Added: 3 years ago
From: yaochoon
Views: 451,918
Sort by time | Sort by thread (beta)

Link to this comment:

Share to:

All Comments (825)

Sign In or Sign Up now to post a comment!
  • wow had no idea... looks like the line method is best really

  • Well done!

  • best on youtube, well done!

  • Witch thermal paste would you recommend?

  • @joines92 go to maximumpc and search thermal paste. They test quite a few compounds but I think the Tuniq-TX4 was their top performer.

  • Cross method is the most effective and prevents air pockets while ensuring even coverage.

  • CROSS FtW!!!

  • I am a pure believer of the line method. The best way to do it is hot-dog style (adding the amount that spans the longest distance of the dimension of the CPU, but not all the way of course).

    Great vid. Thanks!!!

  • This is by far the best video I have watched regarding how to apply thermal paste. Fantastic job! That's science at its finest. Thank you!

  • Cross method is the best

  • Can you make a video showing the proper way to spread thermal compound on a HDD.

  • YOU, good sir, are brilliant

  • what about 5-10 peas spread out, would that be alright?

  • @vicx05

    Still to much just a pea size in the center is fine because the actual chip is in the center.

  • @ZomgACake is the chip always in the center for all CPUs and GPUs? why is the heatsink over the chip so big if the chip is so small? most of the heat would be transferred out by the larger heatsink/fan on top of it, so why not have a smaller heatsink on the chip?

  • @vicx05

    The small chip heats up very fast and the metal case absorbs it but it still needs support from a large or small heat sink(mostly large). Thermal paste helps the transfer of heat between the cpu and heat sink. (Good question idk if all chips are in the center but most ones are from major brands).

  • @vicx05 because the thermal paste would fall from the CPU into the mother board, also there's the need for space for the pins of the CPU

  • @vicx05

    No! You want as little paste as possible. Otherwise, the paste itself will act as an insulator...

    The point of the paste is to fill the air pockets in the contact caused by manufacturing imperfections. This will create better heat dissipation because thermal compound is a better thermal conductor than air. However, these air pockets are microscopic, so you hardly need any paste to fill them...

  • No. You're wrong. it need to speard entire CPU.

  • The amount doesn't matter, except if scoring neatness. The worst thing to do is create air-bubbles and then press that air-bubble filled paste in-between two plates. The air-bubbles will stay there and the micro-volumes with resonate with heat, until the pressure causes the hardening paste next to the bubbles to crack, then the paste is useless. Also, those bubbles in traveling laptops are affected by altitude pressure changes and will crack the the old paste up. Smooth, and slow... !

  • If only half the CPU die is covered by the thermal compound, will it significantly reduce heat transfer, or does the heat and pressure cause it to spread to completely cover it after an hour or so? I ask because my thermal compound (Noctua NT-H1) seems far more viscous than my previous compound (Akasa 450), and I was worried it would not have spread properly.

  • @Flyntendo After cleaning surfaces of crud and dust, use a razor blade to spread a thin SMOOTH layer on BOTH surfaces, then place a small dab of compound on one surface and slowly bring the surfaces together and fasten the fastener or screws slowly. DO NOT CAUSE BUBBLES! Do everything smooth, and slow... to ensure a -solid paste- not one filled with air bubbles. The spring tension determines final thickness, less viscous means a thicker layer will solidify.

  • I don't like the last one too much, imagine if you had a put a tiny bit more and it was Arctic silver

  • witch ones the best

    

  • i like this music what this song?

  • Many thanks for this super experiment. I learned a lot :-)

  • I always warn people about spreading the paste, but they never listen. Now I will point them to this video to see how much air bubbles it causes. DONT FORGET that the actual CHIP is right in the middle of the CPU, not the entire square backing. So all the heat comes from right in the center. The best method and ONLY Method you should ever use is the pea sized right in the center.

  • @staphinfection didn't this video just show Pea size doesn't spread to well and is bad

  • very informative.  thank you for posting this video!

  • Applying thermal paste Dexter style! :D

  • tru to used spread method and then appy the plexiglass from an angle, applying pressure from top to bottom to eliminate as much of the air as possible. It's air vs. more compound...

  • now we just need a person that Knows how much to use to make a video like this LOL

  • hmmmmm not bd

  • I predict super high views for this video - thanks yaochoon ;D

  • is that a ABIT BG7E?

  • Thx for showing

    this is really interesting

  • Over time though the Arctic silver will spread out more once you've settled it in. Paste is normally at it's best performance after a week or two as it's settling in.

  • Great video! Interesting to see how much air bubbles you bake into the paste when using the spread method. Air bubbles can't be good for heat conductivity. I reckon the air bubbles will even create an insulating effect which is not what you want... Personally I think I'll stick with the dot or line method.

  • @KarlHansson it makes me mad when people put on a stupid glove and spread the paste. thats just stupid.

  • Fake!! I tested it with semen. Nothing like your video shown.

  • cut your fingernails, jeez

  • hey, i have a problem, my pc won't start after i applying some thermal paste :( i think my cpu is dead

  • @afieqM haha you probably shorted your CPU. thermal compound can be very conductive, especially if it overflows into the pins.

  • MX-2 > AS5

  • I just reapplied my thermal paste from the finger on the bag spreading method to the line method.. and boy did it make a difference.. at 4.5ghz i was getting 84 degrees easily with prime95 at 1 or 2 minutes.. now i'm getting 65 at 5 minutes now... lol 20 degrees.. WOW!

  • I don't know how i get on here but? its kinda Good i got to admit!! :0

  • I smear it with my finger, although I use alot :P

  • @onionman113 Using too much thermal paste will actually increase the heat. I found this out the hard way when my computer started randomly getting BSOD's. Took forever to figure it out.

  • @ErikLiberty I beg to differ, you must be using a really shitty compound, I use AS5 and a blob a little bigger than a pea does awesome.

  • @ErikLiberty lol i guess that's what happened to me...was getting 85 degrees when i used the finger on the bag method...i used the line method and now i'm getting 65... lol i'm so happy right now.. haha

  • @onionman113 don't use ur fingers.. really bad.. xD

  • @don2xify naaw it's not, i usem to spread it all the time and i never have issues with heat, an i should know i got intel hottest chip ever made (pentium d)

  • @onionman113 have u tried the line method though? maybe i did the spreading method wrong.. but it's way easier to use the line method..

  • @don2xify Yes, actually on the dell I'm working on next to me. Doesn't seem to make a difference I use the same paste for every computer I work on and thats AS5. I think the only factor is the type of paste and the coverage.

  • does anyone know if the spreading vs dot method makes a big enough difference in temps to even matter?

  • You would be better off reading the PDF that is provided by arctic silver. The manufacturer DOES NOT RECOMMEND this method! Not at all!

  • song name plleeease?

  • And now as my coupe de grace... for those of you who do not understand high frequency EMI decoupling... read this:

    intel.com/design/pentiumii/app­lnots/24333402.PDF

  • Just to put you folks at rest, Arctic Silver says, "While it is not electrically conductive, the compound is very slightly capacitive and could potentially cause problems if it bridges two close-proximity electrical paths.)" My meter is a Fluke 170 with capacitance measuring accuracy ± (1.2%+2), Max. Resolution 1 nF, Maximum reading of 10,000 µF and IT says, NO PROBLEM. So the Arctic site is just being extra cautious with nanofarads,

  • ACTUALLY, there ISN'T such a thing as 'too much paste', at least not with silicone oil based, metal filled, non-conductive pastes such as ARCTIC 5. You could smear that stuff ALL OVER your logic board and it won't hurt a thing. I can PROVE IT. Use your VOM set to 20 megaohms and try to read the resistance of that paste... you'll find it is essentially 'non-conducting'... so, where is YOUR proof?

  • @doceigen A 3 1 uF polypropylene film capacitor would have an extremely high resistance if measured with a multimeter too, but you wouldn't want to put one across your CPU's signal lines.

  • @Vegemeister1 Yes, but since 'arctic' is NOT a capacitor, NO PROBLEM. WHAT'S NEXT!? Who's going to jump in here next, and say that the mental power of anxiety caused by postmodern-neurosis, will cause massive quantum tunneling effects, between the junctions of the microprocessor itself, due to an unrecognized 'worrying about energy', produced by misinformed amateurs, possibly getting paste on their logic boards?? *sheesh*

  • @doceigen The back of my NT-H1 thermal paste says "Caution: Applying too much paste can increase CPU temperatures!" I've done it myself in the past and gotten BSOD's from it.

  • @ErikLiberty Nope, no connection between the 'amount' of the paste used, to how well it works coupling the chip top to the heat sink. NT-H1 is very 'thick' and so the spring pressing the heat sink to the chip top needs to be very strong, you want the layer with ANY coupling compound to be as thin as possible, but that has nothing to do with 'amount' since extra paste, squeezes out the sides anywise. Remember, pressure versus viscosity will determine 'thickness' of the coupling layer.

  • i like spreading, or using two lines.

    this video is okay, but he only held the mock heatsink on the chip from the sides with his fingers, forgetting that the actual heatsink when tightened with screws, will put much more pressure on the sides AND center of the chip. and twisting does help, you just can't tell much, again because you are not applying the same force as a screwed on heatsink.

  • Best spreading method... 30,000 units tested. 'Doctor's Knife' method first, on both surfaces... or as some people call it, carding. Spread a thin sheet using a razor or card. Thickness should be about 50 microns. Then before you bring them into contact, place a dipple (dot) of compound at the center of each spread about twice the material you see him use here... now, bring the two together and don't wiggle it, just clamp. Run device 10 min., then turn off 20 min. Now, run and enjoy.

  • @doceigen sounds like a recipe for using too much paste applying on both surfaces. THE BEST way to test is to try different methods and pastes yourself by applying paste, install heatsink, test temps, remove and check for coverage, clean, and repeat by using less or more paste and/or different paste. today's pastes don't need to cure, so results should be near immediate. the heatsink itself and ambient temps are also factors.

  • @jurassic512 Your statement make 'little' sense. First, there is no such thing as 'too much paste', second, if there is... then why would anyone enter a method test program using up paste and more paste and even more paste? In reality, ...coverage is either adequate or not... the method I describe assures coverage. There really isn't any other issue, aside from aesthetics if some of the paste gushes out the side, so WHO CARES?

  • @doceigen Actually, there IS such a thing as "too much paste". The person would have to be an idiot to apply such a massive quantity, but it has happened, and if it does it will destroy the CPU in question.

  • @jurassic512 Well there is one more initialization issue, cleanliness. In order to assure an even 'coverage' the two surfaces need to be alcohol clean to the metal, and as dust, grease, fingerprint FREE as possible to start with, before applying any paste at all.

  • @doceigen thats common sense. it goes without saying. but for the record, if you're not a clutz, when you unpackage the heatsink and fan it should already be clean. you only clean it if you're dumb ass touched it, or you need to remove paste and reapply.

  • @jurassic512 I've watched many of the 'thermal compound paste' videos on You Tube, and they are all demonstrated on surfaces that are not virgins, and in fact, from the comments, I'd say applying thermal compound is kind of a fetish, a kinky electro-sexy thing to do, which is given ritual, ceremony, and myth... so yeah, my saying use an alcohol bath to wash away the nasties first, probably sent thrills and chills along some geek's medulla oblongata.

  • Thanks for such a helpful video.

  • @ezeht Facts about Silver 5 folks never seem to talk about. Silver 5 transfers electrical and can ruin your day if sloppy and in tight spots like gpu's. S5 has a 200 hour cure time. S5 has a 2 year life and gets brittle- bump or twist your cooler a lil can break its seal and it cools less over time. AC MX2 has no cure time- MX2 does not transfer electrical- MX2 has a 8 year life. AC MX2 cools some 6f better than S5. MX2 is just better period. MX3 cools better than mx2- so..

  • Comment removed

  • Is it best to spread AC5 with a card, then?

  • some things you are not taking into account is that the constant pressure the cooler will apply over a LONG period of time also that the heat from the processor will drastically INCREASE its viscosity of the compound.

  • Song name:

    Leviahawk - A Day In The Life Of...

  • Sorry, but the music is horrible.

  • so spreading is bad ????

  • thnks

  • Tx! that was helpful :)

  • My laptop CPU is tiny. What would work best for my laptop?

  • Next time put hummus !! (:

  • Pros use a 5 dot X shape.

  • nice work

    

  • @3DMarkSeries He created this song on Musicshake!

  • Wow, you pressed down so hard.

  • It'd be interesting to see what would happen to the paste in each circumstance while the processor is running. If only we had transparent aluminum to conduct the test while still having some heat dissipation!

    Using that glass or plastic I suppose the processor would just throttle down and the OS/software might crash while doing the test.

  • You need to apply more thermal paste than you used. 2-3x amount than you used, will be enough. Squeeze it hard enough on the middle, not by sides. Squeeze down the heatsink after the montage. Even pressure, so it will spread good. It is also important the paste layer is thin enough, so the metals make contact but also cover the valleys so the paste can transfer heat. It is why thermal pastes work. Also, the overflowing paste on sides. is no danger if it is little. Pea is the best method.

  • Hi, I think the pea method must be the best since the guys building the world's most powerful supercomputer use this method:

  • @gyengez Yes, you have right. I also know that too.

  • I wonder why people insist on spreading... Common sense really is hard to kill...

  • So Cross method da best?

  • would applying the pea sized method for the cooler master hyper tx3 be ok ? thanks

  • one thing I've never understood is why they bother to make a cpu without a heatsink as part of the processor .... it's a ridiculous method to transfer heat with paste when the heatsink could be literally connected to the heat source. They work so hard to make the best processor, but don't bother to cool it correctly. I just don't get it. Wait, it's because they don't want it to last more than a couple of years, so that the consumer market has to keep buying the new ones.

  • @gorillasalads Read about damaged Athlon's XP. Core is too weak to be not covered!

  • @gorillasalads you can buy good heat sinks heat sinks that aren't generic, besides i like to customize, and when i can buy a good looking heat sink, im happy.

  • @gorillasalads wouldn't it be so u can have an option of changing the heatsink. Cause if we didn't have an option we/d just be a mac

  • @gorillasalads I'm glad they don't make processors with non-removable heat sinks! Have you seen how wimpy and poorly-designed stock heat sinks are??? I like being able to remove and trash the stock heat sink so that I can replace it with a good one and overclock my processor. If the stock heat sinks were non-removable, I wouldn't be able to have my water-cooling kit, nor a decent overclock. Many PC end-users want customization; otherwise buy a laptop or a mac...

  • What would be nice to see is AS5 applied in very thin layer + pea drop in center.

  • helpful video, thanks for the info

  • look at how much you put on of the silicon line, and then the arctic silver line.. its like twice as much.. no shit it spreads more

  • @AbominableSnowFish yes, but the arctic silver is also thicker than the silicone as you can see and still would spread less with the same amount.

  • @AbominableSnowFish he probably did that cuz arctic silver is really expensive

  • @AbominableSnowFish yes 'thicker' compound like AS5 spreads less than say a silicone base compound.

  • 49 people don't want to apply thermal paste to their cpus ...lolllll :D:P

  • @sai331995 One person is a douchebag who goes to videos and looks at who doesn't like it then with their enormous brain power copies and pastes that number in their mind and applies it to another sentence that is totally gay.

  • woops.... I forgot to say lolllllll rofl ROFL

  • I also happened to run across a review on heatsink pressure. This also has an effect. It is at Overclockers website and the article title is Heatsink Mounting Pressure vs Performance. The conclusion I came to is that unless you lap (sand) the CPU and Heatsink, then you might as well as use a low viscosity thermal grease like the white stuff. Unless their is high pressure between the CPU and heatsink, the high viscosity stuff can make things worse.

  • This was pretty good, but unless you put it in an over for awhile you can't be sure. Any air could create some large gaps so using the center dot method, and using enough to cover the whole chip, may still be the best method. At Innovation Cooling's website, the "application notes" link has pretty good info on the results of heating and not putting enough grease to cover the whole chip.

  • lol, Just put a dot and spread it yourself. If you're gonna take the time (3 minutes?) to replace or add a cooling system, do it RIGHT.

  • @neblixmusic unfortunatly spreading it yourself isnt the right way of doing it,

    best test i ever saw was a BB size blob in the middle, then seat the cooler over it, the paste doesnt have to cover the whole cpu, and the thinner the spread the better

  • Is it possible that I could've put too little AS mx-3 on AMD Phenom II X4 955 Black Editon (C3) cpu (like a grain of rice?) and it did not spread well?

    My temps are like 40 degrees idle with Zalman CNPS7500 CU LED CPU Cooler :s

  • Is it best to put a dot right on top of each core?

  • which method do you guys recommend?

  • very helpful

  • looks like the cross idea looked like it had the best results

  • Go full screen to click on the annotation.

  • @xXZZomgXx dose not work

  • is thermal cmpound all u need. the white stuff, I put this on and fan is still taking off like a fighter jet

  • Wow ... This is really a helpful vid !

    So ... Its better just to put a cross of paste and then put cooler on ... No need to spread .. Right ?

  • good video

    

  • Nice video thanks for the info 

  • very informative. I think I apply to little tp on my cpu :(

  • Nice, but polymethylmethacrylate or glass aren't metal.

  • Comment removed

  • 48 ppl have more Thermal Compound

  • Where did you get The glass ? With The processor ??

  • Waw if i push on The processor like this, will it break it ? Oo ça pète pas le proco d appuyer comme ça ?

  • 0:07 - 0:27 good! xD

  • Right after the 2:00min mark he's bending the plexiglass letting a lot of air in. The only thing that part of the video proves is that spreading gives u the thinnest layer of paste.

  • @endemion06463 true, but it still shows you that it can happen. A thin layer is good, but that can be achieved by just using less anyway. In the end whatever gets you the best temps is the best, this just helps people make less mistakes.

  • @drx975 It can happen? So metal bends like plexiglas? This video just gives a very distorted view of what happens when u spead it out by hand. Also putting less on spreads less as the video shows (+ really hard to tell how much u really need, especially taking into account those direct touch heat-pipe coolers + try attaching a ~1000g cooler in as smooth a way as the plexiglas is put on in this vid (think bolting it down at the back of the mobo)).

  • @endemion06463 Are you trying to argue its impossible? Problems occur when you have a slightly convex/concave cpu (most are to a degree) or an uneven HSF... maybe you need more TIM in the middle, but you already spread it ultra-thin yourself so you could end up with air gaps over the cores. You can get exactly the same effect as spreading does by just using less paste, or using lines/dots at various locations, so there is no use in spreading. Then there's also no chance for bubbles.

  • @drx975 Nothing is impossible. But concave cpu heat spreaders are very rare simply because of how they are made/attached. If you do have such a cpu, u should return it because it is then impossible to correctly attach a cpu cooler (a thick layer of paste (concave) could work heat isolating as well). So just for the off chance u get a crooked cpu heat-spreader u should always use a method that clearly has some disadvantages?!

  • I use vertical line for my i5 2500k, that works best.

  • pea or x method works best. 50lbs of pressure. synthetic diamond is better than as5 or mx2/4. Twisting and changing pressure points seems to me to create air bubbles. great demonstrative video thanks

  • hi can someone tell me i just took apart my laptop and it back together now and running fine but i replaced the old crusty thermal stuff in between the heat sink and the cpu with arctic silver 5 and there was a very light green pad of thermal Greece i think its called on another platform on the heat sink to the right of the cpu platform i replaced that with arctic silver 5 too and i was wondering was i supposed to do that???? please someone help

  • @unknownplayer03 probably nothing to worry about

  • for thae correction thermal paste there crematic and silver

  • Very informative video, thanks.

  • the biggest fail of the year would be pulling processor out of socket with that glass :D

  • Yeah, but compounds with silver in it work better than some Silicone sh*t :)

  • Diagonal cross method seems to have the best coverage.

  • Does anybody know the best method for a PS3 using Aritic Silver 5?? Please help

  • thx :)

    

  • Wow, can't click the annotation because the Youtube player covers it. T_T

  • @Haroids if you click the expand button (right next to the full screen) it gives you like on millimeter and you can click the annotation. just tried it.

  • @xxbrkdwnxx Youtube should get their act together.

    -smh-

  • @Haroids couldn't agree more. i'm personally vexed with the LENGTH of some of the ads on videos. i know they allow youtube to remain free and what not but a 30 second ad/trailer before a 25 second video clip of a ninja cat or auto tuned baby is ridiculous, maybe even more so than the video itself.

  • @xxbrkdwnxx Rofl, I agree as well. It's especially annoying if your internet is slow and you have to buffer the ad, making it even longer than the set time.

  • @Haroids

    i tried doing it quickly for like half the video loool

  • @saqibd123new Rofl.

  • @Haroids heres a tip: ctrl + ;)

  • @Haroids ctrl + until it u can click the annotation

  • @Haroids go fullscreen to click annotation..

  • @gmax876 IT DOESNT WORK! :( GAAAH

  • so, which the best method?

  • Did he died?

  • I have a few questions because I am about to embark on building my first pc. First, don't you want to refrain from apply pressure onto the cpu in fear of harming the pins? Also, when he showed the spread technique, why didn't he just leave it nice and spread out. It seems like when he went to apply pressure with the glass piece, it just created bubbles and took away from the amount of paste already on the cpu. Thanks for making the video!

  • Okay, i saw the link in the description. I still don't understand how the spreading method is inferior. How could it cover less surface area if it expands to cover the entire cpu, even if the heatsink base doesn't reach that far. Also, if the spread were convex, that means that the area with greatest density of paste is in the center, which would have the most contact with the heatsink base, which is a good thing I would assume. I think I am misunderstanding something.

  • @willsultan because when you push the bead down with the bottom of the heat sink, the paste spreads out with the sink and fits the sink's form, thus all of the microgaps are filled. if you spread the paste before you put the sink on, and the base of the sink is not flat, some of the microgaps away from the center will not be filled because the paste was pre-spread instead of being spread by the sink.

  • @Donut117 Ah okay, thats what I expected. Thanks for clarifying so precisely. What if you were to spread the paste first as a cautionary measure then add a miniscule amount to the center. That way the heatsink is still spreading the paste, yet you still have the entire expanse covered through the spread method. Or is that just a superfluous amount of thermal paste?