Is Intel’s Thermal Paste ACTUALLY That Bad?
Vložit
- čas přidán 23. 03. 2018
- By harvesting TIM from old Intel CPUs we can finally find out just how bad it is.
For your unrestricted 30 days free trial, go to www.freshbooks.com/techtips and enter in “Linus Tech Tips” in the how you heard about us section.
Try Tunnelbear for free, no credit card required, at www.tunnelbear.com/LTT
Buy better thermal paste
On Amazon: geni.us/CroMAlu
On Newegg: geni.us/Ale04G3
Discuss on the forum: linustechtips.com/main/topic/...
Our Affiliates, Referral Programs, and Sponsors: linustechtips.com/main/topic/...
Linus Tech Tips merchandise at www.designbyhumans.com/shop/Li...
Linus Tech Tips posters at crowdmade.com/linustechtips
Our production gear: geni.us/cvOS
Get LTX 2018 tickets at www.ltxexpo.com/
Twitter - / linustech
Facebook - / linustech
Instagram - / linustech
Twitch - / linustech
Intro Screen Music Credit:
Title: Laszlo - Supernova
Video Link: • [Electro] - Laszlo - S...
iTunes Download Link: itunes.apple.com/us/album/sup...
Artist Link: / laszlomusic
Outro Screen Music Credit: Approaching Nirvana - Sugar High / approachingnirvana
Sound effects provided by www.freesfx.co.uk/sfx/ - Věda a technologie
actual tittle:
Thermal paste vs thermal crumbles
exactly, not even diamond's surface is perfectly flat, only a liquid will fill the valleys on the surfaces, using powder pieces and considering it decent, is just nuts.
Two workers of the Chernobyl power plant checking the reactor temperature moments before the explosion. 1986 (colorized)
LMFAO
@@vseghal1 but time travel is complicated and could result in even global extinction
Colourized*
@@dextersilva4411 In the US it's Colorized, but in the UK its Colourized...
^
Why didn't they just replace Intel’s thermal Paste while using the same external paste, then compare results?
because they've done similar things before.
They literally said why at the end of the video
They were either hitting on Intel or being stupid or trying to be funny
I don't think it is the first one but I think it is the right one
Cuz amd sponsors then to spread shit against intel
@@albertkrynski3961 You're an idiot
6:04 When you first start dating her
"but we wouldn't know how well it worked until we hit it with an intense load" XD
Oooooo haha
Boom demonetised
This is really a horrible way to test it..
xorinzor
Welcome to LTT
if you haven't noticed from their "Are review sample CPUs cherry picked?" video, LTT is not very good in these kinds of things, except for regular bench tests.
i mean seriously.. trying to reuse 6 year old hardened paste? wtf
What's a better way to test it?
The intuitive way would be, to compare performance Pre vs After delidding, but they actually did that already.
But I agree. All those other factors they wanted to avoid having to deal with, are still present. Just on top of the IHS instead of underneath. Thickness of the application etc.
But this is exactly the kind of ridiculous nonsense that makes LTT fun to me, so I won't complain.
LTT is basically the Marvel Cinematic Universe of tech channels.
Professionally made, entertaining to watch. Not much of anything else.
I DISLIKED THIS VIDEO XD
The TIM is very likely a form of thermal glue. You managed to spread it a bit, but it is clearly visible that all you did was crush it and not to get it to actually grab both surfaces (which is necessary).
It is kinda like scraping of glue and hoping to re-use it - that is just not how it works.
Thank you, i didn't have to write it... x) This test is not valid at all.
as an arts major at a reputable college I can say you are wrong
LOL
Anyway delidding tests show how shit their TIM is!
Thermal Glue is still *crap* compared to soldering.
Also I’ll be happy if this misinformation spreading caused intel/amd to only use soldering.
+Duck Hardly a constructive reply. Care to elaborate?
I've never seen such a stupid idea! ;-)
you're young...and stupid :)
I'm guessing this wasn't in your Liberal Arts finals.
What are you mad at them for?
Sorry, but this 'test' is inconclusive at best.
You recycled old thermal paste from old CPU's and there's no way of telling how they were stored. You then used IPA to make try make it a paste again, but IPA is a cleaning solvent that could very well have dissolved compounds in the paste. Then you claim the paste application to be a reasonably good job, but you can clearly see big chunks flaking off. The layer was and inconsistent.
Then you compare that to brand new paste, applied in a perfect thin layer...
Regardless of what your conclusion is, the testing method is flawed.
Just because it would have dissolved compounds doesnt mean they disappear when the alcohol evaporates. come on man have you never done the experiment at school to seperate dissolved salt and water. jeez man
Yeah this is for entertainment, they never said it was scientifical
The testing method is flaked
@@Davis38 Your joke deserves at least several hundreds of likes.
Are you going to test 6 years old Noctua paste scraped off an cooler as well?
Great point. Is it really that difficult for Intel to supply someone like LTT a fresh sample of their TIM? The video title really piqued my interest - would Linus finally put this question to rest once and for all? But nope, in the end this content only proved to be a total waste of time and effort.
Philip Cooper an is only used for vowels
YOur a moron..find a new hobby
pam0077 I don't think Intel want make everyone see how bad is their thermal paste in comparison to something else
Its still the paste they have on their CPUs..
Wouldn't a better test to be delidding the CPU and replacing with Noctua TIM? That way you have unmolested Intel paste?
lmfao right
Special EDy . They could have even mitigated the variables using a larger sample size, like I don't know, a bag of CPUs.
There are too many other things that you change when you delid a cpu, most notably the distance between the processor and heat spreader. Not saying that this was a better test, but I've heard that changing that distance is one of the big reason delidding works so well.
Changing the distance isn't much a variable so much as a needed measure in order to make liquid metal work, as its relatively flat and not space filling. I'd imagine if someone were to test with Intel's own paste the control vs removing and using a lot less silicone to seal it closed much more closely, the differences would be margin of error.
probably the ONLY way other then buying the actual stuff, we also find that removing the TIM paste then reapplying with normal thermal paste actually doesn't have that much of an effect on the heat exchange. This is why when delidding CPUs people say to use liquid metal otherwise it's not worth delidding. How do people buy into stupidity
When they said Intel's thermal paste I thought they were talking about the default paste that comes with the stock cooler of the processor.
Same
They should test that too.
Btw why does my i5-10400 with the stock cooler with the pre applied thermal paste go up to 100 degrees Celsius? I have 5 fans in the case or is that normal
@@yoursubconscious1774 its definitely not normal, I have a i5-2400 and have gone through this. Its your cpu not making contact with your heatsink or your thermal paste has degraded so there isn't any thermal transfer between them. Rule of thumb: Your Intel Cpu should never cross the TCase temp specified in the manual page of the processor. (ark.intel.com)
It will definitely reduce the life cycle of your cpu and get you in trouble when claiming a warranty.
@@peacemekka So should I take off the heat sink and and re-apply thermal paste or just take it off and reapply the heat so its in better position
damn, the lighting is so much better, keep it up
Who's Tim and why is everyone hating on him. Jeez
Timtel creator of intel
Assassinext it's a joke
Luke Braker do u even know what I mean
Nice poem
Mason Ferry
You're commenting on a tech channel and expect people to click on your totally not suspicios link. Nice try
Least scientific possible way of testing this...
I'm pretty sure Intel's been giving him money to do all this subliminal type of advertising lately. Results or anything doesn't matter. Psychologically, hearing LTT bring up Intel so much influences consumers. He just coming up with random reasons to bring up Intel.
Richard Lynch
I am pretty sure your ctrl and v keys are very tired.
Let them rest.
What did you use to run your CPU tests
7:29 "decent enough application" my ass
"when delidding and replacing thermal paste, many other variables are changed." Yes, and this alternative just seems like a much more accurate way to test it....
I'm pretty sure Intel's been giving him money to do all this subliminal type of advertising lately. Results or anything about this video doesn't matter. Psychologically, hearing LTT bring up Intel so much influences consumers. He just coming up with random reasons to bring up Intel.
if they could get a fresh application as opposed to using old dried TIM then yes, it would be the best way to test it, but without intel sponsoring them to say how bad the TIM they use is they would have no way of getting it to use.
I've seen better scientific testing methods at 5th grade science fairs.
I'm pretty sure Intel's been giving him money to do all this subliminal type of advertising lately. Results or anything about this video doesn't matter. Psychologically, hearing LTT bring up Intel so much influences consumers. He just coming up with random reasons to bring up Intel.
This video is clearly not in favour of intel so why would they be paying Linus to do it?
Those temps are whack, I got better temps with tim and premium brand paste. I'm sure linus already knew, tell by his laugh.
Take a bunch of old computers which all have Noctua NH-1 (or whatever its called) thermal paste applied to the CPU, scrape it all off and then apply it the same way you did for the supposedly shitty Intel paste. THAT would have been at least somewhat comparable. This was a total waste of my time.
@Jose: True. Another interesting control test would have been just bare metal on metal; clean, without any compound. I suspect that would work better than most people would think. After all, you're really only supposed to use the tiniest of tiny amounts of thermal compound, just enough to fill in the microscopic imperfections between the two surfaces. With modern surfaces like those used in this video being very flat and smooth anyway, I'd be willing to bet you'd get damn good thermal conductivity between them just clean and bare metal to metal.
“Worked to some degree”
Ahhhhhhhhhhhh
Can't thermal conductivity be measured for the TIM and for the aftermarket compound, so you can make a comparison ?
Intel™ Brand Thermayo® - Thermal Mayonnaise - Comes standard on all Intel™ Core™ Series processors, and uses only the best thermally enhanced ingredients such as Canola Oil and three week old eggs!
there's nothing really wrong with it. Tens of thousands (x10 probably) are using non delidded CPUs with no issue, overclocking them. The thing with overclockers is (I know, I'm one of them) they don't like to see high temps. 80-90c is completely safe, but of course we prefer to see 60-70. It's really not something that's necessary
Core™ is actually a trademark for Intel™.
iBoolGuy Added another ™, thanks
TDS
And what if it`s not about overclocking? You could use a cheaper heatsink and still have respectable temps. Or just keeping that stock heatsink - that`s not bad at all
Old pastie
I feel sorry for the Intel engineers, they must be tearing their eyes out watching this haha..
I'd say Intel engineers are geniuses. They have managed to get away with using cheaper materials on CPUs such that they are just a the brink but not overheating for so many years
Well, if this bothers them so much maybe they should send LTT some of their thermal compound.
not really considering this is an extremely flawed experiment. take the dust from one of your old computers paste and smash it on a cpu vs fresh intel paste. intel wins
Abaddan your argument is invalid as their performance is clearly superior to AMD.
They manage to squeeze more performens out of every core, thus it will create more heat, its completely logical and nothing to do with them being "bad"...
Yo genius it’s not about the engineers, it’s about the INVESTORS. Engineers make projects at the best of their ability and knowledge. If the investor says I dont have that much money for this solution or we gotta go way cheaper than this so we will go cheap on that it’s not the engineers fault.
Could you de-lid one and put it back without changing the paste to check for the other variables,
or does opening it once ruins the previous application?
"Chunky style thermal compound" Coming soon to a store near you!
.
honestly this test is worse than testing how much it rains using shoes as measuring cups
I'd definitely use shoes... Is there other way?
Yes, DD bra cups.
This was done... pretty poorly. My recommendation would be to get an ACTUAL mortar and pestle so you can grind the TIM in to a fine dust and then using a small amount of Ester Oil or Silicone grease - something that is *actually used* as part of thermal grease as a filler so it can flow. Or mineral oil, or some kind of other oil instead of isopropyl alcohol which was never going to suspend the TIM particles properly.
Seconded. Would like to see LTT redo the test with this advice in mind
Yeah, I was really disappointed. I did just 10 minutes of research and found out most thermal paste should be insoluble in isopropyl alcohol which was clear in the video
It was never their intention to create a new paste, they just wanted the intel tim to be easier to spread. They even say they want the alcohol to evaporate - i mean did you even watch the video? Come on...
Mineral oil. Use Mineral oil.
Not really recommended, but close enough :P Not sure if those perfumes would break anything down over time, but I am sure you don't want your room to smell of babies every time you stress your system xD
What kind of CPU cooler using mate?
I know it's pretty tough to re-use thermal paste...but this test wasn't very conclusive due to all of the variables from reusing the paste (which you did touch on)
Not valid. Applying dried out paste invalidates the experience.
Agreed - they should have compared it to some 6 year old used dried out "enthusiast" paste and see how much difference there was then... Comparing with brand new fresh liquid paste is ridiculous.
yeah, the whole point of thermal paste is to fill in the microscopic gaps between the IHS and heatsink. that dried out flaky stuff is not malleable enough to fill in the gaps and will do nothing
But intel are opting for this paste because of it's longevity. If they have dried out and perform bad then what is the point of using it. BTW not looking for a comment section war, only trying to understand the rationalization here.
I completely agree.
David Martins it wasn't dry when they applied it though
Is thermal paste an instrument?
Scarlet Squid blow it and find out
As a musician, I can absolutely say that thermal paste is indeed an instrument.
No Patrick mayonnaise is not an instrument
"Horse Radish isn't an instrument either"
One of my favorite spongebob episodes ever
Okay, what do you call that delidding thing and can you buy it in amazon?
Could you do a video about the effectiveness of Gallium based thermal paste vs copper grease (for car brakes) or even silver grease, that'd be great. Since that stuff is pretty inexpensive and wouldn't corrode the heat-spreader/heat-sink.
Plot twist: Intel's new gen chips were never been more efficient. They were just using better thermal compound each year!
Who would've thought imporving just the thermal compound would be cheaper :D
Sadly, that actually sounds like a really probable move from Intel.
That's not bad, improving the paste to get lower temps.
Except delidding doesn't reduce TDP it reduces temp.
No dude, using better TIM doesnt make CPUs go faster. Unless all previous gen CPUs were throttling by default even with high-end coolers on them.
Are you guys actually thinking this is a fair comparison after what you have done to that thermal paste?
Is there any performance variance when graphics card connected to a projector or Monitor ?
i pretty sure they use a setting type pf paste. more like an epoxy than grease
No no, this must be one of the most unscientific "test" ever conducted.
3:52 This about sums up the entire video.
yes!
seems thrown together and shoved out the tubes
What temp program do yous use
Use pine sol or turpentine, it works better for thinning old paste. I actually 'cut' many modern pastes that seem too thick with it -- smells funky for a couple hours, but it knocks 3C off the result of even most premium TIM's if you do it right.
This might just be the dumbest ltt video I've seen in recent memory
Rockmandash12 its hard to pick the dumbest these days... They sure went downhill
The one where they spray-painted a PC black (parts and all) was the worst one imo.
how about making multiple videos about 56 cores cpu,s and act like its a never before seen thing every time
Rockmandash12
Yea i agree it is really bad. He has been putting out a fair amount of nonsense lately. You can tell there is no new hardware for him to review or any new ideas of any really decent content in the past few weeks or so .
DerBurner132 #deleteLTT
25° is the difference I got between no thermal paste and nth-1 on a Chip with similar TDP and with a huge air cooler with a big contact surface. That doesn't mean that intels tim is as bad as no TIM at all it just means that 6 year old tim scraped off from a chip is as bad as no TIM.
you should have tested it as usual but the base test should've been done after opening the processor - so you open it up and dont remove the original paste and then close it so that the seals are broken. and also do it on 7 or 8 series where paste is still in a good condition
Linus should take a new 8700k bench mark it for the temps, then delis it and replace the stock paste with the large tub thermal paste he got from AliExpress, do the test again and see if there is any difference. And use liquid metal on the IHS for both runs.
That doesn't look like a "decent enough" application to me but doesn't matter anyway since no one should take seriously a thermal paste comparison done this way.
If it can't stand the heat stop using it in the kitchen!
lmao i read this in a game character's voice
my soups cold -_-
RandomFlatShowsvideos hey if you’re interested in tech channel, please heck out my channel.. I just started and would appreciate some support! 👊🏼👊🏼👊🏼
Ryan LeFevre wow. Ok man. Thank you
DIM SUM TECH fuck off and stop advertising your channel.
which delid tool are you using?
I don't know why, but when Linus took up the paste near his face i was going to sneze but hold it anyway
You should have done a test with the thermal paste that comes with Intel heat sink....
Gabe Catell do you know they are the same paste or are you assuming?
We can only assume. But it's a good suggestion. At least you can get it fresh.
They even look completely different, so...
This video is disappointing on so many levels.
100% agree, using a couple brand new out of the box Intel heatsinks and a couple with the thermal paste wiped off would have worked so much better instead of torturing those poor 3rd gen chips
With your hook up at Intel...You couldn't get fresh paste from them to use?
that's what i'm thinking as well. As compared to Linus' expensive free toys, a jar of TIM is like a cheap jar of pbj.
I don't they would've, honestly. I think an LTT video is not worth the risk of destroying their claims.
You're an idiot.
There's no hook up, you're delusional.
Intel would never sponsor a video like this.
If anything, they are probably hurting their chances for future sponsorships.
Clement Wong what does it mean a jar of pbj and what's pbj anyway ?
Are the delidded and thermal paste missing CPUs for sale? Pls let me know :)
i have a question, i had delid and lapped the IHS and used arctic MX4 but the temperature remains high at 89-90 on 4770k @ 4.2ghz 1.12~V. do i need to remove the silicon glue to have a better contact from the IHS to the CHip?
Maybe a Better way:-
1. Take 8700k
2. Delid it don't replace TIM to eliminate the variable of contact distance between the IHS and CPU die
3 benchmark
4. measure the volume of Intel's TIM (granted you have to measure the density and mass of the TIM) OR just eyeball the amount
4 take the same volume of NT-H1
5 BENCHMARK
6 GO HOME.
"granted you have to measure the density and mass of the TIM" that is impossible. Why don't you go work for intel and steal a jar of TIM?
Nicholas Lau or just eyeball the quantity of thermal paste ;)
I was thinking this was what they had planned, a little disappointed they went the other way.
Also show temps as delta to ambient and compare to the white silicon grease thermal paste too.
Richard Blackwell *BITCOIN IS DEAD*
What you said at 8:50 is a better comparison than what you did. This was honestly a pretty poor execution of this comparison. I expected LTT to test, then delid and replace the TIM on say, some of your own CPUs. Improving your own thermals as a result. That reconstituted TIM didn't wet out to the cpu at all and didn't spread like the Noctua paste. That junk was 5 years old and crusty. And while alcohol cleans it off, I don't think isopropyl alcohol is what you Should use to try to revive it. That went over like a lead balloon.
I'm pretty sure Intel's been giving him money to do all this subliminal type of advertising lately. Results or anything doesn't matter. Psychologically, hearing LTT bring up Intel so much influences consumers. He just coming up with random reasons to bring up Intel.
Richard Lynch
You do realize that an undisclosed sponsorship would be illegal, and thus one hell of an accusation without a shred of evidence to back it up, don't you?
Not to mention, conspiracy theory or not, this isn't the type of advertising any company would want. There's nothing subliminal about talking about how crappy their TIM is.
No, this was a stupid video, but also nothing more than putting in a ton of work and deciding to see it through. It was just stupid enough to share.
OwlOfBlue
Oh, so he is one of those guys then. Never actually seen this shit before, I just go in assuming full of it one way or another.
Honestly, sometimes you can't even tell these days.
You do realize it doesn't have to be always white or black there are also shades of gray. It is true techniques like the ones mentioned are shady but legally speaking he is not obliged to mention any "sponsorship" if the video itself isn't directly affiliated with the company. For instance, Intel could be a sponsor of LTT supplying them with products and maybe money in exchange of advertisement (there is no bad advertisement). To clarify I am not suggesting this is what LTT is doing it's just that there are so many videos of him talking about Intel products suspicions arise. Don't be absolute for things you don't have insight of and this goes both ways :)
he actually is obliged to.
That harvested TIM looked very much like the old, baked out MX-2 I had to clean out of a really hot-running laptop. (I know it was MX-2 because I applied it about 3 years prior.) Replacing that with fresh MX-2 (from the same tube the old stuff originally came from) produced a 15°C difference. And just like this one, the old MX-2 wouldn't come off with alcohol - it just retained the same rubbery consistency as before, not even discoloring the alcohol -, while fresh MX-2 would dissolve somewhat.
My guess is Intel's using the same binder material found in MX-2, and it polymerizes under excessive heat.
3:33 I like the geeky guy he makes a good parking brake for Linus's Mania.
Couldn't you scratch it out of box coolers? Isn't it the same stuff?
Agnar Selvik it's not
Problem. You are testing paste that has been heated.Then you exposed it to air. After heating the make up changes. Then it is exposed to air. To get a valid test you need to heat up the after market paste. Then expose it to air. The key is air exposure after heat. Your reconstitution is the problem.
Carter Baker I mean, the results are about right from what we see via deluded chips.
Yes Linus can be a dingus, but this isn't really one of those times.
this realy is one of those times, there is a reason you need to replace your thermal paste every time you take your cpu cooler off, so why should it be any diffrent with the intel paste??
I guess by using old, scraped off paste he can kind of "blame" the difference on that fact. If Intel had sent him a tube of brand new one and it still performed that badly, that would hurt Intel's reputation more.
Well, if Intel wants to get some good PR, they only have to send a tube of paste to Linus now. Until they don't do that, I take this video and all the delidding data as evidence that Intel's toothpaste is really that bad.
As an engineer I can say this test was a fucking joke
Can you please share the temperature results without any thermal grease?
It would be awesome if you managed to find a way to solder it together 🙂
Well you said it best “This might be one of the dumbest things we’ve done.”
RIP Intel sponsorship for LTT
At least they will make more AMD builds.
RIP tunnel 🐻
What do you call the delidding tool you used?
Sorry Linus, but your test methodology is rubbish! You really couldn't get Intel to send you a jar of the thermal paste they use? They sponsor your content all the time! Did you even ask?
It seems like Linus doesnt understand how thermal paste works, in the end after they remove the heatsink you can see how thick the paste is and there is 0 metal on metals contact, thermal paste is meant to fill the gaps where the metal isn't making contact due to the imperfections in the materials, if both surfaces were 100% perfect then thermal paste wouldn't be needed, that is why people say to use as little as possible, its not like glue where you want a thick layer all over, just enough to fill microscopic holes.
Bubbauk88 100% correct.
It's probably not bad, it's probably a standard grade material that has like right around 5W/mK, which usually in real use is just a couple degrees off the highest performance compounds. However it is apparently tasked with covering too much gap due to bad heatspreader design. Those are thick slabs of thermal compound they dug out of there, no conventional TIM could ever be effective this way.
Come on you guys, linus does marketing and promotion, material science is out of his reach for such a sudden idea, plus the guy says what is on his mind so you know something as rediculous as this is possible and thankfully a hi- definition camera let's us see the bad and the gooo.......nevermind, I agree, linus does look like a imbecile while mixing old dusty thermal paste with isoprophyl alchohol to try and bring it back to life.
Terrible comparison! Why not test all of the CPUs, get a load of measurement data and then try replacing all of the TIM on all of the CPUs with a different type, test, then try another type on them all, test etc.
Maybe even compare it against other pastes that aren’t for CPUs, like toothpaste, copper grease, red rubber grease, silicone grease, aluminium grease, lithium grease, ibuprofen gel, sexual lubricant etc etc, anything like that will at least give a good comparison and would make an interesting video.
evilution "lithium grease" ayy lmao
shapeshiftingpedro wut
evilution ltt isnt supposed to do this stuff anyways... This is gamer's nexus' job
HAHA this is LTT aka tech for noobs and kids, what did you expect GamersNexus or Level1Techs type of in depth testing and results? That would require more brainpower and effort than LTT has ever shown tbh.
Hell just look at fn Wan show they have problems every gd week and never gaf,audio so bad last week I couldn't even make thru 10 min of the podcast. It's fn sad how far and how fast they going downhill.
my cpu is throttling and i was thinking the thermal paste was badly applied by intel or the cooling paste sucks so should i put on a good brand of cooling paste?
Why are you not using ICDiamond 24 vs Noctua NT-H1?
This doesn’t seem like a good way to test 🤔
Dislike. This test is flawed.
Stefan Yasin no one cares.
Instead of just blurting out stupid shit, why don't you tell us why you think the test is flawed?
Because he did not use the paste pure or properly. It is reused and mixed up. Its like comparing tape to used glue with blu tack mixed in.
You want a very very fine film to fill all the small gaps that could potentually trap air, just watch the spread on the new thermal compound it is even and rather thin, watch the intel compound it looks like spackle is cracked and hard.
end of video explains most of that
What keyboard is that?
Does the stock intel cooler not come with paste pre applied? Its been so long i cant even remember.
It makes sense that they'd choose a TIM that lasts a long time over one that performs better but lasts a short time. Quite a few consumers would be pissed if after a couple year their CPUs throttled over the smallest loads.
Yes it does, but why not keep doing it like they did up to Sandy Bridge and solder the Heatspreader?
Soldering their chips would have infinite durability and be barely more expensive.
Soldering actually performs worse in the long term if the temperature changes frequently.
Microcracks and co will increase thermal resistance and could even damage the die depending on how it is produced.
But mostly - Intel wanted to safe money and the CPUs worked as intended with the thermal paste.
Maybe but they used to Solder the IHS which is the best solution which results in very good temps and is second only to direct die cooling.
"Quite a few consumers would be pissed if after a couple year their CPUs throttled over the smallest loads."
... so instead they use a TIM that lets the chips throttle from day 1 (e.g. i7 7700K). Yay, what a great company intel is! Admit it, intel is just dirt cheap & greedy and saves every penny even on their high performance chips, although it hurts performance and life span when the chips run so hot. intel definitely *does not* use this thermal jizz for the consumers benefit. You gotta be terribly naive to think that, while the competition solders even most of their budget chips!
Really scraping the bottom of the barrel for content.
I'm pretty sure Intel's been giving him money to do all this subliminal type of advertising lately. Results or anything about this video doesn't matter. Psychologically, hearing LTT bring up Intel so much influences consumers. He just coming up with random reasons to bring up Intel.
Two sponsors in one video? This is getting out of hand.
Nah, you'll be fine
hey linus
i think you should have used noctua thermal paste inside IHS(Integrated Heat Spreader) and compared it with stock paste in IHS.
So you're crunching up flakes of air-dried thermal paste with a plastic stick in an alchohol bath, and expect it to actually WORK?? You're just completely destroying the sole purpose of thermal pastes - to fill the micro-gaps between surfaces and providing a smooth, uniform medium to transfer heat.
And do you know where I got that definition from? From your own, previous video explaining about thermal pastes.
I know you're desperate for ideas and that youtube money, but at least have some integrity between videos.
Khoi DX shut up and make your own video or get intel to get your ass a jar of intel brand TIM
It's called Dow Corning. You actually CAN get jars of it.
Amen
i agree but the last line is fucking pointless
You make a valid point, but your logic is flawed. It doesn't matter if he scraped it or if it was dry or whatever, that's still exactly what came out of the CPU. At the end of the day that crap is still what is inside your CPU, even if it is old
Congratulations. You just compared 6 Year old, baldy distributed 'paste' (which also might reacted with the alcohol - who knows) with some excellent new one. Also, it might be interesting if the "better paste" will do it's job after 6 to 10 years (because Intel cannot tell its customers 'well you need to replace your thermal compound every 12 month or so').
I am not saying that Intel is not using bad thermal compound (I really don't know) but this test is telling us absolutely nothing at all. And no, your application of that old dust was *not* even close to being acceptable. WTF guys.
Like many said - one of the worst LTT videos I have ever seen.
I hoped for the chip to burn because of the alcohol applied to the thermal paste, atleast the video would be funny
Lol
They brought up the longevity of intel paste several times in their video.
Linus - "can you send us some of your TIM so we can test it?"
Intel - ".... BAAAAAAAAAHAHAHAHAHAHAHA!"
0:57 when linus did that i was like i dont have a processor half the power of those processors in that bag and he is hitting them i was like pls stop your just hurting me more than the processors
5:02 "i spilled the alcohol" seem like Linus can´t handle his liquor...
So how is this better, and has "less variables", than just changing the thermal paste under the IHS for smtg else and comparing the results?
Well many argue that the silicone glue and the spacing is more responsible for poor temps than the actual paste. Maybe that's hard to duplicate once you take it apart.
Its simpler because they are only testing the difference in thermal conductivity between TIMs and you can learn that difference without needing to delid between each test. It doesnt matter where you test the difference as the only data you need is the actual difference. This test could for that matter could have been done on an electric stove with TIM between the stove top and a pan filled with water and then measured what TIM had the water in the pan reach 80 degrees faster. You would still have gotten the same data AKA the difference.
Except that they weren't just testing the differences in the thermal conductivity. They introduced several other variables, like the thickness of the application (did the powder slurry spread out as thin as fresh TIM?), the amount of air between the cold plate and IHS as a result of the variable size of the powdered TIM, etc, etc. They should have tested the cooler without any TIM to see if they weren't just making the performance worse with their TIM slurry.
They did the best they could with the materials they had, but I do agree that a test without any TIM could have gotten some possibly interesting data(but it would just been a data point that confused the point of the test and needlessly created more work that gained nothing as it is a given that almost anything apart from a straight up insulating material will preform better to some degree than using nothing).
I'm pretty sure Intel's been giving him money to do all this subliminal type of advertising lately. Results or anything doesn't matter. Psychologically, hearing LTT bring up Intel so much influences consumers.
YES , Finally *TUNNELBEAR IS BACK*
A fucking miracle
No more dyson?? I'm so sad... Was it only one timer...?
Bad news mate
SourCan its gone again for good
Lol they just dropped tunnel bear
wait. What was the result again? was the stock tim good or not?
I wish you would have run the heat sink dry on the cpu without thermal paste at all to see it would have been better then just the Intel paste.
can you guys try to see if a copper sheet ( wafer ) makes a good thermal bond between the cpu and cooler ? a wafer thin sheet of coppwer pressed with enough force to cover all the blemishes of the contacts
i tried this with copper tape of several layers with a water cooling block + 3 peltiers + cold dissipation aluminium heatsink and the results weren't amazing. I ran out of thermal paste and the other stuff arrived before the paste so i also tried toothpaste mixed in with olive oil as a "i don't feel like spending 15$ buying thermal paste from a hardware store when my 1$ chinese paste is in the mail" -- i didn't get great results from either beforementioned method but both have some credibility to them if done correctly... which im sure i didn't.
Just test the processor with stock one, then with another pastes, do it with processors of several genarations and based on the difference of stock aftermarket paste of which one, you will know how processors get better or worse with time.
Hi i have a nokia 6 and 1 drop of sweat entered în my jack port and now the sound doesnt work , do my waranty repair it ? I dont know what to do ( sorry for my bad english )
I wouldn't think that you warranty covers sweat damage, but you could give it a try. Or get a new phone that is waterproof. Or an old waterproof one
"intense load"
So lewd...
wow that was so wrong. like srsly this is no paste anymore it got worm and pressed together but cause its this dry its just a hard plate. i mean... if you compare just the looks, you can see its not a fair method. why dont you get a new cpu, test temps and then delid it but you take thermal paste instead of liquid metal. Like you just replace the thermal compound with more expensive, aftermarked stuff....
I have been reusing dried powdery Silver TIM for a decade or 2, whenever I didn't have enough good new TIM. I just add a bit of white thermal compound, (sometimes you get the added benefit of of an oil separating out of the white compound which expedites the process) to the silver TIM and mix well. The result yields a smooth, albeit, viscous compound that's a shade or 2 lighter grey colored and works fairly well. That being said, I never did this with harvested TIM from under the lids, only TIM located between the heat sink and CPU. I use a white small jar of Wakefield compound, which is so old I can't read the part number (it has mostly rubbed/flaked off). There are probably many white TIMs that would work assuming their chemistry is similar. Easy enough to revisit except for the fact you need to delid another bag full of CPUs !
How do they put load on cpu and check the temperature?
Regardless, the IHS should be soldered anyway.
If only Intel thought that way
Why? If its cheaper and easier to use paste why would they solder?
Companies and the humans in them look out for themselves. Get woke.
If you want better paste delid it. If you want to solder it go ahead and try. Its way more complicated than you would imagine.
EHEH RON: AMD uses the solder method for all but their particularly low-end models. All the Ryzen CPUs have soldered IHS's, and this was also true for their older processors like the FX and Phenom/PhenomII series. So they've been doing it for quite some time, which makes me question if the intel thermal mayo method is in any way cheaper or easier.
EHEH RON How much do you think they're saving by using paste? Because it's not much.
It's not that complicated either, they just put a ball or two of the stuff that they use on the die and then heat it up until it melts, just like surface mount components.
>If you want better paste delid it.
Brilliant thinking there. That $300+ processor you payed for, or that $1000+ Xeon, you don't deserve to have it soldered. Just delid it, risk damaging your CPU, and lose your warranty.
Dear lord thank u for the tunnelvear ads being back
linus should use the bulk purchase thermal compound (from the video Is Bulk Thermal Compound Worth It?) as the thermal interface material below the IHS and see if its any different compared to the stock stuff
My i7 3770k is still rocking a 4.5ghz overclock 5 years later, running 42c on idle, 75 on full load with 1.275 volts and a Corsair AIO H105 that's never been replaced or even remounted.
intel cpus would reach worldwide dominance if they used chocolate instead
You should mix together some different brands thermal paste and see the results on the heat disaption.
Or make your own thermal paste
tech ingredients did a video on this
Linus, just FYI, the glue or any type of glue adhesive holding the lid to the cpu is what causes the temp to spike, no glue at all with better thermopaste will lower the temp...
Heya...how about harvesting the consumer-applied paste from years-old CPUs and see whether longevity is even a valid concern?