Do You Even Need Thermal Paste?
ฝัง
- เผยแพร่เมื่อ 22 พ.ค. 2024
- Since my video on thermal paste application method I have been frequently getting suggestions to try it without any thermal compound at all. So, here it is. Do you even need thermal paste?
Arctic MX-4 Thermal Paste: amzn.to/2SHztN2
Kraken X53 Cooler: amzn.to/2RHyFXU
Support us on Patreon :)
/ techilliterate
Or you can hit the easy buttons, just LIKE and SUBSCRIBE. It really helps a lot!
0:00 Application method doesn't matter
0:33 Arctic MX-4 Test
1:37 Removing thermal paste
2:45 No thermal paste test
4:51 Some asterisks
5:41 Is thermal paste needed?
Music - Downfall by Chris Doerksen, Drop off by Chris Doerkson
/ chris-doerksen-1
Circus Of Dread: Robert Austin
Download: goo.gl/6Ivaxs - วิทยาศาสตร์และเทคโนโลยี
![MOBYE - มีใจทำไมไม่จีบ (ASAP) [Official MV]](http://i.ytimg.com/vi/dCOdz92FU18/mqdefault.jpg)
Thanks for watching! If you need paste and want to support the channel get my favourite Noctua paste here, US: amzn.to/3QJdrDF Canada: amzn.to/3CRGcrY
awesome test, thanks!
"What do you expect to happen?" IDK what I expect to happen, but the first image that popped into my mind was my black mage in Final Fantasy 14 casting Fire IV, the most potent fire spell.
Yep
This shows that the cold plate on your AIO has actually pretty good evenness. The temperature without thermal paste is really dependent on the unevenness of the plate, relative to the cpu ihs. If it was more uneven you would see higher temps or even dangerous temps, depending
Even then, 9 degrees delta at peak utilization is a sizable difference, proving that thermal paste always helps.
But I was shocked that idle was no more than 2 degrees delta and light gaming was only 4 degrees delta. Which proves your point: the more even the IHS' surface and the cooler's contact surface are, the less reliant they are on a thermal interface layer between them.
But since you cannot gauge the evenness of your CPU's IHS and your cooler's contact surface without specialized high-precision tools (and exceptional evenness tends to be the exception and not the norm), having thermal paste is still preferable.
That's a very good point. I have a Core i7-6850K cooled by an EK Supremacy Evo waterblock and it would instantly hit 100 degrees and thermal throttle under load at stock speed with no thermal paste. But after I lapped the CPU and waterblock totally flat, temps maxed out at around 75 degrees with no thermal paste. Obviously not nearly as good as under 60 degrees with thermal paste, but still totally safe and completely usable once lapped.
a metal plate has good evenness?? mind=blown
Your thermal paste comparison videos are great! Thanks for your testing and sharing results.
New subscriber here, I really like your approach to these tests and videos. Absolutely no discrimination and 100% honest.
About to binge on all of this channel.
I too am a new subscriber. Great content. Out of habit I always lap my coolers with 1500 grit sand paper and Isopropyl Alcohol 91% or better. I use very light pressure and always make sure the sandpaper is wet. When I'm done...the bottom of the cooler feels glass smooth. Not sure if that adds any lower temps, but I do it anyway. Take care all. Cheers
This channel should have A LOT more subs. Oh, and thanks for "risking" your CPU for us. I expected a much greater temp difference.
Yes. I started up a i5 8400 with no thermal paste and a stock cooler I went into bios saw 91 C and shut it down,
95c
Rookie numbers, mine was 101 C. Get on my level😎
@@mrjazz2570 mine was 121c get on my level
Because stock cooler. This is the cause.
Mine shutdown in 10 seconds
Just subscribed. This was an extremely well made video, far above some channels with 1M+ subs. You've clearly got the formula down with these videos!
I'm honestly surprised it mattered as much as it did. The paste helps pretty microscopic separations connect. It's cool seeing quantitatively how impactful they are. It's definitely necessary but also not something that has to be applied in a lab setting with any level of small tolerance or specific amounts or patterns.
I got a theory that people are mounting the coolers way too loose, hence the wide spreed use of huge amounts of thermal paste, I use so very little and spreed it all over the cpu surface without spilling out.
@@vfn4i83 actually you should do it enough to spread through the ihs. This test just means nzxt kraken has spectacularly even mounting surface and pressure. Thermal paste is necessery when these things are not perfect which in almost all cases they're not. You tightened the screw on one side half a turntoo much? half of ihs is wasted. Did your AiO manufacturer cheaped out on manifacturing cold plate, same story excessive heat. As air is actually good insulator which is why paste is necessery, it removes alot of the variables which almost always happen. Hell, just look at latest intel socket, even the motherboard's cpu mounts are so bad that when you switch that to a TR or TG you gain around 5 degrees of cpu temp just from that. And that is WITH thermal paste.
My cpu is at 60⁰C idle without thermal paste. I think I need it
Thank you so much for your tests and the application dot pattern tests!
Nice video. Surprising results. Thanks for exploring this.
Thanks for this great video! I like the theme, cutting and presentation very much! Keep going! Thank you ✌️
Great video, I love your testing, I cannot believe this channel is not above 500K subs. Keep up the good work man, these are very high production videos, the quality is amazing, from your lighting, to your photography skills to your audio recording, I appreciate the overall image of this channel you have so much good things going on here, no bullshit sponsors, no paid ads, and even if you did have those I wouldn’t be offended by that. A channel like this does not come around a lot. I loved the video where you tested the paste patterns and this one. I watched them both without looking at your sub count and after finishing this one I was shocked to see that your sub count is so low and your views are too. You just gained a long term subscriber and I will be sharing these videos with all of my associates in the tech field. Thanks for the great content brother! Stay frosty. (One thing I will add is that your thumbnails don’t fit exactly to your video style and your channel. They look well made I just don’t think they flow with the channel. It’s a format I’m not used to seeing as a viewer and maybe it’ll grow on me, I would suggest experimenting with new styles but what do I know, I’m a dying 1K subs old washed up channel lol. Your thumbnails are way better than anything I could create myself.)
Thanks for the feedback. I appreciate it.
I struggle with thumbnails. They are the most dreaded part. I usually put my hands up and say "I guess that?"
@@TechIlliterate haha. Nice tho. You’ll go real far I know.
Thank you!! God, I've wanted to see this (and the application video too!) done, to see just how delicate this shit is. EXCELLENT!
Your videos are really good to watch.
Wont take long till you get millions of subs.
Good Job.
*J'adore ces expériences, je trouve qu'elles font avancer !*
Ça rassure de savoir que mettre un peu ou beaucoup de pâte thermique ne change pas grand chose ...
Merci pour cette démonstration .
Merci à vous!
Really good testing presentation
Very nice demonstration of Thermal Paste !!!
OK, my prediction is 10-15 degrees higher temps. I'll continue watching the video now. :)
Wait are you the real kostas!????!1!1?!11
Lol my air cooler gets better temps at full load than many liquid coolers, it seems :) Awesome.
@@DanZhukovin it is well known that good air coolers match mid size AIOs, 280s or 360s are needed to get any real gain in actual peak fully soaked temps but liquid coolers obv do better in burst loads because of the large amount of mass to warm up.
Awesome!! Would not have guessed these results...
Thanks for doing the tests
Actually,about what i expected. Great Video!
You make really good videos! Hope you get more viewers and subscribers to reflect the quality of your content!
Another great and informational video,,,,,, well explained and detailed!
Thanks again!
Not what I expected! I thought it would run much hotter, and throttle heavily. Thanks for doing this test!
Mine died
Damn!! So this where that little voice comes from all the time... Thanks!
You’re underrated, just earned a sub my friend
Really good content!
Love your videos :)
Thanks for this video, took a lot of bravery to run without paste!!!
I am definitely not an expert but I think it mostly depends on what cooler you have, if you have a super cooler than maybe you don’t need thermal paste, but if you have something like a stock cooler than I think adding thermal paste will have a huge benefit
Interesting results, I honestly thought with an AIO, a good one with cool temps on the cold plate anyway, would perform alright without paste.
My takeaway from this though... is that if you have higher temps than with no thermal paste at all... re-applying the paste might not be a bad idea.
Really good video brother
Good video get good comment. Was on the border on whether I should buy some or not. This was a great video too
amazing results
good editing, good video.
It would have been good if you set the CPU fans to a specific number, that increase of 200rpm slightly altered the results.
I went to ITT and we used thermal paste here and there and in my working with hardware through years, you would find it in the oddest places from voltage regulators on old weird places, but it was always said "it just fills the small gaps, it most cases it's not even needed". But you learn that paste gives you the edge to ensure those tiny gaps are filled to ensure contact is made. This video clearly proves that old saying :)
Subscribed! Thanks for the video
Sometimes I reuse my thermal paste and add just a little if remove the heatsink.
Always use "smooth like a butter" to apply the thermal paste.
I think the real difference is the quality of the radiator when dry mount. Else, I also think that applying the paste on it (and spreading) instead of doing that on the CPU, will give slightly better result, as the CPU cap is quite smooth, while the cooler is rough and using too thin layer on the CPU will not fill up all the gaps when pressed
The CPU cap is quite rough, if was so smooth, it'd be glassy and the copper surface would also be glassy, that would be the only way there'd be little to no gaps. If there's even a smidget of a micron gap then thermal paste would continue to win because it fills in gaps. Simply touching the surface of the cpu would cause gapping problems
Thanks for this! I'm doing my first build. Just picked up a Ryzen 5 5600X. But I forgot to get the thermal paste. So I just bought some from Amazon.
really cool video, thank you for doing it on your hardware so I didn't have to! :D
Searched "no thermal paste" just out of curiosity and came across your awesome channel. I always thought the thermal paste debate was a little silly. Just dab some goop on there and move on with life. Anyways thanks for content!
Goop costs money.
Made me feel better about not overthinking how much paste I need exactly on my cpu.
Agreed. I have to put a new AIO onto my CPU, and I was concerned with the best method for applying my paste. I'm still gonna do small-P/spread though, just for my own peace of mind.
Thanks for these videos about thermal paste, good to at last have a go to point for pointing out some proven proof, of ideas most of us has just assumed!
Quick and simple answer, yes absolutely
Great video
Interesting results! I had a PS3 that sat for a few years that wouldn't turn on beyond flashing a heat warning. The fix was fresh thermal paste.
just stumbled upon your chanel, and we suprisingly have a very similar build!
Somewhat warmer, but not like a 1000 degrees C. Because the cooler matters, not the paste, the paste helps to even out the two surfaces. Sometimes they touches.
If you had a perfectly flat heat spreader and a perfectly flat heat sink mount, you could probably get away with no thermal paste. However that is not the case. Never run a system with no thermal paste.
Yeah, after seeing just how not flat the IHS and heatsink are I would never recommend running a system without the paste. Now, if you have lapped them both and there is perfect 100% contact, maybe. But even then, why not just slap on some paste and enjoy your build.
Great content, your channel should blow up soon!
Thx for the info! I'll be adding it to the 13900k :)
I like that epic outro 😆😆
I expect higher temps due to air gaps. No plate is perfectly flat. I think I saw no-paste benchmarks before and it's often a "did not finish the test" type of result... Then again I don't remember for sure... (We all saw that the small dot from before was pretty warm...)
Edit after watching: Wow. Goes to show beefy water coolers can really do a heck of a job. Even with less than ideal conditions.
Loved it
All this video proved was that you have very nice and even mounting pressure :)
Not surprised at your findings at all, Not used thermal paste in any of my past PC's for 25+ years, never had any overheating issues either.
You'll never have overheating since all chips since 20 years ago are thermally throttled. The issue is performance: the higher the temperatures, the worse the CPU performance.
Thanks for the research. I have a non cpu application that thermal paste may not provide a benefit, but thermal epoxy is still an option. These Thermal Interface Materials, TIM, seem to work less by being a thermal conductor, than by eliminating a non-thermal conductor - air. My research will be to attach a Peltier to heat an aluminum plate in a vacuum chamber. With no air insulation gap between the ceramic face of the Peltier and the aluminum plate, heat conduction should be better than at normal atmosphere. There may still be a need for epoxy, but this demonstration is insightful.
As a temporary solution until your thermal paste arrives, none is great
the surface of the CPU and the Cooler might be without gaps and don't need a filler. Other CPUs and their Coolers might have bigger gaps.
Also i wonder if the thermal paste keeps CPU and Cooler a bit separated, so that they don't fuse with each other over time 🤔
I expect your CPU to overheat and your system to throttle and maybe freeze?
Edit: Not even close! Oh well.
The b&w part was great
Thx for this video. I have tested this years ago, and its no suprise to me. Its almost irrelevant which thermal paste you use. You can buy whatever you want, the only difference is the money you spend.
Excellent.
As others have mentioned I think cooler quality matters. I run Pentuim III (both Coppermine and Tualatin variants), Xeon E5450, Xeon E56xx and X56xx series as well as an I7 970 all without paste, and everything works just fine. There are 22 CPUs in total. All the machines are either name brand servers (HP, Compaq, IBM, Dell) with quality parts and good airflow, except the I7 which has a good aftermarket air cooler. After about 8 years in a non-airconditioned room with Australian summer ambient temperature, nothing has shown any sign of heat-related problems.
Would have liked to see this test with a normal stock air cooler.
How do you not have more subs!
this moreso proves how awesome ur aio is
Was the fan speed the same in both of these tests? If it is temperature regulated (which I'm sure it is), you might have it at bad comparison since fan was spinning faster when you tested with no thermal paste.
I'd like to see this same test with an air cooler. Liquid is much better at removing heat so I'm curious if the results would be similar with an air cooler.
Basic science. Thermal paste creates better conduction between two flat surfaces because under a microscope no surface is actually flat. Hence conductive fluid to fill it. The fact people even have to ask this shows where the works is coming to. It's very easy to use Google.
i have dozens of times had to set up a quick thest bench pc to run some hdd recovery software i would just find a mobo, a cpu and a few sticks of ram a psu and sata cables and just a cpu cooler laying around, i would just throw the cooler on the cpu and not use thermal greese or strap it down i just just place it and launch the program i needed and let it work, works fine you just need a chunk of metal to be honest for a cooler
Your videos are great. Well structured and informative. I really like the gag-insterts.
Thank you very much! I just try to make myself laugh at dumb jokes. If someone else enjoys than it's a bonus!
How about test the stock AMD cooler you used in the paste distribution test? Those channels between the heat tubes are sure to cause issues when there's no paste on the CPU.
I'd love to see you do the same testing with a mid-level air cooler. That water cooler is very capable of absorbing so much more heat that yes, no-paste differences are minimal, but due to the nature of water cooling. Could you try it with a $30.00 air cooler? I'll bet there is quite a big difference, but I'll let you try it! I'm on a budget!
NGL I loved that little skit in black in white
I always use thermal paste when building of course. But I did get some celerons and pentiums from different generations for testing boards and I don’t put anything at all on top. I used to, but one day I decided to pull the heat sync off with hwmonitor open and it only went up 5c, so now I just save the time and skip it
Next: what happens if you try to reuse thermal paste: for instance, you put thermal paste, bench your pc and find out that the heat sink isn't making propper contact. So you remove the heatsink, you confirm that it wasn't making proper contact, and then.... everybody says you need to reaply the thermal paste, but maybe just spread it again will do the trick?
I can confirm to you by experience that it changes nothing. I've reused thermal paste several times and nothing has changed. The only difference I found was when the thermal paste was already solid.
Did that just yesterday. I didn't even bother spreading it again, just took the heatsink completly off and back on again. Changed absolutly nothing.
My guess would be that it depends on how old the thermal paste is, as my laptop suffered doing that but the paste was over 5 years old 🤷♂️
@@spagettifelge7191 damn, wish i knew this earlier
It depends on the paste. Some leftover pastes are still reusable, but most hardcore gamers’ thermal pastes’ has some flaky parts, basically some pastes dried out and feels a bit like plasticy instead of a paste.
It's interesting, my kraken z63 sprung a leak in the radiator and the effects were immediate, my cpu usage shot to 100% and overheated within a min, it even had thermal paste. I couldn't even boot into windows before it shut down again. Once I put my air cooler back on it was fine. Yet, here you are with no paste and getting stable temps with normal usage. You should have used an air cooler also, cause maybe my failure was due to no liquid being circulated.
The thing about thermal paste is, once it's getting too old (especially crappy thermal paste) it's going to have a negative effect because it might crack, leaving areas with zero contact.
On my old laptop (about 8 years old) the temps were always quite fine, but one day it suddenly started to get really bad. Turns out the old thermal paste was dry as hell and actually cracked in the middle (leaving an air pocket of a few mm there).
This is a bit weird, in my case, last week I applied my thermal paste 3 times and see a higher difference on temps. I re-applied it because I though it's wasn't well applied before. Using an Bifrost 240mm Water Cooler, made by Riotoro (believe me, it's a very, very good WC). The thermal paste I'm using is the Noctua NT-H2. Now I'm getting temps around 50º and 62º C on full load with my i5 9600k at 5.0GHz and 1.29 on vcore. The difference I have seen in temps before was something around 2~3º C.
that aio is plenty powerful enough to handle a 150w load, that measly 65 watts isn't going to suffer to much with out any thermal compound, so long as you have a solid connection, apply enough pressure and the surfaces mate well, the thermal compound helps transfer the heat in the areas that are not mating that well. so that said, no thermal compound will render higher temps, cause the cpu to loose some performance, but doubt it will damage the newer cpus made in say the last 5 years or so as they will throttle or shut down to save themselves. if all you need to do is post your system, no thermal compound is perfectly fine, just make sure to remove the plastic cover first
Ok after video comment. I expected a throttle, color me impressed.
Don’t do this but I was more curious about no cooler on the chip. To see if it throttled and ran or shut down or became damaged.
Thermal paste is just to ensure a smooth and complete contact surface with your cooler, because the heat transfer is made by conduction and to do that there needs to be surface to surface contact its not actually designed to keep it cool, although over the years some companies do try to provide products that help improve cooling
I mean, one expected there to be an air gap, however minimal, which might be slightly insulating, or just not really well-conducting, and temperatures will tend to be considerably higher, and/or the CPU will considerably down-clock to compensate for terrible [or "not as good"] cooling performance.
Much like the "tiny dot" in the prior experiment showed massively higiher temperatures vs. well, ANY other thermal paste method with a "sufficient amount."
But, who knows, right?
OMG.... squeeky didn't die --- nice mirror finish
My prediction: I think it really depends on what kind of heatsink you're using, and how good the contact is. A good quality, well fitting heatsink should (in theory, I think) function decent without thermal paste (maybe a few degrees higher in temp), since metal on metal contact should conduct heat relatively well along with the dissipation from the fan. I'd guess that in probably 80% or so scenarios, the engineering of the proc/heatsink together would be decent enough to compensate.
I might also be biased because I'm still running a gaming desktop from 2011, and it still runs fairly cool (upper 20s- lower 30s C when not under load) despite me never reapplying thermal paste after building it. But I'm also drunk, so let's see what happens.
my thoughts exactly. Having a perfectly flat contact point from the heat sink limits any real catastrophe, especially when it's copper. Water cooling is also a good reason. Try this with a stock cooler and the results will be very different. I kinda wanna see it happen but for the sake of science I don't want his computer to melt.
I bet with a worse cooler (like the stock amd one) it would be a lot worse
@@squidwardo7074 The stock AMD one (the wraith stealth) is not bad. I haven't examined it thoroughly but I'm sure it's probably machined very well.
If he would've locked the cpu fan speed to a static rpm, I'm sure it would be a few degrees higher than we just saw. I would estimate around 2-4°c more.
I would definitely still use thermal paste. I am new to pc building but I have a pretty good understanding of electricity. So I would think it also protects against corrosion and helps protect from static shock issues.
The IHS is solid copper that's milled and then nickel plated so it doesn't rust. You can wash it with regular water.
Don't understand what you mean by static shock. The actual chips inside the CPU package are covered with non conductive TIP and then they press fit the IHS on top of everything. The IHS does not touch any electronic component.
People love to talk about micro air gaps and all kinds of other dumb shit when they talk about termal paste but the reality is that the paste is mostly there in case the IHS and/or the bottom of the radiator isn't perfectly flat. On many CPUs the IHS is a little bit concave from when it's press fit. Or was that convex?! Anyway it's lower in the middle that's why you put the drop of paste in the middle. Plus that's where the actual chip was for the longest time before these new multi chip CPUs started being a thing.
And then there's the fact that big heavy radiators are going to sag when they're sitting horizontally and they're not bolted on hard enough.
I've been having headaches over my CPU temps recently. I tried everything I could until I decided to try out a different thermal paste brand. Bought a tube of MX-4 and I went from +95ºC under load to 75-ish
were you using Colgate before?
id imagine the cooler wasnt tight or uneven
@@TechIlliterate damn man, that made me laugh way too hard
I thought the temp diff would be higher. I run a 5950x currently using Arctic mx-5 and would never think about trying to run it w/o thermal paste. I use the EVGA CL360 with Antec Prism argb fans and it currently idles around 122°F with all cores set to 4.5ghz.
I think one impotant question is what happens with not totally flat surfaces. A hair, or grain of sand, between the cpu and cooler, and then the effect of the paste.
Whats that model of CPU? I want a talking one model too, how can i get it? Mine doesnt talk like yours :(
So long as the fan works, then the paste is a good idea but going raw is not worth it. Then comes the question, the paste or a thermal pad?
"Alass for you and for science i turned on the machine" , i know the feeling, you get my sub :}
I would like to think efficiency of heat transfer will matter more when we go to higher watt cpu usages like drawing more power for overclocking same die. For default power draw maybe chips are efficient already.
I once didn't change the thermal paste for more that 2 years. I didn't have any thermal paste left. So the temperature difference is much bigger than what you show, I guarantee it. You just had to play for longer.
im currently doing the same thing lol, ive had my pc for 2 years now and ive never used thermal paste
No thermal paste is different story than old thermal paste.
Now do this test with an original Athlon, lol. The system will post, then pretty much freeze a bit after Windows loads. Honestly, this video did surprise me. Gotta hand it to the engineers, modern CPU and motherboard thermal designs have come a long way in the past 20 years.
If this man still likes comments he’s a king
No u