How do you apply your thermal paste

['putertutor]

I've always applied a small dot of the paste at the center and allowed the pressure of the heat sink to spread the paste on the cpu. I haven't ever spread the paste on the cpu first, basically based on my days in chemistry and biology where we had to try and prepare a plate for the microscope. Spreading the sample around always generated bubbles.

I just watched a video (which seemed to confirm my anti-spreading philosophy) showing the results of various methods of applying paste. Basically they used a transparent, rigid plastic in place of the heat sink to show how the paste spread under pressure. One of the techniques was an 'x', rather than a simple dot in the middle of the cpu. This seemed to have a slight edge on the basic dot in getting more coverage. In the end, the difference seemed so slight as to maybe not make a difference, but I think I'll try it out.

So, how do you apply the paste?

 

[BriansTechSvc]

I've heard of a lot of different techniques, i guess it all comes down to preference and technique/skill.. I use the method you mentioned, Small pea sized dot on the center and let the heat sink do the rest, when i built my first computer about 7 years ago, that is what i did (that build is still my machine running 24/7 that i do a lot of my work on)... as far as i can tell it has held up as expected and have not had any issues that would lead me to believe i would have to remove and reapply.. everyone is different tho i guess!

good question, interested to see what other methods everyone here has and their reasoning behind it.

 

[frederick]

In school...some years ago...the army taught us to apply a thin coat to the cpu only. Using a credit card (or in our case, our ID card), leave a thin, almost transparent layer on the cpu. Its worked for me for years. But they also taught us to re-apply thermal paste once every 3 to 5 years based on environment and mean cpu temperature. When in doubt, every 3 years.

My oldest system has been running strong for 8 years, 24/7. I also never liked this computer, as she was my practice system, and has never had the thermal paste replaced. She still operates at 70-80 degrees fahrenheit (air cooled)

I have no objections to the bead method. I feel it does the same, but I fear it might not allow an even coverage area. Same can be said about air pockets and my method as well.

 

[gikstar]

If you use the popular Arctic Silver thermal paste, here are their instructions based on which processor you are using: http://www.arcticsilver.com/methods.html

 

[NYJimbo]

I always spread enough to cover the part of the chip that comes in contact with the heat sink. Either spreading it or "dotting" the contact part with enough dots to ensure there will not be any empty space when the sink it put back in place. Its not a thick spread but its almost all covered. I never trust the "grain of rice" business.

 

[frederick]

If you use the popular Arctic Silver thermal paste, here are their instructions based on which processor you are using: http://www.arcticsilver.com/methods.html

Good to know. Thank you.

 

[16k_zx81]

grain of rice. never had an issue with it.

 

[damitman]

mayonnaise as thermal compound

Just for fun, read this article,

www.hardwaresecrets.com/article/Does-Mayonnaise-Last-as-a-Thermal-Compound/1793

then this to answer, how to apply paste, my alltime favorite is MX-2 or 3 or 4, non conductive, no curing time,ect.

http://www.hardwaresecrets.com/article/What-is-the-Best-Way-to-Apply-Thermal-Grease-Part-2/1392

this web-site has a lot of interesting tid bits, enjoy

 

[angry_geek]

Wait, you're not supposed to use the entire tube?

Kidding. I use a pea-sized dollop in the center, apply heatsink nice and steady, lock it down. I've experimented with just about every technique. No noticeable difference unless you apply too little or too much. Even experimented with some serious copper paste stuff a while back. That stuff was crazy expensive, and, under load, only made about a 3° c difference. Not enough to justify the cost.

 

[NYJimbo]

From the article:

So, does mayonnaise last as a good thermal compound? No, it doesn't.

Well, duh. That's why real techs use Miracle Whip.

 

[angry_geek]

Well, duh. That's why real techs use Miracle Whip.

Who in their right mind would ever use mayo over Miracle Whip? Except Canadians, of course.

 

[brandonkick]

I've spread it thin using a junk promo credit card, I've used the pea sized drop and I've also "dotted" it on there by dabbing the tip of the tube all over the heat sink.

I usually prefer to dab it these days, as I'm sure there will be a pretty uniform coverage. But the pea size hasn't failed me either. On my first build I did the pea size method and pulled the heat sink to check it two weeks later. There was a circular expansion of the drop to all but the last quarter inch or so of the corners of the heat sink.. which I feel was more then good enough.

I also prefer the MX stuff, I've noticed a good 10-15C drop in temps vs stock paste/pads. Which IMO is worth it.

 

[jbartlett323]

I spread with my finger. I know your not supposed to spread, but every time I try a drop or x or any other configuration, I use too much. If I use my finger, I can spread what appears to be a perceptibly thin, even coat every time. Been doing this for years, never had an issue one... All cores, typically even temps (within error gradient anyway). Everything nice and cool for years.
I do figure though, my finger will rot off from a newly discovered type of cancer before I'm done...

 

[SpaceSquad]

I use the grain of rice method for any modern(capped) CPU. I use MX-4 now as Arctic Silver 5 wants a burn in unlike MX-4.

 

[frederick]

I spread with my finger. I know your not supposed to spread, but every time I try a drop or x or any other configuration, I use too much. If I use my finger, I can spread what appears to be a perceptibly thin, even coat every time. Been doing this for years, never had an issue one... All cores, typically even temps (within error gradient anyway). Everything nice and cool for years.
I do figure though, my finger will rot off from a newly discovered type of cancer before I'm done...

Oh well...You get regular exposure to lead if you work on computers enough, so whats the difference? Either you go mad or you die of cancer...

 

[Boston Pro Comp Serv LLC]

Water/Pea drop sized amount in the center, no spread. Reapplay based on when it needs it. When does it need it? When it dries out or gets baked from heat. I replace on my own machines every 3 years.

 

[SilverLeaf]

I spread with my finger. I know your not supposed to spread, but every time I try a drop or x or any other configuration, I use too much. If I use my finger, I can spread what appears to be a perceptibly thin, even coat every time. Been doing this for years, never had an issue one... All cores, typically even temps (within error gradient anyway). Everything nice and cool for years.
I do figure though, my finger will rot off from a newly discovered type of cancer before I'm done...

I do the same thing as well, and never had a problem. Often I'll wrap my finger in thin plastic wrap or use a latex/nitrile glove so as to not contaminate the stuff....or get "finger-rot"

 

[Altster]

I spread with my finger. I know your not supposed to spread, but every time I try a drop or x or any other configuration, I use too much. If I use my finger, I can spread what appears to be a perceptibly thin, even coat every time. Been doing this for years, never had an issue one... All cores, typically even temps (within error gradient anyway). Everything nice and cool for years.

Ditto...ditto...ditto...

Years ago I bought a small container of silicon thermal compound (label is unreadable now - painted on outside of plastic container). And it has been serving my service work splendidly over the years. And I too, (using my finger) spread a very thin coating on the CPU before I latch down the heat sink.

Which reminds me --- a few years ago I took on a new customer who had used a great deal of heat sink compound between the heat sink and CPU so that it would "literally" glue the heat sink to the processor. Of course I fixed that and told the young man that he was lucky he hadn't burned up the CPU. AND this heat sink / compound (glue) / CPU were on a tower computer. I thought the fan ass'y. would have fallen off but it never did.

 

[RedFoxComp]

I came to the same conclusion as most others. Expensive and 'exotic' pastes are not worth a few degrees difference. I bought 200grams of this and haven't looked back.

http://www.newegg.ca/Product/Product.aspx?Item=N82E16835103080

http://www.hwreviewlabs.com/index.php?option=com_content&view=article&id=64&showall=1

 

[MikeLierman]

So many times I've pull off the heatsinks to find little or no thermal paste on the CPU. I use high density silicone paste from Arctic Silver. Not as expensive, but boy that stuff really eliminates any chance of air. I've done the dot method in the paste, also spread it out a little before throwing on the heatsink, but I prefer the X method. Always have had good luck with it. Honestly, I think anything would be better than what manufacturers use!