i did that to my laptop, i applied what seems like the right amount but was not getting the temps I want, then I just put a giant blob and it was cooler
Yes exactly, however when you're using liquid metal or an electrically conductive thermal paste (like the Arctic Silver 5) you have to make sure the compound doesn't leak onto the motherboard when screwing back the heatsink. Most thermal compounds like the Arctic MX-4 or Noctua NT-H1 however are not conductive so this isn't a problem in most cases
The die under the IHS is much smaller than the IHS, the little dot covered most of the area where the die under the IHS is located, which is what matters the most, but it is always better to have most of the IHS covered in paste to optimize heat transfer
Good experiment. Little or no filler, straight to the point with no mucking about. This would be really useful to someone who hasn't built a PC before.
Chances are barely anybody will encounter this but there are super thick thermal pastes, which do yield worse temperatures when applied too thick! Specifically for instance the Phobya Nanogrease. In my tests the single best thermal paste outside of liquid metal and that stuff is so so thick, that you can't even use the blop method. However, when applied correctly awesome results and that explains (to me) why there's so much deviation between tests concerning this paste.
@@dylansmidt6853 Thats only if you use conductive thermal paste I mean he used an unholy amount of it on this short ruclips.net/video/vUOawJ7XkZY/видео.html&ab_channel=mryeester
What about the artic mx 6, i used it on laptop GPU and I couldn't spread it evenly as it was very thick so I put a bit more than moderate amount but not too much
Essentially you just need enough to cover the surface area (or the vast majority of it) in order to have even heat transfer. As long as you achieve that there isn’t much more to do
@@nunya3163 That is what im doing for past 15 years. I just like to be sure that its spread evenly over the cpu, regardless if heatsink covers whole cpu. Excess of the paste will spread on parts here its too thin and its all good.
@@nunya3163 Agree. I've always put a medium bead in the center then used my finger to finely spread around the entire surface. My machines always worked well and stayed cool for the entire time I've owned them. Never having the need to redo the job plus I never got any excess over the sides.
I accidentally put too much thermal paste on my acer nitro 5 and for the first week it was perfect but after it started getting more and more worse ending actually more worse than old thermal paste. It seems to me that at the beginning it is good because the paste is still in a condition like from the "packaging", but after use it starts to harden slightly, which causes hardening on the surface of the processor and the heat sink, and there is still a "packaging" consistency in between, which blocks the heat flow and keeps the heat between the heat sink and the processor (and its MORE LIKELY problem of processors without IHS like in laptops)
@@IamGod534 Don't, I used it once on my PC and forgot to change it out for real thermal paste and after a year, mould formed on my IHS. After cleaning it up the surface was smooth but the colour of the IHS turned green in some spots, other than that minor cosmetic damage, it didn't effect thermals as the surface was relatively flat.
That could explain my findings, I guess... I always used a bit more rather than a bit less, but my temps would get higher than they should be. Removing a bit of paste has always done the trick so far!
Thanks for this video. I always thought I really damage my motherboard and CPU actually just because a bit of thermal Paste in the socket but now I don‘t even think about it anymore so thank you very much!
Some thermal paste is electrically conductive and can potentially cause damage if it gets onto components, so it is something you should definitely still keep in mind when choosing and applying thermal paste.
@@Sevicify Well let’s hope Be Quiet!‘s thermal paste isn’t electrical conductive 😅. But as my computer runs as normal for about half a year now I don‘t think it’s going to be a problem in the future. But thank you!👍
SO many videos have shown that amount of paste and pattern are almost irrelevant as long as the whole of the IHS is covered. I always use the 'butter' method with a thin layer of MX-4. 'Better' pastes or liquid metal make 2 or 3 degrees difference at best, and that's on Prime95 Max Heat which will never be hit in normal running.
Paste is only for surface imperfection, radiators blocks are almost perfectly machined nowadays. Even that small amount of paste would work if he would spread it before instaling radiator. Thats my 5 cents.
@@szaka9395 JayZ2Cents did a test from one spot to 1/2 a tube. Made no noticeable difference (except to the 'overflow' of paste :) ). GPU's especially don't have an IHS so 'spreading' is vital. I found that out when watercooling my 3090!
I think using more and more won’t have an effect because the mounting pressure of the cpu cooler will be the same making the layer between cpu and cooler the same and just pushing more put to the sides
I always spread the thermal paste evenly over the *entire area* of the top of the CPU case. Otherwise I would not know whether the "squishing" caused by the pressure of the heat sink has been adequate to reach the edges of the CPU case. Also I believe that the layer of thermal paste should be *as thin as possible* . It is only there to fill the microscopic roughness between molecules of the metal surfaces. The thermal conductivity of thermal paste would be *much inferior* to that of the metal itself.
The problem with manually spreading it even is you run the risk of trapping air bubbles instead of having a tight spread. It's not horrible but not optimal either.
can a heatsink be too tight? Im doing a passion project, using a big heatsink for from a dell 2400 Dimension onto a Gigabyte motherboard, the hold-down level is tight due to the height of the heatsink.. can this be bad? Would a tight join between sink and cpu affect the heat and paste conductivity?
@@Defender78. I'm inclined to say that the tighter the better - within reason. Don't break or bend anything! Tightness would tend to guarantee that the layer of thermal paste is as thin as possible. It's only there to compensate for the fact that the metal surfaces are less than perfectly flat.
Exactly. I can confirm it from my own experience. Recently I was forced to re-paste my Ryzen 5 7600 because the mobo went wrong (and repaired under warranty). Although I replaced the mobo because it went wrong again as far as I can judge it does not make a difference. Last time I used a lot less paste (same Noctua NT-H1 AM5 Edition as before, did spread it thinly evenly) and the max. temperature during Cinebench R23 dropped by 10 Celsius (!). Thus I concluded the same, both quantity and application matter and the best way is to spread a very thin layer of paste.
It would be interesting to see how this compares to making the top of the IHS, and the cooler plate, totally flat and smooth by sanding them down. I mean, in reality a tiny amount of thermal paste (or none) should be just fine if both surfaces are totally flat. You just need more if your CPU has a convex or a concave shape. When I apply thermal paste I generally like to wiggle the cooler around a lot on the CPU to make sure it gets to the right places.
@@raven4k998 You mean because the paste all squishes out the sides? I did it with my old 4790k a few years ago, but I can't remember what my before and after temps were, but I do remember that it was quite concave in the centre. Might give it a shot with my 5800x sometime, because it does get pretty toasty. P.S. Nice profile pic, she's one of my favourites!
@@JoeBob79569 There has been benchmarks by a bunch of the "world records" setting nerds out there. Basically, polishing those components DOES increase performance, but it's very small margin. It's not worth it unless every other solution (better chokes for less voltage variance, minimizing said voltage for your clock speed, a more efficient cooler) has already been accomplished, especially with the risk of damage on the IHS or the cost of having someone else do the work for you. Hell the cost of Binned vs non-Binned CPU's to ensure you have a more stable chipset (and can run lower voltages) is going to benefit you more.
@@asymsolutions I don't think it's that risky, or that difficult to do it. Just grab a sheet of glass and put a sheet of sandpaper on it. And rub the CPU around on it for a few minutes.
You'd maybe get 2-3 degrees better on this CPU. Just a number I pulled out my behind, but it reflects similar tests that some tech channels have done. Honestly, not really worth it. Not unless you've got a golden chip, and at that point you are probably setting a record as your average joe has maybe a 1% chance at best to stumble onto a chip that good.
I spread a very thin smooth amount onto the cooler, then apply a small(rice) to medium(pea size) to the cpu. This works perfect everytime, good even coverage, little to no overflow, and pretty much dummy proof method.
The medium amount already covers the whole die once pressed down, and the large amount covers the same amount of area and thereby, having the same thermal performance. Conclusion: the performance isn't related to the amount of paste in volume but how much area it covers once pressed down. Simple as that, what's there to investigate?
The only thing I can think of is just that too much is ok just messy? It's an age old argument about how the thermal paste should be applied, how much ect
@@RK-252 the reasoning behind that is because the thermal paste is actually less heat conductive than the IHS and the copper/metal plate of the cooler, therefore actually creating more insulation for the CPU and it being able to dissipate heat slower. However, when you screw in the cooler correctly, the pressure will press out any excess thermal paste, thus removing the insulating factor.
Why would you send them a video showing this? Layer thickness is several times too thick, mainly because the host of this channel thinks placing a blob on the CPU is the best way to spread it. The thinnest layer possible while coating the surface of the CPU is the best. You just want to eliminate unevenness at a microscopic level, so the very good thermally conducting aluminum surfaces can touch!
I usually put slightly less than the 2nd one in this video and spread it manually with my finger to make sure all cpu surface is covered and nothing spilled out when I put on the heatsink
Thermal paste is used to fill in tiny cracks between two metal surfaces. It actually has worse thermal conductivity than pure metal-to-metal contact, but of course much better conductivity than air. Once the air is pushed out, any excess paste simply spills out and that is that.
Yes, and there is so many people who don't understand it and spread nonsenses and myths. I have a lot of conflicts with these people all the time on internet, they call me a liar and boomer who knows nothing and say today it's totally diferent time than 20 years ago and my experiences mean nothing....so I guess laws of physics have changed during last 20 years. 😀 I actually had heating and cooling at school and these people very often don't even understand difference between heat in watts and temperature in °C, which is shocking. Most of these nonsense spreaders are AMD fanatics by the way, when you say it has 90°C in full load they call you a liar because their CPU has only 60 in game, they don't know what 100% load means and this is how stupid myths and nonsenses are spreading. 😀
I think we have another factor to consider, and that is the viscosity of the chosen thermal pasty. Yours is a somewhat runny one, I would guesstimate based on looking at the video. Some of them spread less easily and might turn into more of a problem in the "too much" scenario - though I haven't tested this.
yup. helped a friend diag a prebuilt with same gpu as me. swapped gpus. my mobo is fried. new mobo. got thick concrete like paste. debating just gettting another mobo and intel, and gpu. lol nothing about this makes sense, pc worked for years. first thermal paste was runny and did not give me hell like ths one.I'm almost positive there is too much. Luckily it's a decent processor from 2018. not an i9 or something.
I think the issue is more mounting pressure. Paste must fill air gaps, so if you've got poor pressure between your CPU and the cooler, then it's not going to cool well, because then the paste will be thick regardless and act as more of a barrier rather than a conductor for the microscopic air gaps that it's meant to fill.
I go on the safe side and apply 5 smaller dots, 1 in the middle and 4 around it but not too close to the edge. That way it spreads better. Ive found that to be the best result. You wont have a mess of it in the middle, but a more even spread throughout.
The CPU metal case is not engineered. It is not true flat, and if it is not true flat, then if looked at with magnifiction, you will see hills and valleys *grooves/bumps*. The paste is just to fill the valleys so there is an equal disappation.
Mechanical engineer here; I deal with all kinds of practical heat transfer problems daily. This was a lot of testing with no explanation of the relevant heat transfer issues. The conclusion of too much paste not impairing cooling cannot be correct. (1) Both the cooler and the chip surfaces have microscopic asperities, that in cross section look like an irregular sawtooth. Without thermal paste to eliminate the air gap, solid-to-solid contact is minimal, heat transfer is very inefficient and CPU temps will soar. (2) In the analysis assume no air gaps (death for CPUs because gases are poor thermal conductors compared to liquids and solids). (3) All heat transfer occurs by conduction through a chain, as follows: CPU --> Thermal Paste --> Copper base of CPU cooler. The copper in the CPU cooler has the lowest resistance to conduction of all the elements in the chain. The efficiency of the CPU cooler (more heat pipes, bigger fans and radiators) provides the temperature differential (or delta) that is the driving force for heat conduction from a warm body to a cooler body. (4) The thermal paste has lower thermal conductivity than the copper, but whatever it is, having a thicker layer adds resistance to heat flow because the heat has farther to travel. That does not become noticeable (measurable) until a fairly large excess is used, as shown in this video. My guideline is to spread a thin, semi-transparent layer using a razor blade, on both contacting surfaces. Seating the cooler properly on the CPU is crucial. BTW, I saw a detailed study that found little or no differences between brands of thermal paste. However you probably want to avoid off-brand stuff from Temu.
If the cooler is properly mounted, excess thermal paste would just be squeezed out the sides (like the example in this video). It’s not like there would be a materially thicker layer of thermal paste that the heat has to travel through.
Always hated applying thermal paste. Always made me paranoid. I remember a while ago, Sony I believe, developed a thermal paste alternative which was a graphite film I think. Shame it never caught on.
I typically just use 5 small dots of thermal paste on the CPU before I install my CPU fan and that method has never ever failed. The only other method I use is using about a grain of rice thermal paste and then installing the CPU cooler. Both methods appears to work just fine.... If you are unsure just watch a couple DIY CPU installation videos on RUclips and you'll get a good idea on how much thermal paste to use.
i thought for decades the standard agreed to be best method to apply thermal paste to CPU is the standard "pea shape/ammount" directly in center of CPU, no spreading by anything, and put a heathing with fan directly onto it... Is it still the best method...?
@@romannavratilid as as it spreads out enough to cover the contacted area between the heatsink and surface it doesn’t matter what pattern or amount you use
My i5 11400f hit 45-50c IDLE using stock cooler, im pretty sure the thermal paste is not a small amount, i have using various cpu and this is my first time seeing that temperature when idle
it can be because its a ''test bench'' where there is no case at all so it might do better at getting rid of the heat, and the room tempreture can be a factor as well; but don't take my word for it, i can be wrong
@@egordo7262 its just weird hearing the fan ramps up while doing literally nothing on desktop, well just gonna replace the stock cooler tho its never been great
I think the middle is the best. I did the same thing but what I did was actually spread the thermal paste around edge to edge of the cpu to transfer as much heat to the heatsink. I use an Evo212. In the end, I think when you put thermal paste on, you want to cover the entire surface with out spilling over.
That's what I tended to do. I never just placed a blob in the middle but rather in a pattern that had lines that covered the entire chip before placing the cooler down.
I usually just eat the thermal paste and then once my body processes it and turns it into compost then I take that compost and smear it onto the CPU and actually get a cooler CPU then just regular (non-compost) thermal paste. In fact, I was able to get the temperatures down on my CPU so much that I was able to overclock it to 6.9 GHz, AND IT'S AN i-1 100k 1 GHz CPU by default. If you don't believe me you should just try it sometime
this has been very interesting!, while it is kind of obvious that too much thermal paste can't lower the temperature below a certain point bc the fan can only get so much heat away it is nice to see that confirmed with an experiment. also thanks for adding the bit with the thermal past in the socket at the end bc again it's very reassuring to hear as someone who has rently build their first pc and is not sure if they did everything right
If you use thermal paste with a metal in it (some have copper or silver flakes) putting too much on can short your board. So if you are using something like Arctic Silver be careful
It dosent matter really, you dont need much to cover the plate, too much dosent matter it's going to squash regardlewss, it's electo insulative to so if it ooses over it dosent matter. I'd say use a teeny bit more than you think you need. I go for a 1cm blob in the middle.
For all of you using liquid metal paste that is conductive you must completely disregard this video. any of it touching your motherboard will short it and fry your system immediately
lol wait wait wait hold on you use liquid metal to destroy your computer that's on you buddy I use nactua thermal paste it's non conductive and harmless to the solder as such no issue with taking my mobo apart with thermal paste
Great video. I was worried about applying too much thermal paste, but now I see what worries is applying too little. Always good to have a middle ground.
When I run out of thermal paste a cut a small square of aluminium foil and place it on the cpu, being careful not to touch anything but the tip of the foil and the cpu and the cooler head has been cleaned completely. The foil can warp slightly to act like thermal paste and it works! Aluminium foil will also never have to be replaced.
@@bradleydcruze Probably. You're only supposed to use the thinnest layer of thermal paste possible to cover the surface. It should be much thinner than any of the blobs he placed on the chips in this video. The person saying he used aluminum foil probably did use the right amount of "thermal paste" as a piece of aluminum foil is very malleable and ductile, and will easily fill the microscopic voids between the chip and heat sink. Any Electrical Engineer will tell you the same thing.
@Depressed Aussie There's videos on youtube but they fold the aluminum in half and create air pockets. I tried it with my Clevo laptop with GTX 980m and i7 6700k but temps are very high. My heat sink probably isn't mounted properly or just doesn't have the force to press down the aluminum foil enough.
I've applied the thermal paste 3x on my first build and only the first time it was a little too much, still a little is on the mb but nothing bad happened and works just fine, after that I applied less and nothing leaked yet.
normally you would spread out the paste over the area manually before applying the heatsink. This gives 100% coverage and thickness is only to fill the roughness in the 2 metal surfaces. Without paste almost no contact area would result due to the hard and unflexible surfaces. So very very tiny layer is needed. Paste is not as good conductor of heat as metal and should be as thin as possible and ideally just to fill the "no contact" spaces between the 2 metal surfaces.
The best cooling option would be direct metal to metal. Unfortunately, this will never happen: the surfaces are too uneven; air will be trapped between, and air is a very good thermal insulator. So, you need thermal paste, but only so much as to replace the air between the CPU and the cooler. More paste will isolate again, as it is less heat-conducting than direct metal-to-metal. I usually use an amount close to your option no.1, and spread it around before I place the cooler. I will move the cooler aroud under pressure, so the paste will be forced to go where the two surfaces have a gap. I never had cooling issues from this spot...
I could see in the future really finely polished IHS and heatsinks that mate like gauge blocks, I think that would work really well (might still want a super tiny bit of thermal compound though)
@@TheAechBomb That would be ideal. You only need to replace the last bit of air by paste, not more. The parts should touch each other, and where they don't, a bit of paste is better than air...
Yes, there will always be small imperfections or honing/machining marks and the paste fills those imperfections with something thermally conductive. I spread a thin layer across the entire surface then tighten down the cooler until I don't get any movement of the cooler with moderate side pressure.
Test after test shows that excess paste just gets squeezed out. The "thermal insulator" thing only comes into play if you have insufficient clamping force on the heatsink mechanism.
@@TheAechBomb LOL gauge blocks, I have real gauge blocks and doubt we will ever get to that point because it cost way too much to do that with parts made separately. Cost would go up. Its a nice fantasy though.
I always put a little bit on, but then I spread it out with a small piece of plastic (an old bank card) to make sure every single corner is covered. Manny applications leave some space left at the corners. I'm amazed that so few people do this and just use the heatsink to spread the compound.
@@Kamtar34 Pretty sure it's measurable. If all the heat/cooling can be done 100% the die would also get cooler as a effect. If the cooler can only use 85% of the die it makes sense you lose (some) cooling performance.
I tested the paste on the Ryzen 6350.At idle, its temperature is 17-19 Celsius.I used Cooler Mastergel Maker paste.The temperature has been the same for more than 1 year.With Artic pastes and the like, processor temperatures were once or twice as high.There is a huge difference in quality between the traps.Example paste 2g for 4-6-7 dollars it can’t have the same quality as 2g paste for about $ 19.Cheap pastes are also all fake they are not original and they also suck.So when using a quality paste, spread is not very important but it should not be too small and not too large. The paste is designed to compensate for microscopically small irregularities in the refrigerator therefore it must be lubricated very thinly.If the paste is too thick a layer then the temperature is much higher.
How come such a simple basic (but very general user leveled) comparision took all these years to be found? One could easily think too much won't be as effective to evacuate hotair. Thanks Mr.
Because people never actually bothered to test their hypothesis as they were too afraid that they’d wreck their expensive hardware. So they just passed off their assumptions as fact and people took their word for it. It’s similar as to why people thought the Earth was flat and the sun revolves around the Earth for so long. Very few people stepped up to the plate to challenge what was considered common sense, and in some cases, the ones that did ended up suffering for it.
I love how google and reddit will tell you too much thermal paste will conduct electricity and short your system , while youtube will show you actual results that too much thermal paste is fine - just a pain to clean up .
Well yeah, there's only so much thermal paste that can compress before coming out the side. So, at some point close to the "normal" amount, after you tighten the cooler back down, you'll be left with a normal amount
@@jcspotter7322 they both work the same, the X and the dot are both viable options. your cpu might be very slightly covered a bit more with an X formation but the dot is easier to judge the amount
Thanks thanks thanks a lotttttt for the last part. I was really worried about thermal paste getting spilled into the Processor's pores. You cleared the doubt.
I really like the complete paste method, apply a thin layer all over the IHS using a spreader and then there is no part of the IHS that has a gap to to the cold plate of the cooler. This is how you can get consistency each time without uneven pumping/squeezing out of thermal paste..
On an LGA1700 socket, remove the stock intel bracket and replace it with a ThermalRight or Thermal Grizzly bracket and no paste will spill over onto the motherboard or socket connectors. Side note: Do NOT use liquid metal. Removal of the cooler will let liquid metal run all over and can screw up the socket or short the board if you don't get it cleaned up properly.
Think about it. He added more and more paste until eventually there was no further benefit. And that was achieved with normal paste. So how can using ceramic paste be any better? Even if the ceramic paste has better thermal conductivity, firstly he demonstrated you don't need it, second the elements involved in dissipating the heat is not just limited to the paste, it is also dependent on the design of the heatsink and performance of the fan attached to the heatsink. And once the chip is operating consistently within its normal operating temperature range, what benefit is there in trying to reduce the temperature further? The chip doesn't become faster because you reduce its temperature. It may give more room for overclocking, if that's the way you want to go. The temperature does affect the longevity of the chip, and chips that run too hot will have a shorter life but that shouldn't normally be an issue.
What a perfect timing of the video, watching this just after few hours of changing the thermal paste on my laptop cpu & gpu. I applied a little more than moderate amount since it already took so much time and hesitation to open everything up while back in the mind thinking about if this thing doesn't turn on again (yeah, my first time applying thermal paste and it went perfectly). Also laptop cpu and gpu surface are very small compared to a desktop version so that's why I was hesitating to put on more but in the end I did.
Can you do the same test but instead of just putting the cpu cooler on, use your finger to spread the thermal paste all over the IHS. I think the smallest amout of paste may do pretty well with the finger spread method.
You are exactly right, and it's provable. The thinnest layer of thermal paste that you can evenly spread over the surface, is the CORRECT amount. Unless you leave it in a blob like this guy did. I wouldn't post videos if I had his lack of knowledge and experience. I get his point that there's a such thing as too little or too much, but he used too much for the layer thickness in EVERY example.
It's never too much. I tested it myself, the excess is just squeezed out when you tighten the screws. I had max 1C temp difference between normal amount of paste and excessive amount of paste.
Define "never too much". I would say too much is the point at which adding more paste achieves no worthwhile benefit. And he demonstrated that point in the video. So yes, there is a point where too much paste can be added.
The paste is supposed to fill any gaps that maybe present between cooler plate and CPU. These gaps on older CPU's were basically tiny scratches so it was always best practice to apply a very small amount of paste and spread it around with a scraper or credit card then scrape off the excess leaving an extremely thin layer. What has changed that PC builder now apply a whole tube of the stuff and make a mess?
I'm sure this video will find out the same conclusion the rest I've seen did. There's never too much, only too little. The performance loss on using too much is negligible but the loss from too little is overwhelming.
I had a 3700x before and now upgrading it, noticed a dry spot since the grooves in my heatsink seem to eatup alot of it. Generosity with sanity is the way to go.
Thoroughly clean the underside of the heatsink and the mating surface of the cpu, with isopropyl alcohol if available. Rub a tiny amount of thermal paste well over the entirety of both mating surfaces. Fasten the heatsink on securely to ensure flush and tight mating. This will ensure optimum heat transfer from cpu to heatsink. Thermal paste is to fill in the microscopic imperfections of the metal that end up not contacting the mating surface. The ideal is metal to metal with no thermal paste in between. Thermal conductivity aluminium/copper > thermal conductivity thermal paste. But a metal/thermal paste/metal sandwich transfers heat much better than a metal/air/metal sandwich. Only reason to use more paste is if surfaces don't mate flush. I use dow corning heat sink compound.
Yeah, it doesn't actually take alot of thermal paste to work at all, especially if the CPU has an integrated heat spreader. You just have to apply it relatively evenly.
Hey mryeester! I also love doing experiments on pc's and i love your content.I ordered a older pc to do experiments on it and to learn also how to build a pc.I watch your videos everyday and your videos always make me smile.Keep up the good work😁
Great video. Im about to upgrade a 5 year old laptop (Dell G3) and will be pulling the double duty cooler from the soldered on CPU and GPU to put on new thermal paste. The fan runs more than I think it should when theres not really a load on the machine so cleaning the fans, heatpipes and applying new high quality thermal paste, as recommended by Tom's Guide. I would never consider using conductive thermal paste because I dont do PC work myself very often and worry I might put too much on and short/fry the motherboard with excess paste. Your video lets me know that being a bit overly generous with the paste won't hurt cooling, and choosing a non conductive paste means I won't zap the machine.
As long as your paste isn't conductive and your cooler has enough pressure to make good contact, then "too much" shouldn't matter. If you are using conductive thermal compound, then getting any squeezed out the sides could be a disaster. Just something to keep in mind if any of you out there decide to go the liquid metal route
I would never put thermal paste like that. One of the things I most enjoy while building or cleaning a desktop PC is to manually spread the paste all over the CPU until it is evenly distributed in a thin layer (I use CoolerMaster's accessory that comes with the thermal paste to do that). If you don't do that, you are letting the heatsink have all of the fun.
i read decades ago that you should spread the paste on CPU in any way, not by tools, not by finger (in a plastic bag/sack/film)... AFAIK the "pea shape" blob in center is the best method...?
When you read that you should not spread the paste do you recall any reasoning given for that? There is a lot of outdated advice about PCs that do not apply anymore, and many didn't even make sense when they became a thing.
Nice that you did the experiment - I recently made a very rough estimation of the force and pressure applied by the cooler once it's screwed on tightly and arrived at ~500-5000 N/cm^2, i.e. ~50-500 bar. [I used the recommended 0.6 N*m screw torque of my Noctua cooler, the number of screws, an estimated screw-thread density of 2/mm and ~1/8 to 1/2 screw rotation at full (0.6 N*m) torque as starting points] Since the "sides" are open, only 1 bar air pressure acts against the thermal paste squeezing out, plus a difficult to estimate amount of pressure due to the viscosity, capillary effect and surface tension of it. But I assume the latter to be
Nice and short video. I put something like between 2nd and 3rd option shown at 0:27 and never thought to check if it is correct. I'm glad I am not that far off from optimal.
Same, except I use a nitrile glove to prevent skin oil contamination. I have no idea if this makes a difference or not, but it seems to make me feel better, so whatever. lol
Been building PC's for years, even professionally for many of those, and always found that adding a little extra (and I do mean a little, not the entire tube, like here) for piece of mind was never a bad move... To this end, I've always liked the X application style, as sometimes a dot in the middle just doesn't spread to all the corners...
i tried the medium size dot for an aio. lets just say, temps went close to 90c under full load and only spread in the middle. spreading the paste covering the cpu kept the heat down around 70c under full load.
You are lying or something happened during your tests. It's just impossible to the application style make that huge difference, even using strange things like toothpast dont produce a huge difference like that.
Been building PCs since 1998 and I was told that too much would insulate heat so I've always used a paper thin layer on the entire contact surface. Good to know that too much doesn't effect it negatively other than goop everywhere.
Core 0 and 1 are running at 54°c idle is that a problem the gpu at 37°c idle My computer keeps restarting at random It's even worse on hot days I'm starting to see a pattern So i'm concerned it might be one of the two
You should have touched on the location of the actual CPU die beneath the heat spreader. If the design is monolithic, you will get a single silicon die that's right in the middle beneath the heat spreader. Such a CPU will still properly cool even if the edges of the heatspreader are not being covered with paste. However chiplet designs like what AMD uses may have a I/O die in the middle but the actual CPU die in the top left or top right. In that case, if you don't cover the corners of the heatspreader, you may experience massive temperature increases. So instead of worrying about how much paste you use, first know how the CPU silicon is actually laid out beneath the heatspreader or just cover the entire thing with paste using a plastic card or something.
Very simple. The tolerance of the heat sink screw down yields the same pressure, tension and distance when fully seated. So, after squeeze out you end with the same amount needed as a proper amount needed to operate the CPU at its optimal level.
Most manufacturers of CPU's have a document that explains the correct amount of thermal paste to use. My experience has always been to put a spot a little bigger than a cooked rice grain. I've also, manually spread the thermal paste over the surface of the CPU and have never had a bad experience with that.
Great comparison. I would like to see the same test but with the spreading technique. Looking at the volumes you applied, I typically apply in between the "small dot" and "moderate dot" (Arctic Silver 5 also) but when spread you get even coverage. Side note, if anyone reads this - please don't apply the large/excess dot. I spent 1hour cleaning thermal paste from my PCB when I first bought my workstation and to this day I am thankful I decided to check the CPU paste before booting it up!
Same here, I put a small dot and spread the paste with a credit card to make a thin layer that cover evenly the surface of the CPU. The thermals are good and no cleaning needed after. I want to see the small dot spread evenly how performs.
@@sebastian19745 In my hands that has just always worked better. Placing a blind drop never covers the whole IHS. One recommendation, place a small drop in the middle once done with spreading. It is mean to help reduce the chances of an air pocket forming in the middle. I also spread both the CPU IHS and the cooler mating surface to ensure a durable finish.
@@racerrrz That is what I do. About the final small drop I will test soon; it is approaching the time to clean my laptop. When the CPU is not very accessible, I just clean it and spread thermal paste on the cooler part only. I`m kinda lazy... But I never had thermal throttling or high temperatures in GPUs, laptops or desktops. Thanks for the tip.
@@sebastian19745 No trouble, I am certain it will help extend the life of your laptop! It does take ages to spread both the IHS and Cooler surfaces but I think it is worth it for the long service interval. I run my machine for 10 hours a day, likely 6 days/week, and the Arctic Silver 5 lasts ~2 years! Not the best thermal paste out there but it does the job, and it has been slower to fatigue from the pump-out effect compared with other pastes I have used.
@@racerrrz I've run CPU's for up to 5 years before and never seen thermal paste pump out. I've even used the stock paste that comes on coolers just fine.
I thought that more thermal paste would make it a little hotter but after seeing this I realized that this it is wrong because you are still tightening the CPU Cooler all the which makes the paste have about the same thickness with excess going onto the sides while I was thinking it would be more like a thermal pad.
I was getting thermal throttling on my CPU and decided to go through the pain of reapplying thermal paste and it was totally worth it. Dropped my temps massively and confirmed that I didn't put enough on there. The AM5 recommendation is actually to use a pea-sized glob in the middle and 4 tiny globs on the corners of the processor!
Just putting a blob in the middle then seating the cooler just makes a circle. If you were spread the past to the corners it would increase contact surface area and dissipate the heat better.
I've always used a plastic spreader to cover the entire contact point with a very thin layer of paste, then checked to be sure it has full contact to the cooler, never an issue..
Thanks! I recently installed a better CPU and found it getting up to 59 celcius. I had my suspicions that I maybe hadnt put enough paste on. I'll take it out, clean it up and give it another try. When I installed it I was worried about putting too much paste on but after your video I think I wont be so nervous.
الطريقة الثانية هي الافضل والصحيحة ويفضل ان تضغط قليلا على المبرد لانتشار المعجون الحراري بتوازن ولا تشغل جهاز pc الا بعد نصف ساعة او ساعة على الاقل وستحصل على نتائج مذهلة
It's the space between the cooler and Cpu. Once you press down, however much can't fit squeezes out the sides and you end up with the same thin layer of paste. You just need to use enough to cover entire surface of the Cpu. The thickness is irrelevant
Let's talk about what thermal paste actually does. Thermal paste serves as a thermal bridge between your CPU's heat spreader (the case protecting your delicate CPU), and the cooler on top of your CPU. Its job, to put it simply, is to eliminate all tiny air gaps, as air is a very poor conductor of heat. When ran with no paste, the heat will get trapped in these imperfections instead of transfer to your cooler.
![[FREE] Veeze Sample Type Beat - "Mohogany"](http://i.ytimg.com/vi/FNZc-I-FM9Q/mqdefault.jpg)
Just goes to show that if you're unsure, using more thermal paste than you think you need is better than using less than you think.
i did that to my laptop, i applied what seems like the right amount but was not getting the temps I want, then I just put a giant blob and it was cooler
Yes exactly, however when you're using liquid metal or an electrically conductive thermal paste (like the Arctic Silver 5) you have to make sure the compound doesn't leak onto the motherboard when screwing back the heatsink. Most thermal compounds like the Arctic MX-4 or Noctua NT-H1 however are not conductive so this isn't a problem in most cases
Yeah, if you don't understand physics.
@@tcolondovich2996 .... all i did was apply the amount of thermal paste the videos recommended and that didn't work so I added more.
@@vanskis7618 That's completely beside the points I've made. I'm not sure if you read my comments at all.
The fact that a tiny amount of thermal paste prevents thermal shutdown was nuts to me
I know its kinda nuts but if you think about it metal to metal, no matter how flat they are, its always more gaps than real contact.
The die under the IHS is much smaller than the IHS, the little dot covered most of the area where the die under the IHS is located, which is what matters the most, but it is always better to have most of the IHS covered in paste to optimize heat transfer
Its preventing the hot metal from the heat sink to make contact to the cpu.
@@dedpul5071 It's the other way around, the heat sink is supposed to sink (move) heat away from the CPU (where the heat is coming from.)
@@wizzrobe3896 sorry, thats what i meant.
Good experiment. Little or no filler, straight to the point with no mucking about. This would be really useful to someone who hasn't built a PC before.
That's how I ended up here 🫡
Chances are barely anybody will encounter this but there are super thick thermal pastes, which do yield worse temperatures when applied too thick! Specifically for instance the Phobya Nanogrease. In my tests the single best thermal paste outside of liquid metal and that stuff is so so thick, that you can't even use the blop method. However, when applied correctly awesome results and that explains (to me) why there's so much deviation between tests concerning this paste.
But if it gets into your motherboard, you are royally screwed.
@@dylansmidt6853 Thats only if you use conductive thermal paste
I mean he used an unholy amount of it on this short ruclips.net/video/vUOawJ7XkZY/видео.html&ab_channel=mryeester
What about the artic mx 6, i used it on laptop GPU and I couldn't spread it evenly as it was very thick so I put a bit more than moderate amount but not too much
Essentially you just need enough to cover the surface area (or the vast majority of it) in order to have even heat transfer. As long as you achieve that there isn’t much more to do
Spread it around like butter on bread, corner to corner.
@@nunya3163do that if you want to have to clean off the motherboard cpu housing!
@@nunya3163 That is what im doing for past 15 years. I just like to be sure that its spread evenly over the cpu, regardless if heatsink covers whole cpu. Excess of the paste will spread on parts here its too thin and its all good.
@@nunya3163 Agree. I've always put a medium bead in the center then used my finger to finely spread around the entire surface. My machines always worked well and stayed cool for the entire time I've owned them. Never having the need to redo the job plus I never got any excess over the sides.
I accidentally put too much thermal paste on my acer nitro 5 and for the first week it was perfect but after it started getting more and more worse ending actually more worse than old thermal paste. It seems to me that at the beginning it is good because the paste is still in a condition like from the "packaging", but after use it starts to harden slightly, which causes hardening on the surface of the processor and the heat sink, and there is still a "packaging" consistency in between, which blocks the heat flow and keeps the heat between the heat sink and the processor (and its MORE LIKELY problem of processors without IHS like in laptops)
it also depends on what paste it was, crappy thermal compounds will harden quite fast.
@@baraka629 arctic mx4
@@Interxoxo Bro. Use Colgate or Pepsodent. There the best.
@@IamGod534 Don't, I used it once on my PC and forgot to change it out for real thermal paste and after a year, mould formed on my IHS. After cleaning it up the surface was smooth but the colour of the IHS turned green in some spots, other than that minor cosmetic damage, it didn't effect thermals as the surface was relatively flat.
That could explain my findings, I guess... I always used a bit more rather than a bit less, but my temps would get higher than they should be. Removing a bit of paste has always done the trick so far!
what would happen if you just decided to NOT put the paste onto the board and instead put it in your mouth?
😂
Haha too funny
Brain freeze, do not try.
That would prevent YOU from overheating. But your motherboard will be toast.
both will burn down to ashes
Thanks for this video. I always thought I really damage my motherboard and CPU actually just because a bit of thermal Paste in the socket but now I don‘t even think about it anymore so thank you very much!
nope you won't it's just a bit messy that's all
Some thermal paste is electrically conductive and can potentially cause damage if it gets onto components, so it is something you should definitely still keep in mind when choosing and applying thermal paste.
@@Sevicify Well let’s hope Be Quiet!‘s thermal paste isn’t electrical conductive 😅. But as my computer runs as normal for about half a year now I don‘t think it’s going to be a problem in the future. But thank you!👍
Its better to use too much than too little. Especially when using open die but at that point i just spread it.
honestly no one cares that's it in a nutshell🤣🤣
SO many videos have shown that amount of paste and pattern are almost irrelevant as long as the whole of the IHS is covered. I always use the 'butter' method with a thin layer of MX-4. 'Better' pastes or liquid metal make 2 or 3 degrees difference at best, and that's on Prime95 Max Heat which will never be hit in normal running.
same for me-thin layer of paste (almost feels transparent) and paste works good for years.
Paste is only for surface imperfection, radiators blocks are almost perfectly machined nowadays. Even that small amount of paste would work if he would spread it before instaling radiator. Thats my 5 cents.
@@szaka9395 JayZ2Cents did a test from one spot to 1/2 a tube. Made no noticeable difference (except to the 'overflow' of paste :) ). GPU's especially don't have an IHS so 'spreading' is vital. I found that out when watercooling my 3090!
"better" is indeed not much better than "good," but "good" is a whole lot better than "mediocre" and below though.
@@MrKippie15 Agreed, but you don't need to go beyond MX-4 or my old favourite Arctic Silver. 2-3 degrees for liquid metal just isn't worth it IMO.
0:44 forbidden love
with exchange of bodily fluids
enemies to lovers
And now, make them kiss
I think using more and more won’t have an effect because the mounting pressure of the cpu cooler will be the same making the layer between cpu and cooler the same and just pushing more put to the sides
Agreed, I just put a large amount on.
I put a little bit, spread it then put a bit again in the middle.
I always spread the thermal paste evenly over the *entire area* of the top of the CPU case. Otherwise I would not know whether the "squishing" caused by the pressure of the heat sink has been adequate to reach the edges of the CPU case. Also I believe that the layer of thermal paste should be *as thin as possible* . It is only there to fill the microscopic roughness between molecules of the metal surfaces. The thermal conductivity of thermal paste would be *much inferior* to that of the metal itself.
The problem with manually spreading it even is you run the risk of trapping air bubbles instead of having a tight spread. It's not horrible but not optimal either.
can a heatsink be too tight? Im doing a passion project, using a big heatsink for from a dell 2400 Dimension onto a Gigabyte motherboard, the hold-down level is tight due to the height of the heatsink.. can this be bad? Would a tight join between sink and cpu affect the heat and paste conductivity?
@@Defender78. I'm inclined to say that the tighter the better - within reason. Don't break or bend anything! Tightness would tend to guarantee that the layer of thermal paste is as thin as possible. It's only there to compensate for the fact that the metal surfaces are less than perfectly flat.
Exactly. I can confirm it from my own experience. Recently I was forced to re-paste my Ryzen 5 7600 because the mobo went wrong (and repaired under warranty). Although I replaced the mobo because it went wrong again as far as I can judge it does not make a difference. Last time I used a lot less paste (same Noctua NT-H1 AM5 Edition as before, did spread it thinly evenly) and the max. temperature during Cinebench R23 dropped by 10 Celsius (!).
Thus I concluded the same, both quantity and application matter and the best way is to spread a very thin layer of paste.
It would be interesting to see how this compares to making the top of the IHS, and the cooler plate, totally flat and smooth by sanding them down.
I mean, in reality a tiny amount of thermal paste (or none) should be just fine if both surfaces are totally flat. You just need more if your CPU has a convex or a concave shape.
When I apply thermal paste I generally like to wiggle the cooler around a lot on the CPU to make sure it gets to the right places.
what happens is simple you make a mess that's what happens otherwise it's all good
@@raven4k998 You mean because the paste all squishes out the sides?
I did it with my old 4790k a few years ago, but I can't remember what my before and after temps were, but I do remember that it was quite concave in the centre.
Might give it a shot with my 5800x sometime, because it does get pretty toasty.
P.S. Nice profile pic, she's one of my favourites!
@@JoeBob79569 There has been benchmarks by a bunch of the "world records" setting nerds out there. Basically, polishing those components DOES increase performance, but it's very small margin. It's not worth it unless every other solution (better chokes for less voltage variance, minimizing said voltage for your clock speed, a more efficient cooler) has already been accomplished, especially with the risk of damage on the IHS or the cost of having someone else do the work for you. Hell the cost of Binned vs non-Binned CPU's to ensure you have a more stable chipset (and can run lower voltages) is going to benefit you more.
@@asymsolutions I don't think it's that risky, or that difficult to do it. Just grab a sheet of glass and put a sheet of sandpaper on it. And rub the CPU around on it for a few minutes.
You'd maybe get 2-3 degrees better on this CPU. Just a number I pulled out my behind, but it reflects similar tests that some tech channels have done. Honestly, not really worth it. Not unless you've got a golden chip, and at that point you are probably setting a record as your average joe has maybe a 1% chance at best to stumble onto a chip that good.
I spread a very thin smooth amount onto the cooler, then apply a small(rice) to medium(pea size) to the cpu. This works perfect everytime, good even coverage, little to no overflow, and pretty much dummy proof method.
so bottom line there is a point of diminishing returns with thermal paste🤣🤣🤣🤣🤣🤣🤪🤪🤪🤪🤪🤪🤪🤪🤪🤪🤪🤪🤪🤪🤪🤪🤪🤪🤪🤪🤪🤪🤪🤪
@@raven4k998special education comment
The medium amount already covers the whole die once pressed down, and the large amount covers the same amount of area and thereby, having the same thermal performance. Conclusion: the performance isn't related to the amount of paste in volume but how much area it covers once pressed down. Simple as that, what's there to investigate?
pregnancy is government tho
The only thing I can think of is just that too much is ok just messy? It's an age old argument about how the thermal paste should be applied, how much ect
Some people believe too much thermal paste actually degrades performance. This video neatly dispels that.
@@RK-252 the reasoning behind that is because the thermal paste is actually less heat conductive than the IHS and the copper/metal plate of the cooler, therefore actually creating more insulation for the CPU and it being able to dissipate heat slower. However, when you screw in the cooler correctly, the pressure will press out any excess thermal paste, thus removing the insulating factor.
@@truefeelke yup, exactly right.
Send this to a newer pc builder to help them with Thermal Paste
💀
@@justin7837 ???
@@justin7837 ?
Why would you send them a video showing this? Layer thickness is several times too thick, mainly because the host of this channel thinks placing a blob on the CPU is the best way to spread it. The thinnest layer possible while coating the surface of the CPU is the best. You just want to eliminate unevenness at a microscopic level, so the very good thermally conducting aluminum surfaces can touch!
@@tcolondovich2996 bruhhh
I usually put slightly less than the 2nd one in this video and spread it manually with my finger to make sure all cpu surface is covered and nothing spilled out when I put on the heatsink
I do the same way - my finger is in thin plastic bag 🙂
Just don't shoot the entire tube on the poor thing.
Thermal paste is used to fill in tiny cracks between two metal surfaces. It actually has worse thermal conductivity than pure metal-to-metal contact, but of course much better conductivity than air. Once the air is pushed out, any excess paste simply spills out and that is that.
Yes, and there is so many people who don't understand it and spread nonsenses and myths. I have a lot of conflicts with these people all the time on internet, they call me a liar and boomer who knows nothing and say today it's totally diferent time than 20 years ago and my experiences mean nothing....so I guess laws of physics have changed during last 20 years. 😀 I actually had heating and cooling at school and these people very often don't even understand difference between heat in watts and temperature in °C, which is shocking.
Most of these nonsense spreaders are AMD fanatics by the way, when you say it has 90°C in full load they call you a liar because their CPU has only 60 in game, they don't know what 100% load means and this is how stupid myths and nonsenses are spreading. 😀
I think we have another factor to consider, and that is the viscosity of the chosen thermal pasty. Yours is a somewhat runny one, I would guesstimate based on looking at the video. Some of them spread less easily and might turn into more of a problem in the "too much" scenario - though I haven't tested this.
yup. helped a friend diag a prebuilt with same gpu as me. swapped gpus. my mobo is fried. new mobo. got thick concrete like paste. debating just gettting another mobo and intel, and gpu. lol nothing about this makes sense, pc worked for years. first thermal paste was runny and did not give me hell like ths one.I'm almost positive there is too much. Luckily it's a decent processor from 2018. not an i9 or something.
I think the issue is more mounting pressure. Paste must fill air gaps, so if you've got poor pressure between your CPU and the cooler, then it's not going to cool well, because then the paste will be thick regardless and act as more of a barrier rather than a conductor for the microscopic air gaps that it's meant to fill.
In no case did you apply the paste correctly. Correctly is the thinnest possible spread on both cpu and sink and then join.
You are correct, no one knows how to apply paste anymore.
I go on the safe side and apply 5 smaller dots, 1 in the middle and 4 around it but not too close to the edge. That way it spreads better. Ive found that to be the best result. You wont have a mess of it in the middle, but a more even spread throughout.
Cleaver and safe 👍
The CPU metal case is not engineered. It is not true flat, and if it is not true flat, then if looked at with magnifiction, you will see hills and valleys *grooves/bumps*.
The paste is just to fill the valleys so there is an equal disappation.
Mechanical engineer here; I deal with all kinds of practical heat transfer problems daily.
This was a lot of testing with no explanation of the relevant heat transfer issues. The conclusion of too much paste not impairing cooling cannot be correct.
(1) Both the cooler and the chip surfaces have microscopic asperities, that in cross section look like an irregular sawtooth. Without thermal paste to eliminate the air gap, solid-to-solid contact is minimal, heat transfer is very inefficient and CPU temps will soar.
(2) In the analysis assume no air gaps (death for CPUs because gases are poor thermal conductors compared to liquids and solids).
(3) All heat transfer occurs by conduction through a chain, as follows: CPU --> Thermal Paste --> Copper base of CPU cooler. The copper in the CPU cooler has the lowest resistance to conduction of all the elements in the chain. The efficiency of the CPU cooler (more heat pipes, bigger fans and radiators) provides the temperature differential (or delta) that is the driving force for heat conduction from a warm body to a cooler body.
(4) The thermal paste has lower thermal conductivity than the copper, but whatever it is, having a thicker layer adds resistance to heat flow because the heat has farther to travel. That does not become noticeable (measurable) until a fairly large excess is used, as shown in this video.
My guideline is to spread a thin, semi-transparent layer using a razor blade, on both contacting surfaces. Seating the cooler properly on the CPU is crucial.
BTW, I saw a detailed study that found little or no differences between brands of thermal paste. However you probably want to avoid off-brand stuff from Temu.
This toke me the entire vid to read your comment
Thank you for the insight
If the cooler is properly mounted, excess thermal paste would just be squeezed out the sides (like the example in this video). It’s not like there would be a materially thicker layer of thermal paste that the heat has to travel through.
Your "guideline" works if we are talking about a laptop CPU where there is little pressure and usually irregular shape (ie non square ish CPU).
This is a very great explanation behind working of thermal paste.
Thank you. This is a great video for someone who is overly anxious about applying thermal paste!
just use the whole tube it'll be fine🤪🤪
Same tbh
Always hated applying thermal paste. Always made me paranoid.
I remember a while ago, Sony I believe, developed a thermal paste alternative which was a graphite film I think. Shame it never caught on.
I typically just use 5 small dots of thermal paste on the CPU before I install my CPU fan and that method has never ever failed. The only other method I use is using about a grain of rice thermal paste and then installing the CPU cooler. Both methods appears to work just fine.... If you are unsure just watch a couple DIY CPU installation videos on RUclips and you'll get a good idea on how much thermal paste to use.
i thought for decades the standard agreed to be best method to apply thermal paste to CPU is the standard "pea shape/ammount" directly in center of CPU, no spreading by anything, and put a heathing with fan directly onto it... Is it still the best method...?
@@romannavratilid as as it spreads out enough to cover the contacted area between the heatsink and surface it doesn’t matter what pattern or amount you use
@@romannavratilid we no longer have just square CPU's so 5 dots is good for newer intel CPU's
My i5 11400f hit 45-50c IDLE using stock cooler, im pretty sure the thermal paste is not a small amount, i have using various cpu and this is my first time seeing that temperature when idle
my 5800x aswell
it can be because its a ''test bench'' where there is no case at all so it might do better at getting rid of the heat, and the room tempreture can be a factor as well; but don't take my word for it, i can be wrong
its not unusual for those temps, check some websites or ask some people what temps they get with a stock cooler and an i5 11400f
@@egordo7262 its just weird hearing the fan ramps up while doing literally nothing on desktop, well just gonna replace the stock cooler tho its never been great
@@alvag6 yeah i would just replace it with sometjing else. What cooler are you buying?
I think the middle is the best. I did the same thing but what I did was actually spread the thermal paste around edge to edge of the cpu to transfer as much heat to the heatsink. I use an Evo212. In the end, I think when you put thermal paste on, you want to cover the entire surface with out spilling over.
That's what I tended to do. I never just placed a blob in the middle but rather in a pattern that had lines that covered the entire chip before placing the cooler down.
Very quick and comprehensive, just what I like when I want to get informed on a subject. Thanks!
I usually just eat the thermal paste and then once my body processes it and turns it into compost then I take that compost and smear it onto the CPU and actually get a cooler CPU then just regular (non-compost) thermal paste. In fact, I was able to get the temperatures down on my CPU so much that I was able to overclock it to 6.9 GHz, AND IT'S AN i-1 100k 1 GHz CPU by default. If you don't believe me you should just try it sometime
this has been very interesting!, while it is kind of obvious that too much thermal paste can't lower the temperature below a certain point bc the fan can only get so much heat away it is nice to see that confirmed with an experiment. also thanks for adding the bit with the thermal past in the socket at the end bc again it's very reassuring to hear as someone who has rently build their first pc and is not sure if they did everything right
Omg, sameee
I always do what the Noctua manual says, one pea in the middle, 4 tiny peas in each corner with some offset (like a 5 on a die).
If you use thermal paste with a metal in it (some have copper or silver flakes) putting too much on can short your board. So if you are using something like Arctic Silver be careful
It dosent matter really, you dont need much to cover the plate, too much dosent matter it's going to squash regardlewss, it's electo insulative to so if it ooses over it dosent matter. I'd say use a teeny bit more than you think you need. I go for a 1cm blob in the middle.
For all of you using liquid metal paste that is conductive you must completely disregard this video. any of it touching your motherboard will short it and fry your system immediately
lol wait wait wait hold on you use liquid metal to destroy your computer that's on you buddy I use nactua thermal paste it's non conductive and harmless to the solder as such no issue with taking my mobo apart with thermal paste
@@raven4k998 cool this obviously wasn't for you but thats a huge piece of information for you to leave out
Great video. I was worried about applying too much thermal paste, but now I see what worries is applying too little. Always good to have a middle ground.
When I run out of thermal paste a cut a small square of aluminium foil and place it on the cpu, being careful not to touch anything but the tip of the foil and the cpu and the cooler head has been cleaned completely. The foil can warp slightly to act like thermal paste and it works! Aluminium foil will also never have to be replaced.
Wait this 100% works??
@@bradleydcruze Probably. You're only supposed to use the thinnest layer of thermal paste possible to cover the surface. It should be much thinner than any of the blobs he placed on the chips in this video. The person saying he used aluminum foil probably did use the right amount of "thermal paste" as a piece of aluminum foil is very malleable and ductile, and will easily fill the microscopic voids between the chip and heat sink. Any Electrical Engineer will tell you the same thing.
@@bradleydcruze No.
@Depressed Aussie There's videos on youtube but they fold the aluminum in half and create air pockets. I tried it with my Clevo laptop with GTX 980m and i7 6700k but temps are very high. My heat sink probably isn't mounted properly or just doesn't have the force to press down the aluminum foil enough.
I've applied the thermal paste 3x on my first build and only the first time it was a little too much, still a little is on the mb but nothing bad happened and works just fine, after that I applied less and nothing leaked yet.
don't be stingy overkill won't fail like underkill may
Make a pattern writing Intel on an AMD chip in AMD logo's style, use a intel heat sink as well.
Ultimate universal collapse
normally you would spread out the paste over the area manually before applying the heatsink. This gives 100% coverage and thickness is only to fill the roughness in the 2 metal surfaces. Without paste almost no contact area would result due to the hard and unflexible surfaces. So very very tiny layer is needed. Paste is not as good conductor of heat as metal and should be as thin as possible and ideally just to fill the "no contact" spaces between the 2 metal surfaces.
The best cooling option would be direct metal to metal. Unfortunately, this will never happen: the surfaces are too uneven; air will be trapped between, and air is a very good thermal insulator.
So, you need thermal paste, but only so much as to replace the air between the CPU and the cooler. More paste will isolate again, as it is less heat-conducting than direct metal-to-metal.
I usually use an amount close to your option no.1, and spread it around before I place the cooler. I will move the cooler aroud under pressure, so the paste will be forced to go where the two surfaces have a gap. I never had cooling issues from this spot...
I could see in the future really finely polished IHS and heatsinks that mate like gauge blocks, I think that would work really well (might still want a super tiny bit of thermal compound though)
@@TheAechBomb That would be ideal. You only need to replace the last bit of air by paste, not more. The parts should touch each other, and where they don't, a bit of paste is better than air...
Yes, there will always be small imperfections or honing/machining marks and the paste fills those imperfections with something thermally conductive. I spread a thin layer across the entire surface then tighten down the cooler until I don't get any movement of the cooler with moderate side pressure.
Test after test shows that excess paste just gets squeezed out. The "thermal insulator" thing only comes into play if you have insufficient clamping force on the heatsink mechanism.
@@TheAechBomb LOL gauge blocks, I have real gauge blocks and doubt we will ever get to that point because it cost way too much to do that with parts made separately. Cost would go up. Its a nice fantasy though.
Thank you for demonstrating the parameters of application
I always put a little bit on, but then I spread it out with a small piece of plastic (an old bank card) to make sure every single corner is covered.
Manny applications leave some space left at the corners. I'm amazed that so few people do this and just use the heatsink to spread the compound.
Because you are trying to cooldown the die, not the metal cover on the CPU. Leaving out corners of the metal shielding wont do that much.
@@Kamtar34 Pretty sure it's measurable. If all the heat/cooling can be done 100% the die would also get cooler as a effect. If the cooler can only use 85% of the die it makes sense you lose (some) cooling performance.
I tested the paste on the Ryzen 6350.At idle, its temperature is 17-19 Celsius.I used Cooler Mastergel Maker paste.The temperature has been the same for more than 1 year.With Artic pastes and the like, processor temperatures were once or twice as high.There is a huge difference in quality between the traps.Example paste 2g for 4-6-7 dollars it can’t have the same quality as 2g paste for about $ 19.Cheap pastes are also all fake they are not original and they also suck.So when using a quality paste, spread is not very important but it should not be too small and not too large.
The paste is designed to compensate for microscopically small irregularities in the refrigerator therefore it must be lubricated very thinly.If the paste is too thick a layer then the temperature is much higher.
How come such a simple basic (but very general user leveled) comparision took all these years to be found?
One could easily think too much won't be as effective to evacuate hotair.
Thanks Mr.
Because people never actually bothered to test their hypothesis as they were too afraid that they’d wreck their expensive hardware. So they just passed off their assumptions as fact and people took their word for it.
It’s similar as to why people thought the Earth was flat and the sun revolves around the Earth for so long. Very few people stepped up to the plate to challenge what was considered common sense, and in some cases, the ones that did ended up suffering for it.
I love how google and reddit will tell you too much thermal paste will conduct electricity and short your system , while youtube will show you actual results that too much thermal paste is fine - just a pain to clean up .
Well yeah, there's only so much thermal paste that can compress before coming out the side. So, at some point close to the "normal" amount, after you tighten the cooler back down, you'll be left with a normal amount
if pressure is too low it will hit performance becouse layer of paste would bee too thick
last bit about thermal paste seeping under the cpu was super informative, thank you for taking your time to make this video cheers
This is why I do the X pattern. Just run from corner to corner leaving 3mm from each corner. Perfect CPU temperatures. First build! 🙌
Orrrr, you could just do a dot in the middle to keep the paste from spilling over
@@jcspotter7322 they both work the same, the X and the dot are both viable options. your cpu might be very slightly covered a bit more with an X formation but the dot is easier to judge the amount
@@joshdeconcentrated2674 you missed my point, it literally was to suggest a method just to prevent the paste from spilling over
@@jcspotter7322 the X doesn't spill over though
Thanks thanks thanks a lotttttt for the last part. I was really worried about thermal paste getting spilled into the Processor's pores. You cleared the doubt.
I really like the complete paste method, apply a thin layer all over the IHS using a spreader and then there is no part of the IHS that has a gap to to the cold plate of the cooler. This is how you can get consistency each time without uneven pumping/squeezing out of thermal paste..
That's the right way to apply thermal compound. But it takes some effort so no one's going to do that. Because they're all a bunch of lazy bastards.
On an LGA1700 socket, remove the stock intel bracket and replace it with a ThermalRight or Thermal Grizzly bracket and no paste will spill over onto the motherboard or socket connectors.
Side note: Do NOT use liquid metal. Removal of the cooler will let liquid metal run all over and can screw up the socket or short the board if you don't get it cleaned up properly.
What about ceramic thermal paste? How much does it differ from liquid metal and is it a bigger difference in keeping temps down
Think about it.
He added more and more paste until eventually there was no further benefit.
And that was achieved with normal paste.
So how can using ceramic paste be any better?
Even if the ceramic paste has better thermal conductivity, firstly he demonstrated you don't need it, second the elements involved in dissipating the heat is not just limited to the paste, it is also dependent on the design of the heatsink and performance of the fan attached to the heatsink.
And once the chip is operating consistently within its normal operating temperature range, what benefit is there in trying to reduce the temperature further? The chip doesn't become faster because you reduce its temperature. It may give more room for overclocking, if that's the way you want to go. The temperature does affect the longevity of the chip, and chips that run too hot will have a shorter life but that shouldn't normally be an issue.
What a perfect timing of the video, watching this just after few hours of changing the thermal paste on my laptop cpu & gpu. I applied a little more than moderate amount since it already took so much time and hesitation to open everything up while back in the mind thinking about if this thing doesn't turn on again (yeah, my first time applying thermal paste and it went perfectly).
Also laptop cpu and gpu surface are very small compared to a desktop version so that's why I was hesitating to put on more but in the end I did.
Can you do the same test but instead of just putting the cpu cooler on, use your finger to spread the thermal paste all over the IHS. I think the smallest amout of paste may do pretty well with the finger spread method.
Lmao finger spread method. I’m very familiar with this method but not in CPU terms🤔🤣
@@bigdog8989 it works tho
@@bigdog8989 woah 😳😳
You are exactly right, and it's provable. The thinnest layer of thermal paste that you can evenly spread over the surface, is the CORRECT amount. Unless you leave it in a blob like this guy did. I wouldn't post videos if I had his lack of knowledge and experience. I get his point that there's a such thing as too little or too much, but he used too much for the layer thickness in EVERY example.
@@tcolondovich2996 Sigh... the pressure will flatten the thermal paste, that's why spreading out thermal paste is not necessary.
Nice and short and straight to the point. Thanks dude, very informative for a noob who is freaking out doing this for the first time!
It's never too much. I tested it myself, the excess is just squeezed out when you tighten the screws. I had max 1C temp difference between normal amount of paste and excessive amount of paste.
Define "never too much".
I would say too much is the point at which adding more paste achieves no worthwhile benefit. And he demonstrated that point in the video.
So yes, there is a point where too much paste can be added.
I've seen a local pro applying paste, and it was a huge amount. Seems like he knew there's never too much.
The paste is supposed to fill any gaps that maybe present between cooler plate and CPU. These gaps on older CPU's were basically tiny scratches so it was always best practice to apply a very small amount of paste and spread it around with a scraper or credit card then scrape off the excess leaving an extremely thin layer. What has changed that PC builder now apply a whole tube of the stuff and make a mess?
I'm sure this video will find out the same conclusion the rest I've seen did. There's never too much, only too little. The performance loss on using too much is negligible but the loss from too little is overwhelming.
@@manchuriancandidate8907 Bro stfu. I work with computers and no customer has ever broke a cpu with paste
@@manchuriancandidate8907 Very hard to do if the CPU is in the socket. Also, thermal paste is usually not conductive
For me the less the better. I used too much and my temperatures went up. Small dot, spread it evenly, instal radiator and forget for a year.
I had a 3700x before and now upgrading it, noticed a dry spot since the grooves in my heatsink seem to eatup alot of it. Generosity with sanity is the way to go.
Use power fulll fan with thermal syastems and heat shrink try it your old pc will blow out by overclocking
Thoroughly clean the underside of the heatsink and the mating surface of the cpu, with isopropyl alcohol if available. Rub a tiny amount of thermal paste well over the entirety of both mating surfaces. Fasten the heatsink on securely to ensure flush and tight mating. This will ensure optimum heat transfer from cpu to heatsink. Thermal paste is to fill in the microscopic imperfections of the metal that end up not contacting the mating surface. The ideal is metal to metal with no thermal paste in between. Thermal conductivity aluminium/copper > thermal conductivity thermal paste. But a metal/thermal paste/metal sandwich transfers heat much better than a metal/air/metal sandwich. Only reason to use more paste is if surfaces don't mate flush. I use dow corning heat sink compound.
Thank you. You really applied science simply
Yeah, it doesn't actually take alot of thermal paste to work at all, especially if the CPU has an integrated heat spreader. You just have to apply it relatively evenly.
Hey mryeester!
I also love doing experiments on pc's and i love your content.I ordered a older pc to do experiments on it and to learn also how to build a pc.I watch your videos everyday and your videos always make me smile.Keep up the good work😁
Great video. Im about to upgrade a 5 year old laptop (Dell G3) and will be pulling the double duty cooler from the soldered on CPU and GPU to put on new thermal paste.
The fan runs more than I think it should when theres not really a load on the machine so cleaning the fans, heatpipes and applying new high quality thermal paste, as recommended by Tom's Guide.
I would never consider using conductive thermal paste because I dont do PC work myself very often and worry I might put too much on and short/fry the motherboard with excess paste.
Your video lets me know that being a bit overly generous with the paste won't hurt cooling, and choosing a non conductive paste means I won't zap the machine.
As long as your paste isn't conductive and your cooler has enough pressure to make good contact, then "too much" shouldn't matter.
If you are using conductive thermal compound, then getting any squeezed out the sides could be a disaster. Just something to keep in mind if any of you out there decide to go the liquid metal route
too much rather than too little, at least its hard to screw that up
I would never put thermal paste like that. One of the things I most enjoy while building or cleaning a desktop PC is to manually spread the paste all over the CPU until it is evenly distributed in a thin layer (I use CoolerMaster's accessory that comes with the thermal paste to do that). If you don't do that, you are letting the heatsink have all of the fun.
i read decades ago that you should spread the paste on CPU in any way, not by tools, not by finger (in a plastic bag/sack/film)... AFAIK the "pea shape" blob in center is the best method...?
When you read that you should not spread the paste do you recall any reasoning given for that? There is a lot of outdated advice about PCs that do not apply anymore, and many didn't even make sense when they became a thing.
Tests have shown this is the way to go.
Nice that you did the experiment - I recently made a very rough estimation of the force and pressure applied by the cooler once it's screwed on tightly and arrived at ~500-5000 N/cm^2, i.e. ~50-500 bar. [I used the recommended 0.6 N*m screw torque of my Noctua cooler, the number of screws, an estimated screw-thread density of 2/mm and ~1/8 to 1/2 screw rotation at full (0.6 N*m) torque as starting points]
Since the "sides" are open, only 1 bar air pressure acts against the thermal paste squeezing out, plus a difficult to estimate amount of pressure due to the viscosity, capillary effect and surface tension of it. But I assume the latter to be
Realistically the best method is to use a spreader to make sure you get full coverage for the best heat transfer.
Nice and short video. I put something like between 2nd and 3rd option shown at 0:27 and never thought to check if it is correct. I'm glad I am not that far off from optimal.
I just apply a small amount and spread it with my finger across my i7 4770
Same, except I use a nitrile glove to prevent skin oil contamination. I have no idea if this makes a difference or not, but it seems to make me feel better, so whatever. lol
As someone kinda new to pc building, this helped me a lot. Thanks!
Been building PC's for years, even professionally for many of those, and always found that adding a little extra (and I do mean a little, not the entire tube, like here) for piece of mind was never a bad move... To this end, I've always liked the X application style, as sometimes a dot in the middle just doesn't spread to all the corners...
The X is definitely my favorite method, since an X will naturally expand into a square when it gets pressed down by the CPU cooler.
The corners don't get as hot so that's not an issue..
The only thing thermal does is flatten the surface.
What happens if you use an egg instead of thermal paste
i tried the medium size dot for an aio. lets just say, temps went close to 90c under full load and only spread in the middle. spreading the paste covering the cpu kept the heat down around 70c under full load.
You are lying or something happened during your tests. It's just impossible to the application style make that huge difference, even using strange things like toothpast dont produce a huge difference like that.
Been building PCs since 1998 and I was told that too much would insulate heat so I've always used a paper thin layer on the entire contact surface. Good to know that too much doesn't effect it negatively other than goop everywhere.
2:28 that looks scary. If I was that CPU I would be screaming in absolute terror. :P
Core 0 and 1 are running at 54°c idle is that a problem the gpu at 37°c idle My computer keeps restarting at random It's even worse on hot days I'm starting to see a pattern So i'm concerned it might be one of the two
Asus p6x58d-e motherboard with 16gb ram (2x 8gb) and i7 cpu 930 @ 2.80ghz 2806mhz 4 core and a amd radeon hd 5700 series
Any information would be helpful for finding out the cause of this problem
You should have touched on the location of the actual CPU die beneath the heat spreader. If the design is monolithic, you will get a single silicon die that's right in the middle beneath the heat spreader. Such a CPU will still properly cool even if the edges of the heatspreader are not being covered with paste. However chiplet designs like what AMD uses may have a I/O die in the middle but the actual CPU die in the top left or top right. In that case, if you don't cover the corners of the heatspreader, you may experience massive temperature increases. So instead of worrying about how much paste you use, first know how the CPU silicon is actually laid out beneath the heatspreader or just cover the entire thing with paste using a plastic card or something.
Very simple. The tolerance of the heat sink screw down yields the same pressure, tension and distance when fully seated. So, after squeeze out you end with the same amount needed as a proper amount needed to operate the CPU at its optimal level.
Most manufacturers of CPU's have a document that explains the correct amount of thermal paste to use. My experience has always been to put a spot a little bigger than a cooked rice grain. I've also, manually spread the thermal paste over the surface of the CPU and have never had a bad experience with that.
It's very interesting for me to compare method "spreaded rice grain" and medium dot. I am using a first one and it's works cool.
Great comparison. I would like to see the same test but with the spreading technique. Looking at the volumes you applied, I typically apply in between the "small dot" and "moderate dot" (Arctic Silver 5 also) but when spread you get even coverage. Side note, if anyone reads this - please don't apply the large/excess dot. I spent 1hour cleaning thermal paste from my PCB when I first bought my workstation and to this day I am thankful I decided to check the CPU paste before booting it up!
Same here, I put a small dot and spread the paste with a credit card to make a thin layer that cover evenly the surface of the CPU. The thermals are good and no cleaning needed after. I want to see the small dot spread evenly how performs.
@@sebastian19745 In my hands that has just always worked better. Placing a blind drop never covers the whole IHS. One recommendation, place a small drop in the middle once done with spreading. It is mean to help reduce the chances of an air pocket forming in the middle. I also spread both the CPU IHS and the cooler mating surface to ensure a durable finish.
@@racerrrz That is what I do. About the final small drop I will test soon; it is approaching the time to clean my laptop. When the CPU is not very accessible, I just clean it and spread thermal paste on the cooler part only. I`m kinda lazy...
But I never had thermal throttling or high temperatures in GPUs, laptops or desktops.
Thanks for the tip.
@@sebastian19745 No trouble, I am certain it will help extend the life of your laptop! It does take ages to spread both the IHS and Cooler surfaces but I think it is worth it for the long service interval. I run my machine for 10 hours a day, likely 6 days/week, and the Arctic Silver 5 lasts ~2 years! Not the best thermal paste out there but it does the job, and it has been slower to fatigue from the pump-out effect compared with other pastes I have used.
@@racerrrz I've run CPU's for up to 5 years before and never seen thermal paste pump out. I've even used the stock paste that comes on coolers just fine.
I realized that my liquid cooler is a crap, my CPU reaches 80°C on stress test, and it's one month old
Good vid man! Loved your test. I still think spreading it out is still the best option after applying.
Just submerge whole PC into silicone oil aquarium
Great video. Expertly explained with no wasted time.
I thought that more thermal paste would make it a little hotter but after seeing this I realized that this it is wrong because you are still tightening the CPU Cooler all the which makes the paste have about the same thickness with excess going onto the sides while I was thinking it would be more like a thermal pad.
URGENT:
What software do you use to measure the heat and the other stuff.
Thank you
Bro please tell the name of music you used from 2:43
Night Stalker by wave saver hope this helps 😊😊
I was getting thermal throttling on my CPU and decided to go through the pain of reapplying thermal paste and it was totally worth it. Dropped my temps massively and confirmed that I didn't put enough on there. The AM5 recommendation is actually to use a pea-sized glob in the middle and 4 tiny globs on the corners of the processor!
not to be negative but pls but music volume a bit lower otherwise good vid
Just putting a blob in the middle then seating the cooler just makes a circle. If you were spread the past to the corners it would increase contact surface area and dissipate the heat better.
Day 5 of asking for a thermal paste cooled pc
day 5 already see you on day 6
@@danf1025 sure
I've always used a plastic spreader to cover the entire contact point with a very thin layer of paste, then checked to be sure it has full contact to the cooler, never an issue..
Thanks! I recently installed a better CPU and found it getting up to 59 celcius. I had my suspicions that I maybe hadnt put enough paste on. I'll take it out, clean it up and give it another try.
When I installed it I was worried about putting too much paste on but after your video I think I wont be so nervous.
الطريقة الثانية هي الافضل والصحيحة ويفضل ان تضغط قليلا على المبرد لانتشار المعجون الحراري بتوازن ولا تشغل جهاز pc الا بعد نصف ساعة او ساعة على الاقل وستحصل على نتائج مذهلة
It's the space between the cooler and Cpu. Once you press down, however much can't fit squeezes out the sides and you end up with the same thin layer of paste. You just need to use enough to cover entire surface of the Cpu. The thickness is irrelevant
Let's talk about what thermal paste actually does.
Thermal paste serves as a thermal bridge between your CPU's heat spreader (the case protecting your delicate CPU), and the cooler on top of your CPU. Its job, to put it simply, is to eliminate all tiny air gaps, as air is a very poor conductor of heat. When ran with no paste, the heat will get trapped in these imperfections instead of transfer to your cooler.