Best Way to Apply Thermal Paste

This is probably the most asked question, and I've researched the internet but I cannot find a a solid answer for it. Everyone has their own opinion and opinions are not facts or a valid benchmark.

Does anyone know of a scientific test that was used which proves the best way to apply thermal paste? Did Tom's ever do a review?
43 answers Last reply
More about best apply thermal paste
  1. I just used my finger, carefully spread out a thin layer :)
  2. I have just found this video which it's very well done. Here you can see the correct way to apply the thermal paste:

    http://www.youtube.com/watch?v=o7rPqCvCt0g

    He is a young guy but He knows well what he is doing.
  3. blackhawk1928 said:
    I just used my finger, carefully spread out a thin layer :)


    this is what i do...make sure it is a thin layer. too much thermal paste is bad
  4. All that I write are my opinions. Nothing more. You can take them or not.
    For me that is the correct way. If you have another one which is better for you, good for you.
  5. Quote:
    This is probably the most asked question, and I've researched the internet but I cannot find a a solid answer for it. Everyone has their own opinion and opinions are not facts or a valid benchmark.

    Does anyone know of a scientific test that was used which proves the best way to apply thermal paste? Did Tom's ever do a review?



    Your JOKE-IN? Right? Use your finger and put it on there. If your not running a radiator or anything like thermal gel in a fish tank you shouldn't worry about it. All you need is a finger and some Artic Silver compond. All your answers are on newegg.com

    FYI- All you have to do is make sure you don't put to much or to little and that it dosn't go over the sides of the CPU.

    Thanks.
  6. Not your finger, unless it's sealed in plastic wrap or a baggie. You can get skin cells on the paste. But you do have to make sure there are no spaces when the 2 pieces meet A-S id gummy, doesn't flow like the white stuff. I had to do my first Arctic Silver 3 times till it was just right.
  7. An important note for begginers: DON´T apply thermal paste over the mosfets. Most of the thermal pastes (Arctic silver 5, for example) conducts electricity and you will kill your mobo.
  8. graywolf said:
    Not your finger, unless it's sealed in plastic wrap or a baggie. You can get skin cells on the paste. But you do have to make sure there are no spaces when the 2 pieces meet A-S id gummy, doesn't flow like the white stuff. I had to do my first Arctic Silver 3 times till it was just right.


    It dosn't matter trust me I run a real gaming rig:

    Specs:

    GTX 260 1GIG of DDR3

    AMD ATHOLON 2 x 2 Dual Core @ 3.0Ghz

    3Gig's of Ram Dual Channel DDR2 800Mhz

    160Gig Veloc-raptor 10,000RPM

    and I'm playing on the Crysis editor Resolution 1680 by 1050, all VeryHigh settings and I'm writing a paper on Office and on here telling you that using your finger to apply the thermal paste WONT HURT IT I PROMISE YOU!!!! You know what I DONT EVEN USE A STATIC WRIST STRAP!
  9. ct1615 said:
    frankly i could care less how anyone applies thermal paste but wtf is a "real gaming rig"? Are there fake gaming rigs? You have a certificate with a stamp stating "REAL GAMING RIG"??


    If I did would you be surprised! lol...

    Well I can tell you I think its funny when were having a thread about the most perfect way to apply thermal paste I went to college for 2 years and thermal paste wasn't ever and I mean ever talked about in this extreme. Slap it on there and your good to go!
  10. ct1615 said:
    besides intel/amd,
    ATI/Nvidia
    I would say thermal paste would be the next most popular argument. The fact is as long as it works it makes little difference on the application method.



    Thank you. End of our Discussion.

    If it aint broke don't fix it.
  11. graywolf said:
    Not your finger, unless it's sealed in plastic wrap or a baggie. You can get skin cells on the paste. But you do have to make sure there are no spaces when the 2 pieces meet A-S id gummy, doesn't flow like the white stuff. I had to do my first Arctic Silver 3 times till it was just right.


    Buddy, believe me...i applied the paste with my finger and my i7 idles at 26 Celcius....and doesn't get past 40 (maybe very rarely) during various loads.
  12. Anyone suggesting a one size fits all method doesn't understand that each thermal paste is different. Some are ok with multiple application methods, others only work properly with one. I can think of one paste where spreading with your finger will do nothing but polish the stamp off the IHS.

    As a general rule I would suggest using pressure to spread the paste, but you should still research the product you bought.
  13. iholbrookv2 said:
    It dosn't matter trust me I run a real gaming rig:

    Specs:

    GTX 260 1GIG of DDR3

    AMD ATHOLON 2 x 2 Dual Core @ 3.0Ghz

    3Gig's of Ram Dual Channel DDR2 800Mhz

    160Gig Veloc-raptor 10,000RPM

    and I'm playing on the Crysis editor Resolution 1680 by 1050, all VeryHigh settings and I'm writing a paper on Office and on here telling you that using your finger to apply the thermal paste WONT HURT IT I PROMISE YOU!!!! You know what I DONT EVEN USE A STATIC WRIST STRAP!


    I would LOVE to know how you running a "REEEL GAMING RIG" makes your opinion any more valid, let alone how well it will run crysis.


    Personally I'd use a plastic bag over your finger if you're spreading it that way, some of the thermal paste is nasty ***.
  14. iholbrookv2 said:
    It dosn't matter trust me I run a real gaming rig:

    Specs:

    GTX 260 1GIG of DDR3

    AMD ATHOLON 2 x 2 Dual Core @ 3.0Ghz

    3Gig's of Ram Dual Channel DDR2 800Mhz

    160Gig Veloc-raptor 10,000RPM

    and I'm playing on the Crysis editor Resolution 1680 by 1050, all VeryHigh settings and I'm writing a paper on Office and on here telling you that using your finger to apply the thermal paste WONT HURT IT I PROMISE YOU!!!! You know what I DONT EVEN USE A STATIC WRIST STRAP!


    Yeah how is that a real gaming rig...? Is there a definition that clearly states what is and isn't.?
  15. I have always used a credit card when applying the paste to ensure I get a nice even thin coat.
  16. ocz freeze, just dab an squash as recommended by ocz
  17. how good of a job you did on applying thermal paste always shows itself in temperatures :)
  18. flaminggerbil said:
    I would LOVE to know how you running a "REEEL GAMING RIG" makes your opinion any more valid, let alone how well it will run crysis.


    Personally I'd use a plastic bag over your finger if you're spreading it that way, some of the thermal paste is nasty ***.


    I'm saying my hardware alone will most likley be running hotter for the fact I'm game with a intensive application an in = Crysis sandbox 2 editor! I applied the thermal paste not even with Artic Silver compound just some stock SH** that came with my AMD II x 2 3.02 GHz Dual Core and it runs cooler then most OEM put together projects that you would be purchasing from a store and I USED MY FINGER TO APPLY! Well I guess if you don't want to get those poor hands a little dirty then use a little baggy over your finger. I'm done talking about this!
  19. I always apply a thin layer to the CPU. I have always had good results in terms of temps that way.
  20. iholbrookv2 said:
    I'm saying my hardware alone will most likley be running hotter for the fact I'm game with a intensive application an in = Crysis sandbox 2 editor! I applied the thermal paste not even with Artic Silver compound just some stock SH** that came with my AMD II x 2 3.02 GHz Dual Core and it runs cooler then most OEM put together projects that you would be purchasing from a store and I USED MY FINGER TO APPLY! Well I guess if you don't want to get those poor hands a little dirty then use a little baggy over your finger. I'm done talking about this!


    There's dirty and then there is toxic, the difference between the two is considerable.

    Also an Athlon 2 running at 3.0 isnt even going to run hot in the first place...
  21. iholbrookv2 said:
    I'm saying my hardware alone will most likley be running hotter for the fact I'm game with a intensive application an in = Crysis sandbox 2 editor! I applied the thermal paste not even with Artic Silver compound just some stock SH** that came with my AMD II x 2 3.02 GHz Dual Core and it runs cooler then most OEM put together projects that you would be purchasing from a store and I USED MY FINGER TO APPLY! Well I guess if you don't want to get those poor hands a little dirty then use a little baggy over your finger. I'm done talking about this!


    Aside from the fact they can be toxic. The bag is to protect the thermal paste, not you.
  22. flaminggerbil said:
    There's dirty and then there is toxic, the difference between the two is considerable.

    Also an Athlon 2 running at 3.0 isnt even going to run hot in the first place...


    First of all anything running under a benchmark or a load can get hot! I've gotten things from RAM to overheat to Harddrives!
  23. qbit said:
    Aside from the fact they can be toxic. The bag is to protect the thermal paste, not you.


    Don't worry the thermal paste will be fine, your not going to hurt it! I know thermal paste is really fragile!

    *lol*
  24. iholbrookv2 said:
    Don't worry the thermal paste will be fine, your not going to hurt it! I know thermal paste is really fragile!

    *lol*


    We are arguing over how to apply thermal paste, really it makes a fraction of a degree difference. So on the subject of a fraction of a degree difference all those oils, dirt, (probably alot of dust, especially if your using any old parts) and skin cells can make a fraction of a degree difference.
  25. Cleaned pre applied off fully with alcohol then use the back of a steak knife apply compound
    to the heat sink it wont be as this as thin as using a a razor blade and you can avoid adding
    any on your CPU
  26. lol this is great

    first off don't use your finger your fingers are oily
    second off AS-5 is not conductive though it is slightly capacitive

    Doesn't really matter how you do it, the goal is to get the thinnest layer you can spread across the entire surface without going over or having any gaps. Personally I use AS-5 and follow the directions that they were kind enough to provide for the product they designed.
    http://www.arcticsilver.com/arctic_silver_instructions.htm
  27. thinnest layer........wrong.......a thin layer people will take this wrong "thinnest layer"
  28. actually it is quite correct. seeing as TIM (thermal interface material) is not nearly as heat conductive as the metal of the heat spreader (metal thing on top of the cpu) or the heatsink (thing with a fan you mount on the cpu) you want the thinnest layer possible without any gaps. TIM is an insulator when compared to the metal components it is used to connect, the purpose of TIM is to fill in the minuscule air gaps (air being an extremely efficient insulator) between the two and form a constant layer of contact. If that layer is thicker than necessary you are reducing the thermal conductivity of the union.
  29. and if the layer is to thin then ultimately its not doing anything
  30. it is in fact a liquid and helps to disperse the heat more evenly //and can dry and has dried yes I have seen this and it causes ....overheating
  31. Pea methods are the best, use ur finger and u`ll have an air bubble stuck between ur heatsink and your cpu! check it on youtube video on how thermal paste work. The video shows some methods and LIVE of the results!

    But then you do not want anything else between a cpu and a heatsink, its only a thermal paste. Then another thing to consider is your own thermal compound, is it premium? or just an average thermal paste. My personal choice is a thermal compound from noctua NH-T1.

    Some thermal paste needs time to actually works between the CPU and heatsink, but with NH-T1 from noctua, u dont need this stuff.. it works straightaway when u put ur paste in between. its safer, its more efficient and u can apply it 15 times (0.5 mm/ each install).
  32. spentshells said:
    and if the layer is to thin then ultimately its not doing anything

    The ideal thickness of TIM is the thickness of a single particle. But that's practically impossible to get.
  33. randomizer said:
    The ideal thickness of TIM is the thickness of a single particle. But that's practically impossible to get.


    nope, I see your a moderator...please be moderate
  34. Yea, ok... :heink:
  35. Spentshells is right. Unless you have a lapped cpu and heatsink then if the layer is too thin there will be gaps. This is why not spreading it is a better method. If it spreads itself then it will be thicker where there is gaps (and its needed) and thiner where there isnt, providing you use the smallest amount necessary you will get very good coverage from pea/lines method.
  36. qbit and spentshells, thermal paste is intended to fill in microscopic imperfections (and also make sure air doesn't get in those gaps)when the heat spreader of the cpu and HSF were machined...the thinnest layer you and i can possibly get...even with a razor blade would already be more then necessary as it needs to fill in MICROSCOPIC gaps...these small gaps don't look like much but do make a difference in conductivity. Anything more then the "thinnest layer possible" is to much as the thermal paste isn't as conductive as the metals of the HSF and CPU so more then needed Thermal paste would actually cause temperatures to increase. So you are wrong.
  37. blackhawk1928 said:
    qbit and spentshells, thermal paste is intended to fill in microscopic imperfections (and also make sure air doesn't get in those gaps)when the heat spreader of the cpu and HSF were machined...the thinnest layer you and i can possibly get...even with a razor blade would already be more then necessary as it needs to fill in MICROSCOPIC gaps...these small gaps don't look like much but do make a difference in conductivity. Anything more then the "thinnest layer possible" is to much as the thermal paste isn't as conductive as the metals of the HSF and CPU so more then needed Thermal paste would actually cause temperatures to increase. So you are wrong.


    Your presuming that the cpu and heatsink are flat. Look at any lapping guide and they will show you just how big bumps can be on the surface of the processor. Your right if its a lapped processor and a lapped heatsink but the stuff out of the factory floor is nowhere near as smooth.
  38. I never said they were flat...if they were flat we wouldn't need thermal paste now would we?...the bumps are tiny, microscopic mostly i said unless the hsf is deformed. Therefore thermal paste needs to extremely thin, the thinnest a normal person can get it with a blade or whatever. This thin layer isn't much but its more then enough to compensate the imperfections.
  39. Your argument is that the bumps are tiny, microscopic even. Im telling you there not, unless you lap your processor. Heatsinks are often much better at having a flat surface but the processors themself you can run your finger along and feel the bumps. For this reason the thinest layer possible isnt always best instead it should be thicker in some places than others (which is why most methods spread by pressure not by finger/creditcard/razor). Although it makes such a small difference either way.
  40. ^ Partially true.

    I'd say that generally heatsinks are worse off then the heat spreader though, a lot of sinks dont even seem to be flat, they always seem to be curved in places.
  41. If your willing to void ur warranty that is! go lapping on your own risk :p even then u cant sell ur cpu
  42. Don't do what i do and use a Q-tip. I'm sure i leave fibers which get incinerated. lol

    Eh, who cares. It works and I've got my CPU OC'd 1ghz. Fact is, if you get too much you run the risk of excess oozing out onto your motherboard when the heatsink is applied. If you get too little, your not maximizing the thermal transfer aided by the paste. I think if you're really trying to OC the max out of your CPU, use the business card method and make sure it's covered.

    Then again... i did get 1ghz out of my CPU by using a Q-tip. So how much does it really matter? lol
Ask a new question

Read More

CPUs Thermal Compound Benchmark Internet