Best Way to Apply Thermal Paste?
ฝัง
- เผยแพร่เมื่อ 3 พ.ค. 2024
- Comparing the different methods of applying thermal paste using the Intel Alder Lake i7 12700K. In the video I compare 7 different methods of applying thermal paste and examine how well they spread to cover the CPU and test the thermals of each method using an AIDA64 stability test to determine which is the best method to use.
#pcbuild #ThermalPaste #Benchmarks - วิทยาศาสตร์และเทคโนโลยี
Love all your videos man. Keep them coming whenever you feel the need. Love your tests, information, explanation, video editing. You have a channel that indeed is guaranteed worth to subscribe. Thanks Christopher. I'll go with the pea sized.
Commonsense approach, backed up with clear tests and no-hype results. Love your video mate.
Thank-you sir. I just watched approx 10 videos on this, but yours is by far the best. You are clear and include real world testing (full heat).
Thanks Christopher! I've been waiting years for someone to do this very comparison.
What works for me is the X with very thin paste and adding thin dots between the lines. This is really nice video.
Yes, I think this is about right. People spend a lot of time and get bogged down in pointless recrimination when there's really only two takeaways from these tests for me. You certainly want to use thermal paste, as it makes a significant difference in promoting conductivity. Yet repeated testing has yielded results that show that the difference between various thermal paste brands, their tiered product lines, and application methodologies is a scant few degrees. I used to painstakingly use the back of a plastic bag, taut around my finger, to apply paste evenly in the 1990s. Now some brands give you the handy plastic applicator. These days I usually use the pea method, but that has been with the square CPU dies. I'll probably use an applicator when I build with Intel again, but in any event, don't sweat it too much. The difference is negligible.
Thank you so much for testing this!!!
Awesome video as usual. Thermal pads are getting SUPER popular now days. Would love to see this!
Thank you for preparing and sharing this video. I found it clear and informative. Thank you for the patient, detailed explanations.
Why is nobody trying a two pea sized or two rice grain application, a third and two thirds of the way up the CPU
Fantastic analysis, thank you kindly.
Great video Mr. Flannigan!
Finally someone with some common sense. Great advice sir!
No matter how much I know it doesn't matter, it's just so satisfying to watch thermal paste videos. Lol
I spent quite a bit of time trying to figure out which method is best for me, over the years. The way I do it is with a very thin spread and tiny dot in the middle (much smaller than the pea-sized dot). With a thin spread there is a potential for air pockets, and I think the tiny dot helps with that. I agree that it's a small difference in temperature between methods. You're also right that if you use a lot of paste then the heatsink can get glued, I've had that happen. Thumbs up for great videos.
Make a build guide for the Velka cases, love your content dude
Awesome, thanks for the info
I bent the pins on my Ryzen cpu removing the cooler just as you mention, fortunately was able to straighten them and it works fine. Lesson learned, don't pull it straight out, give it a wiggle and slide it out gently to the side!
another thing would be to run your pc, let it heat up then remove the cooler..
This was super helpful
If you look at a infrared view of a CPU the hottest spot is a circle in the middle which gets cooler further to the edges
What if the X is smaller ?i think it won’t go over the CPU ?either way excellent video as always , don’t understand why your channel doesn’t have more subs
To me what counts is the max temps over anything else - it seems the pea drop and x provide best results, 72C, the reason you are getting higher temps with more, is that when the application is too thick, the transfer of heat is not optimal, you need a thin application but not too little just the rest amount for optimal transfer of heat, SO if you put too much, you are not optimal, spreading and other methods could create more air bubbles, the pea sized seems to provide a thinner spread, so more optimal heat transfer in an area that counts the most, center area.
something ive always done was just make a thing layer and spread it on the entire top of the cpu.
i think if you apply in two places with the "too little" method,
one a bit higher than the center and one lower, it would be the best in terms of good spread and not spilling out of the socket.
That is a good idea
The easiest way is to cut the PTM7950 out of the sheet and not play with any dots, dashes or crosses 😏
What about two pea size amounts at the 1/3 spots?
TLDR: Stop the buttered toast, just a pea sized amount is enough. Any other method may reduce your temps but is too insignificant to matter; but what is significant is more messy work.
I pay full price for my thermal paste. So, I'm going to use 100% of it. 😁
@@little_boy_blue I would suggest to instead keep the thermal paste for future reapplications. After a few years, thermal pastes tend to dry up and reduce in efficacy. The thermal paste you saved up can be reapplied when you feel your cpu runs hotter than usual.
@@konachan01 am very scared to apply thermal paste idk if my pc will be damaged which method should i use? some people say that using pea sized amount makes ur cpu temp high
@@revex4951 "Some people" is just opinions on the internet which could be subject to a lot of bias. This video is shows that there isnt any significant benefit to using more thermal paste than needed. Pea-sized is enough. If your CPU is still overheating, it is more likely the cooler needs to replaced rather than it is a thermal paste problem. Another video shows the same result, two different videos on this topic yielding same results is enough evidence to satisfy my confirmation: th-cam.com/video/ofyNgJyhGuc/w-d-xo.html
Subbing because you actually did a video on this!
It’s crazy how litttle you really need . I’m doing this today and I’m overthinking this so bad. I’ve always just got a new aio but this aio isn’t even a year old so I’m reusing it.
please make a video on the different lga sockets/cpu chipset from 7th gen to 14th gen and similar for ryzen on how to apply paste on them
thank you for talking about the ryzen stock cooler sticking to the cpu pulling it right out of socket. It happened to me man I tried so hard to get the damn thing off iso,heat floss, By the time i got it off just a few pins bent and with that my first build was dead... all because i wanted to replace that stupid box cooler.
wouldnt the best be 3 smalles dots you could apply then ? 1 in the middle, 1 on the top and 1 at the bottom (still not too close to the edge of and on the longer side)
Love the test, and the topic is one for the ages. BUT, the results from the 'spread' method MUST be effected by something, a variable unseen maybe? This method should provide the best results, not the worst?!?! Might need to re-run this test and try to identify the issue, something is going on in that test that messed it up...I look forward to you retesting in the future.
1:52 - Thermal paste covers the bottom more, because , looking at 1:49 the grain is more aligned with the bottom of the CPU and not centered. The Ryzen popping out of the socket when replacing the cooler/paste actually happened to a friend of mine. Poor lad had to buy a new CPU :(
Is it possible the spread method had an air pocket? I’ve seen people recommend spread + smallest dot in the middle to 100% prevent bubble at the hotspot
Thanks for the help, ended up doing pea method, but I have an amd cooler that came with my ryzen 7 and im pretty sure i smeared that shit everywhere because its a miserable fucker to install.
Appreciate the guide though, you deserve credit for your time so have my comment anr a like.
Being doing 5 dots method for years and it never failed me. If you being careful with the amount of thermal paste you put on the dots, you'll have an even spread with no excess spill on the sides. Spreading them with spatula supposed to be the best method but I just hate seeing the amount of thermal paste left on the spatula, just being wasted.
What about the + paste spread, i've noticed the + doing better on this 12/13/14 gen series chipset
i always do x with four dots. but with fairly thin coverage. you dont need a tone of thermal paste. you jsut want a light layer between the two metals.
Should be noted that the reason you might be getting some spill out of the sides from some paste configurations is because the chip isn't square. Spills seem to be happening on the sides perpendicular to the shorter width than the length.
no shit
Should include "for intel 12th gen platform". Because different kind of mounting or direct die without heat spreader have different outcomes. Without that in the title beginners will not get the full context.
This is 12gen Intel. Motherboard is a z690. Looks to be an Intel 12700k. Bracket even says lag 1700 🤙🏻. The problem is. Thermopaste needs more time to sit before pulling of the aio to see the real coverage
With 13th too?
It might be possible to have too much, but in general the only consequence is that it spills over the sides. A lot of people seem to think you can end up with a thicker layer by adding more, but that is not the case because applying the heatsink pushes it off the sides. I'm inclined to doubt you get air bubbles either. At least without some evidence.
It is quite unlikely to get into the socket itself though.
Thanks
How about 2 small peas or rice grains, one centred in the top half and one centred in the bottom half.
if by mistake we touch our thermal paste laden hand to the underside of intel cpu can we use 99 percent alcohol to clean it off ? mine got a little bit of paste underneath.......... will it cause short circuit or less conductivity ?
im going to try a very light and narrow x
The method I was curious about was 2 smaller peas in center about 1cm apart. Which may perhaps be better suited to the rectangular shape and spread more evenly?
What about spread, then dot on top?
I use the pea sized drop and then the tiniest drop that I can about 3/4 up the cpu, which takes care of that spot that was missed by the pea. Definitely because I have a 5 oz tube and I’m trying to use up 🤣
Two line and pea center ?
5:36 isnt it more about the socket-cpu connection type( PGA (pin grid array) or LGA (land grid array) ) rather than the brand ? Yes intel seems to favor LGA right now but that doesen't mean that they don't got any PGA s left or might even switch to them again ... What i'd really hope they would not do
maybe its just me but its very hard to see the gray paste on silver background. The lighting needs to be adjsted?
Thought: Would you consider 1 additional test? Since the 'spread' method gets 'best' contact, but higher temps, could you redo that with a tiny extra dollop in the center and report that? I suspect that the bending of the chip cover in the center (previously well-documented by Intel) requires a bit more to make sufficient thermal contact although it could also be related to the mounting pressure by the cooler mounts.
Do you expect a huge delta in doing this ?
@@ashleyjaytanna1953 Huge? No . But a statistically significant and consistent rise across tests? Possibly. I did that here on my 12th gen., but I don't have the instruments to do an adequate monitoring run. Worth a consistent 3*/4*/5* ? Negligible, but curious if it is so. Do you think they'll adjust either chip shell thickness or mount pressure going forward?
@@edwardsomers5037 There are dozens of tests on the Intel 12K and 13K. The difference in average temperature between using no paste at all, different brands of paste, and different methods of paste applications is irrelevant.
Noctua method is the most optimal one which is the 5 dots 2:12
Hey! I’m just wondering, where’s are you from? Sounds like a Irish accent to me.
X marks the spot
and its also depending on the thermal paste quality itself
I hope you get what you're looking for!
what about 3 small pea drops, vertically. like the Line method
I was thinking something similar but only 2 pea sized drops vertically not too far away from each other but not too close either, it should have pretty good coverage of the IHS and not be too messy, but I think I'll just go with the Cross X shape (but making sure they are not very thicc lines).
I got nearly perfect coverage by applying 2 small dots about 8mm apart on my i7-13700K.
Takeaway from all these tests done by countless youtubers: average temperature doesn't matter, 1 degree is not gonna make a difference, what matters is peak temperature when the CPU goes from idle to full load, cooler takes a while to spin up while there's a 50 watt burst of power going on.
Nice video. Another method is to use the finger to spread the thermal compound.
2 dots would be ideal
Was thinking the same thing.
2 dots on what spots?
I use 4 dots , each at the 1/3 of it own diagonal lines and the size of a grain of rice, or the exact dot that he tested into the towel. Once the sink is pressed on and screwed in, it will automatically float and spread almost perfectly
Hate messy paste… so I have been using a thermal pad for years.
X, 5 Dots, Spread
Those are the best 3 methods from what im seeing!
Arctic themselves that make the paste recommend to use pea method. Now we see why
А просто тонким слоем размазать. это уж не прокатит ?
You would think in this day and age, we all would be using thermal patches cut to the near perfect size and amount.
The issue with thermal patches/pads is that they have a (much) lower conductivity than thermal pastes due to their solid design and amount of air pockets. For this reason, applications that require high thermal conductivity never use pads but liquid or semi-liquid interfaces. This applies not only to semiconductors but also to electrical and mechanical applications. What you would expect in this day and age is that CPUs, CPU sockets and cooler baseplates would be designed to allow a consistent application of thermal paste.
Jesus gonna fix this. Rounded CPU. That's all. 😂
Sooo. 2 pea size dots on each side!
You should have done one without the paste for a baseline...
Thermal paste is an insulator, that’s why spread method has bad temps. Too much paste is as bad as too little (sometimes even worst).
The one and only purpose of the paste is to fill air microgaps (like scratches and metal deformations), not to get an even and pretty paste surface.
People have a hard time understanding that direct metal contact has better heat transmittance that introducing an extra layer of thermal compound.
That's not how you do 'spread' method. It supposed to be equally thin layer, not spread it all over. This method uses the least amount of thermal paste.
Way too much paste with the spread method. it does not hurt to have a bit too much, but definately not necessary. Max temp is NOT a good indicator of the effectiveness. You need to look at the hot spots (i.e. high temps/specific cores).
X is 100% the best method
I swear my brain is to small to understand anytNice tutorialng he’s doing.
This motherboard has been pasted too much
bad advice, could work only for old 1 chip cpu, but now with 2 or 3 chips under the case you can end up with some having no paste right on them..
You simply used too much in most cases. That makes your results useless.
This was super helpful