Thermal Paste Removal/Installation Guide

[amuffin]

Edited, sort of..

 

[Galego]

Can I use any kind of thermal paste or there's an specific type for each processor?

Step Three: Application of New paste.

Once you have removed the old paste from both your processor and heatsink, you are ready to apply your new thermal paste!

Since we are using Arctic Silver 5, the application will be slightly different. Unlike other pastes, Arctic Silver 5 requires us to apply the paste in a vertical line method on specific processors, in this case we are using the Intel Core i7-2600k.

For a list of processors Arctic Silver specifies, click these links:
Intel: http://www.arcticsilver.com/intel_application_method.html#
AMD: http://www.arcticsilver.com/amd_application_method.html#
http://img171.imageshack.us/img171/4406/59010998.jpg


Arctic Silver 5 Performance Compared to Other Compounds:
http://i130.photobucket.com/albums/p250/4ryan6/TRAC2.png
Line Method:
http://img40.imageshack.us/img40/5923/29567930.jpg

Right after applying paste you should install your new cooler! Personally, I use the weight of my cooler to spread the paste. However, you can spread it around to a thin layer with things such as an old credit card.

Step Four: Installation of New/Old Cooler.
Now we have installed the Hyper 212 EVO, we are good to go!
http://img708.imageshack.us/img708/5995/27468507.jpg

Thermal paste and cooling fan temperature results belong to 4ryan6.

 

[amuffin]

Any, I just listed the processors Arctic Silver specifies for their AS5 paste.

Every thermal paste is different in it's own way, make sure to read the instructions before using it!

 

[e56imfg]

Great guide but paper towels? No. Coffee filters are the way to go.

Can I use any kind of thermal paste or there's an specific type for each processor?

Anything but liquid metal TIM would work. Liquid metal will just destroy your aluminum HSF and the top of your CPU (along with its warranty)

 

[randomizer]

I use kitchen wipes myself. They don't leave fibres and are big enough to have plenty of surface area for several iterations of the cleaning process. I usually go over each surface with a new application of alcohol three times (possibly more for IC Diamond because it's so thick). It's surprising how much residual paste can cover each surface even though it looks clean.

 

[amuffin]

I wonder how well alcohol wipes would perform.

 

[amuffin]

Q/A Section anyone?

 

[-Fran-]

For how much time we have to replace it ?

What do you mean by that?

Once you apply the paste, you have time until dirt gets on the paste or something like that. And once applied, it should last a good 10 years before even worrying, if not more. It really depends on the compound used, but most should last a long time.

Cheers!

 

[amuffin]

For how much time we have to replace it ?

It really shouldn't have to be replaced, unless you are getting intense temperatures. If not, paste can last a VERY long time before drying up. As well, it depends on the quality of the paste.

 

[amuffin]

Guys, help me come up with some common questions and some answers for them.

 

[ddan49]

Common questions (from an actual dummy ):

1) Why replace TIM?
2) How often should you replace TIM?
3) What's some good TIM?
4) What are some other types of TIM? (There's paste and like... a liquid metal)
5) Is there any difference on the application methods? (Lines vs one big glob)
6) What are some common mistakes?

That should get you started!

 

[Ladamyre]

Maybe some pastes dry out in 6 -12 months but I've pulled coolers off of my CPU's after years of usage and found the stuff still sticky, still working. Seems to me if the paste is such poor quality that it dries out that quickly, I'd quit using it.

Then again, there's a company that's selling some sort of plastic/metal type material that needs to be 'burned in', that is to say melted into the interface after it's installed and is supposed to be better than any paste and permanent to boot. I plan to use it on my next build.

I won't name it, don't want to be accused of spamming, but I will say it has a colorful name. Don't want anybody to get the 'blues' over me promoting a product here.

 

[ddan49]

Ah... Ladamyre? This entire forum is built on recommending products

For example, recommending an i5-2500k is perfectly fine... and many people recommend AC5 for TIM. I think you'd only risk getting banned if you did something like nonstop promote an irrelevant product.

 

[AntiZig]

some questions to the section of "common mistakes"
1. what happens if you put too much paste and it comes out the sides? what are the risks?
2. what happens if you put too little and it doesn't cover the whole surface of cpu case /HS

 

[Ladamyre]

Ah... Ladamyre? This entire forum is built on recommending products

For example, recommending an i5-2500k is perfectly fine... and many people recommend AC5 for TIM. I think you'd only risk getting banned if you did something like nonstop promote an irrelevant product.

OK. It's called Indigo Extreme. (Get it? 'Colorful'? 'Blues'? :lol: )

Don't hold them to any claim on my part, but the more I think about the stuff and how it's applied the more sense it makes. Then again, I'm one of those overclocking cowboys that'll spend $20 just to get two degrees cooler and thus get 20 more MHZ out of the CPU. When you think about it, that's less than a penny per KHZ! Whoopee!

 

[-Fran-]

Relevant link for today:

http://gizmodo.com/5920224/brilliant-spinning-heatsink-cools-cpus-30-times-more-efficiently

Cheers! xD

 

[ddan49]

That is pretty awesome. I'm surprised no one did it before (I've thought of it), and even more surprised it actually has benefits. It takes more power to move that thing than a fan.

 

[AntiZig]

nice find, the only problem i find with it, it has to be mounted without thermal paste at 1/1000th of an inch above surface. Practically that's impossible to achieve in non-lab environment.

Also, it will collect dust and that monstrosity spinning at 2k rpm will make noise, looking at the form it will chop the air so much

 

[amuffin]

Relevant link for today:

http://gizmodo.com/5920224/brilliant-spinning-heatsink-cools-cpus-30-times-more-efficiently

Cheers! xD

 

[AntiZig]

I'm not following..?

 

[Ladamyre]

I'll believe it when Tom does a review.

Oh, I get the idea about the 'air bearing' and no, it won't take a lab environment to make that reliable, your cars engine does that using oil.

What I don't believe is the claim "X times more efficient than some air coolers".

Hell my 6 year old Zalman is 30 times more efficient than some air coolers too.

That claim is what we used to call 'dooky', don't mean sh... uh... cra... can we say crap?

 

[bigcyco1]

Great job amuffin i do the size of a pea it works that way fine as well though i know some disagree which is fine to each their own my way works fine for me lol.

 

[amuffin]

Great job amuffin i do the size of a pea it works that way fine as well though i know some disagree which is fine to each their own my way works fine for me lol.

That would depend on the paste and processor [:lutfij:4]

 

[bigcyco1]

That would depend on the paste and processor [:lutfij:4]

Very true

 

[Jonathanese]

http://www.frostytech.com/articleimages/Dy1206bh638_finger2.jpg

I wonder if that red stuff can be used as TIM?




Anyway, even with systems with heat spreaders, wouldn't there still be an advantage to spreading the thermal paste over the simple line method? It seems like no matter the TIM, the line method would still place too much in the middle, and not take advantage of all that extra real-estate for transfer. I figure, if you go ahead and spread it out, you only have as much TIM as you need in the center, but also allow better heat "bandwidth" so-to-speak. I say this, because I'm about to get a direct touch heatpipe cooler, and if I only use the line method, I'm making decent contact with only one of the three pipes. But in spreading it, I take advantage of every square-inch of the cooler and IHS.