011 1:02:35 AM

It looks like Indigo Extreme is ~ 2C cooler than Arctic Silver 5, so nothing significant in reality, mostly just hype. It's a unique TIM that has to be bought to fit a specific CPU model. It hardly seems worth the hassle and price when 2C means nothing in the grand scheme of cooling.

http://www.indigo-xtreme.com/
a c 206 K Overclocking
September 16, 2011 12:29:38 PM

@digitalraven

If you do carry through with your lapping plans, keep in mind you want a thermal compound towards the thinner side, because it takes so little under those circumstances, you don't want a layer of thermal compound between the two surfaces under those circumstances.

You can actually take a sandwich bag and use your finger to film coat both contact surfaces individually, no buildup just a film coat, and if the two surfaces are perfectly flat, that's all you need.

That will give you the maximum thermal conductivity, you are only filling the microscopic imperfections between the two surfaces that air would occupy if it were metal to metal so you're replacing the poor conducting air gaps with the excellent thermal conducting properties of thermal compound.

Remember air is an insulator and a very good insulator, that's why the new thermal pane windows have a space between the two glass layers.

Todays thermal pane windows are not filled with air but argon gas, because air has it's bad side with condensation, and did not prove reliable enough to use long term in thermal pane windows, but it's the insulating principle I'm trying to get across here.

That's the entire purpose of thermal compound to replace the air, so very little air means very little thermal compound needed.



October 3, 2011 12:04:53 AM

Sorry for the late reply,

Thanks for the explanation 4ryan6. My current plan right now is lapping my i7 920 and then putting in the Corsair H100 all in one liquid cooler. Ive seen other people just using the pre applied TC on the plate and i might go that route. I might also buy the Cool Laboratory Liquid Ultra, clean off the pre applied stuff from the cpu block. Using the brush method wont create air bubbles will it?

I also plan on buying COUGAR CF-V12HP fans for the radiator in a push pull config.

http://www.newegg.com/Product/Product.aspx?Item=N82E168...

!