Thermal Paste Spread Under Glass 

Thanks for watching :) Please leave a LIKE, it really helps the channel! Best Way to Apply Thermal Paste: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-ofyNgJyhGuc.html

I usually just fill my whole bathtub with thermal paste and completely submerge my whole pc in it. It had yet to exceed -48° i did notice a 92FPS drop due to not being able to see my RGB through the thermal paste, so i made sure to cover the entire bathroom in RGB lights to eliminate FPS loss. 10/10 most effective method so far. Highly recommended.

"The bubble disappears, but where did it go?" These are the thoughts that haunt you after and make you want to remove it and take a peek.

X was found to reduce bubbles considerably and was quick and easy to paste on.

This video is great, to me it shows no matter the preference thermal paste will do its job, even the little spot that covers only 70% of the surface will do just fine.

Last year during lockdown I took a weekend to do a bunch of testing with different spread methods. The pea, rice, X and buttered toast methods. After each application I ran a one hour benchmark to compare benchmark results and temps. I did this three times as I had so much time haha. The results were all within

Ah yes. A 10-dollar tube of goop that is the difference between a freshly assembled 2500 dollar computer, and a heat-treated brick.

It's important to ensure you apply the force in the center, to compensate for flexure of the acrylic.

Incredibly helpful/interesting. The effect on temps is debatable, but it's still nice to have a better idea how well the paste is applied, simply a matter of OCD more than actual results. Very cool idea and very well illustrated to us! Thank you!

Keep doing these sorts of reviews about PC internals and components and results in real-world. You don't need to try and compete with the sites that are test benching everything to within a micro-millisecond. Just answer the average users' queries about the best way to do things and whether it makes any difference which way it's installed. You've found your niche ! Thanks.

try doing it while the acrylic sheet is covered so you don't adjust the pressure as you press based on what you see. then take off the cover to see which method is more prone to making bubbles

I used to apply the x method, but Noctua actually recomments the penta pea method as it saves on the paste but does perfect coverage as outlined in your video. Noctua has a couple of videos on how to apply their paste based on the type and size of CPU.

I think mounting a cooler puts a lot more pressure than your tests because I've always used the pea method and the coverage was 100% when I pulled the cooler off.

At the start, I thought you were going to push the acrylic against the heatsink rather than the CPU. So that way you can see how the grooves and grain might affect the way it spreads out.

The key observation of this experiment should not be the amount of coverage or the pattern, but how easily Thermal Compounds are designed to spread under pressure. You do not need very much compound to achieve an effective amount of coverage. A common error users make is to use more than necessary and overestimate the amount required. You don't actually need as much as many users typically think and they end up with excess that might actually impede thermal conductivity. The job of thermal compound is not to transfer heat, per se, but to displace air by filling in the microscopic imperfections in the surfaces of the CPU heat spreader and cold plates of the cooler assembly. Heat transfers best with direct surface to surface contact and air is a good insulator. The less air trapped between the surfaces the better; and the more direct contact between surfaces that exists; the more heat energy can be transferred through the spreader into the heatsink and away from the CPU core. Thus total coverage of the heat spreader with thermal compound is not critical, so long as a good surface to surface contact directly above the CPU cores is achieved. Caking thermal compound to the point of it oozing out the sides of the assembly, not only wastes thermal compound, but may actually create a barrier that degrades the quality of the direct surface to surface contact and lessen the efficiency of the heat assembly as the thermal compound hardens naturally into a wax like consistency over time.

Having worked in the electronics industry and in other industries... application of the HSC to BOTH faces in a thin even layer then apply both together with the added 'wiggle' to seat them well.

For years I've done an X with small dot in each open space between the lines

Might help to start by saying what the objective should be - which is to cover the whole of the CPU with as thin a layer as possible. The paste is only there to fill the gaps between imperfectly flat surfaces and, wherever possible, direct metal to metal contact is much better. if the surfaces are not perfectly flat there would otherwise be air gaps and air doesn't conduct as well as thermal paste but thermal paste doesn't conduct as well as direct metal to metal.

The reason that the air bubbles look like they disappear is because the glass is getting covered with thermal paste. Shakey hand movements cause it to reach more and more surface, until eventually you just see gray. The bubbles are exactly where they used to be, just right underneath an ultra thin layer of gray goop.

Great idea using the transparent plexiglass for showing the thermal paste spread. Thank you

THIS IS THE BEST VIDEO, shows you exactly what happens with the paste AWESOME

I think a lot of the misconstrued mythologizing about this is that every application is never the same, like the butterfly effect, In my opinion I think for an inexperienced person the X marks the spot to play it safe.

Thank you so much! I was very scared of doing anything myself inside my laptop but it was overheating and wheezing recently, it just needed a good clean up and a bit of new paste. It took me forever (my sinks were attached with two big fans, it was a struggle) but finallyI managed to clean everything up, add new paste and now laptop is running quieter and doesn't get so freakishly hot. THANK YOU!

Having never had the inclination to do this - Kudos to you!

Great video. I would suggest you cover the glass with tape when you put it on and take the tape off, once you think you are done. This way you prevent yourself from cheating due to the visual feedback. Some methods might work better when you don't see how the paste spreads.

I do the 5 dots, center dot being a bit bigger. Then I don't clip the heatsink yet, I twist by like 10 degrees or so, back and forth a few times, the heatsink to make sure I spread and squeeze as much out from under there as possible.

This is a really cool visual of what approximately happens. Great vid!

I've seen several videos on thermal paste application (some from major tech tubers) but I don't remember any of them doing something a simple as using a bit Perspex so you can actually see how the paste spreads. Great Vid 👍

I've been doing the x method for several years. Seems to work fine. As long as you don't put too little or too much, you're fine.

I wonder if in these tests if your acrylic is flexing and adding more to the bubbles, and reducing the spread from the middle of the cpu to the edges. In either case this gives a decent representation.

This was great to see i have always spread my thermal paste i have done it that way for 20 or more years but the x pattern and many others all work fine .

The video we've needed since forever

I used the pea method, based on Noctua's cooler manual suggestion. Mine was about half the size of your pea, and seems to have done the trick. Right now my pc is at 29 C. Never goes beyond about 72.

As long as you use red thermal paste for AMD, green thermal paste for nVidia and blue thermal paste for Intel, you will get the right RGB proportion.

Thanks a lot. Will be using the X. :) Covers. No Bubbles. No spilling over the edges.

very interesting! i usually go for the X with one dot in the middle of each V shape

Been using the cross method for about 8 years on Intel and amd, Always Works the best

The reason tile mortar is applied line and groove is so that air gets squeezed out as the tile sets, it's best to not do shapes that fence in air (circle, infinity...)

Very good test of the methods. And also very professional way doing that. (The sound quality also adds to the experience)

What a great idea using glass

This was soo enjoyable to watch kinda like the feeling you get peeling off peel. So incluconsion, about the only thing you don't want to do with thermal paste is use too little.

Through my own experimentation I've found my favorite is a super thin, but very even buttered toast. Then a small dollop (of jam) in the middle. I also try to wiggle the mount a bit to help even out any differences from an off center placement. I love the MX5 that Arctic recently released.

This was enjoyable. I used to do the spread method, however the spread with the dot seems like a great method.

Very nice to see this under glass!

Amazing video! Thanks.

Pea dot method has never failed me, as long as theres enough paste. Any air escapes out the sides as pressure is applied.

Thank you. This is EXACTLY what I was looking for.

There really isn't a wrong or "better" way to apply thermal paste. It's just grown into more of an internet meme that people take too seriously. Also, it is a myth that you must cover the entire IHS. CPUs have hot spots. Those are the areas that need the thermal paste. Kudos on the ingenious way to test this though.

I like to rotate the cooler after I put on the paste, maybe 20-30 degrees each way a few times, and then rotate it in a little "wax on/wax off" fashion. My thinking is that I want to try to get some of the bare metal to touch in a few places, and let the paste fill in any dips or humps. As opposed to having a layer of paste everywhere when it's not really needed.

The "bubbles" worry usually is moot because a lot of coolers are putting way more pressure down on the CPU than your hands will. The air bubbles get squeezed out with all the excess paste generally.

I have yet to build my own PC (family / time/money/etc)... but it's on my bucket list. This was fascinating. Love that it's practical, and not basically theoretical [ like: this (x) is .01 millisecond better than (y) ]. Also - The outtakes at the end. Subscribed. 👍

Nice, thats what is the internet needed, another of thousands videos about thermal paste spreading. Does now every hardware channel make a "new" test how thermal paste looks like under glass?

I do the X method with little dots in each open space :)

Nice comparison :) It would have been nice to make sure that the pressure / pushing force is right (right in respect to the pressure which the colling element, mostly a cooling fan, will apply). As long as you increase the pressure it will lead to an ongoing spread of the thermal paste and sometimes you could missimulate the effect a bit by applying too much pressure just to get the paste to each corner (here again, too much in respect to the cooling element which would have a set pressure if you mount it in the right way).

Keep feathering it, brother.

Really glad to see this video. I work in the professional computer building industry. This is always discussed amongst the techs & engineers. I have always done the big X, corner-to-corner. Then a smaller x to make it like a snowflake, those smaller ends extending from the center, outward to the sides. I always start each line from the center, so a bit extra is left there each time. Giving me that extra dollop in the center.

Great idea to do tests with a clear material so you can see the results! Value of a steady hand? When you are putting on a big aircooler how steady can you be with it? Especially if you are changing a cooler while everything is in the case and you are dropping the cooler down into the big hole. So this leads to another question- how do these various methods work if you put the cooler down so that it contacts on one side of the cpu then rocks into position towards the other side?

I normally go with 5 sided dice, but I like the X. Regardless I rarely apply paste or overclock.

Why the hell hasn't someone else done this video before? You are the real MVP. Subbing just for this.

This video convinced me to stop doing the "buttered toast" as it has more risk of trapping the air bubbles. I still like a good coverage in the corners, so the X and the 5 dot seem to be the best overall.

I've been using Ye Olde debit card method since ~2002 when i started working in a PC shop where we assembled and repaired systems. I can't ever recall a customer coming back in with any CPU overheating problems, and we never discovered any, and we used to sit and play games on some customers PCs as they were higher powered than anything me and my coworkers could ever afford, so they were well tested! LOL As with all things, we may have missed a handful out of the 1000's of systems we handled. On my own personal assemblies, i of course stuck with using the debit card method, and again, never had any overheating issues at all. We had done a similar test with a glass plate that you'd use with those microscopes, thinking that the glass would be of a higher surface polish standard (not that it mattered) and used a cylindrical weight from an old set of scales that almost perfectly touched all four sizdes of the CPU (again, not that it mattered). We would start this test just before we closed the shop for the night at around 5.30, so it would be an over night test that was checked when we opened again at 9.30am, and besides the X, the rest of the methods you've shown here turned out at about the same coverage. The "Penta dot" as you called it, was simply done like the 5 on a dice face, with the dots closer grouped together than shown in your own testing. My coworker even just Picasso'd (random) thermal paste onto the CPU, and it turned out about the same, but with a more spilling over the side than even a large blob LOL (don't do either since it's just wasteful). In the end, we decided among ourselves to eyeball a pea-sized blob and then use a debit card to spread it out. This was our way of ensuring we didn't use too much paste, which in turn would translate to higher costs over all from having to buy extra thermal paste. However, when you're doing your own PC, then you don't need to worry about over all costs like that since you only need one tube. So even though i have known for a very long time, that it really doesn't matter all that much, i still can't kick that habit of using the pea sized blob + debit card method. It's like an addiction at this point in my life! LOL

hey little suggestion if its not too late: if you wanna see how the grooves and channels on coolers affect the spread, use the acrylic on said heatsinks as in apply some thermal paste to the heatsink as if it was the CPU, then press the acrylic on it

Thanks! I am glad someone made this cause I’m overthinking some stuff and this helped me have a good sleep 😁

The X marks the spot!

You should try what tilers do. They use notched tile trowels to avoid bubble. You could probably make one suitable for spreading thermal compound by just cutting notches into an old credit card or something.

Really good test! Thanks for this

It's an epic theme that will continue for generations. 🤣 I in particular used to distribute an even layer with a spoon of ice cream or similar.

I've just always done the penta dot method for years and haven't had any noticeable issues.

Excellent showcase. How flexible was the clear cover plate? Could pressing that far from the cpu cause it to dome up in the center? Maybe you didn't have to press very hard to get the spread.

Great way to test. Doing the X right now!

That was very informative 🙂Thanks for the science!

been using X method with 4 small dots for years

Your videos are terrific. Very useful stuff.

Very informative thanks for this! Building a computer at christmas so Im glad this came up

Good test bro. I use the pea method and it has worked well for years. I use a bigger pea size for AMD chips.

I've always used what everyone calls a "small ammount", and when I took the cooler off(weeks later, sometimes months) it's spread across the entire IHS, edge to edge, corner to corner, in a layer so thin that you can almost read the writing on the CPU. _That's_ what I call "perfect ammount". With my 2200g with Wraith Stealth or with my brother's 2700X with a large TUF cooler (can't remember the exact name), I applied the paste on both of them myself, and monitoring the temps they never went over 65°C in stress tests, and lower in normal gaming regime. And for the record, I have tested them with the "recommended ammount" and there's little to no difference, so why bother using more when less is more? More paste is bad too, there needs to be just enough to fill the gap, any more and it's an insulator. But noo, if there's not thermal paste seeping in my RAM slots then I have too little.

This was fascinating. Great job.

Nice Channel! Nice Topic! Just subscribed! Keep it ON!

It's all fine looking at the spread while pressed with glass, but what many forget is that cpus/gpus get hot, heats up the paste and the paste then spreads even further acorss the thermal surface. You really only need a thin layer of paste, as the paste is Only used because of the microscopic pits and other impurities of the metals used. When it comes to thermal paste, less is better. Ive been a tech & applying paste now for over 35yrs.

I tend to use X method, but I add really small dots in the open parts of the X. 🙂

I like the pea in the middle method the best

I clicked this video in less than a second while making a sort of honking sound. Excellent stuff.

The X or 5 dot method is the best, and yes it does make a difference more so the paste you use as long as you have a large enough coverage, it won't be perfect every time and the heatsink (air cooler or AIO) will apply a better pressure across the IHS when you screw it down so that helps spread it better.

@Tech Illiterate This was excellent! It would be interesting to do this same type of test with Glass instead of plexi, mostly because glass doesn't flex like plexi does. Might better show what actually happens to the air pockets..

That's a fantastic way of seeing what's happening 😀

great video, thanks for making it! X method is indeed the best!

It probably depends more on the volume applied rather than the shape

bro how hard are you pressing down???

Gunna start using the X ! Thanks mate

I use plastic wrap on the finger and spread it evenly across the IHS and never had any issues since. In the beginning I tried rice or pea method but ended up having to re mount the cooler until I got an OK result.

As said. It doesn't matter. Gamers Nexus looked at all methods and quantities. The diference in temp was miniscule.

Nice one. Great channel, great quality content, love it. It would be interesting to not apply the pressure completely by hand, but rather have the transparent sheet rigged with holes so that one could install it using the cooler mount screws. This would get the results very close to real life. Not to mention that these tests vary alot depending on the thermal compound used, with their different viscosities. Last but not least, der8auer, after also making these kind of glass sheet tests, said on a GamersNexus video that due to the very high mount pressure between the heatsink and the CPU IHS, trapping air bubbles of a size that actually matters is almost impossible. Here: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-CCqxE-5Ct3w.html I would personally add that in a DIY home situation like this, it is actually impossible to NOT trap microscopical bubbles in the compound, which even out, and they do not really matter in the end thermal performance and measurements. AND they will eventually escape thanks to the pump out effect through thermal cycling.

Thanks for posting this video. My PC has been shutting-down with greater frequency. I started monitoring the CPU temps and saw that they were averaging in the high 60's C, spiking into the low 80's. Time to re-apply thermal paste! (And thoroughly clean everything else) I followed your advice of using the X method. CPU temp is currently averaging high 20's.

Nice & Very interesting ! Thumbs up ! Thanks ..

I think spread patterns mattered a lot more back in the day when thermal paste was more clay-like in consistancy. With the more fluid pastes we have today, even if you have a bubble it will work itself out to the edge over time as things heat and cool. I think having too much was also a bigger deal with older paste, because the thermal conductivity of the paste wasn't good... it was just better than an air gap. Modern pastes are almost as good as a metal connection, so having a blanket of paste that might be a little too thick isn't a big deal. I do think it is important to have some idea of the layout of the chip underneath the lid, not so much for the paste, but for the cooler. I use an Intel chip, and an air cooler with several heat pipes. If mounted in the traditional direction where the heat pipes are laying up-and-down parallel to the RAM, and the fans are blowing front-to-back, then only the center 3 pipes are sitting over the CPU because Intel chips tend to be long and skinny. Rotate it 90* so the fans are blowing bottom-to-top and then all 5 heat pipes are going across active elements of the CPU which drops system temps on my CPU by about 10*c which is pretty sweet. It does make RAM install a pain, and looks a little weird with the fans facing 'the wrong way', but it is very effective. Last winter I picked up a 3090 and started mining, which caused all sorts of thermal fun because the RAM modules on the back of the GPU would hit 112*c very easily (and then the card would thottle). So I tool an old Pentium 2 heat sink, filed off the alignment nubs, and pasted it to the back of the GPU. Then I changed the CPU fans to blow top-to-bottom so there was active airflow blowing on the back of the GPU. This has knocked temps down to ~94-96*c which is more than acceptable as the modules shoudl operate just fine under 110*c, and should only take on damage near the 115-120*c range. Fast forward 5 months and everything is still going strong! Sadly with the coin price crash my GPU has only paid for 1/4 of itself instead of half of itself like I was hoping for by now... but prices will come back... eventually. And it looks more and more like the merge will have a delay, so we will just keep on going until we cant, and wait for a rice spike to pay for the bulk of the fancy GPU. It has still been a fun experiment, and having a GPU actually capable of playing 4k games has been a lot of fun!

My thoughts are to use as little paste as possible. The ideal condition is to have direct CPU to cooler contact and let the paste fill in the gaps. I lap my CPU if it needs leveling, apply a pea sized bit of paste and when I gently press the cooler in place I also slide it back and forth slightly until I can feel the metal to metal contact. Based on dis-assembly for my last few upgrades, most of the paste is squished out and I get perfect coverage. Never lost a CPU to heat and I've been building PCs since 1983.

Penta dot and X looked really nice

I'm always gonna be an X and dots guy, never failed me yet.

I also use the buttered toast method but instead with a dollup of marmalade

And to think I was just doing the X method this whole time just to make pirate puns to myself when I was building PCs.