Does Thermal Paste Need To "Break-In" ?
ฝัง
- เผยแพร่เมื่อ 20 มิ.ย. 2024
- I have received several messages about thermal paste break in time. Specifically in regards to arctic silver 5 thermal compound. Does it really need to set, cure or break in? Let's take a look.
Support us on Patreon :)
/ techilliterate
Or you can hit the easy buttons, just LIKE and SUBSCRIBE. It really helps a lot!
0:00 Does thermal paste need to break in?
1:00 Arctic Silver 5 Recommendations
1:50 Test Parameters
3:12 Cinebench R23 CPU Temps
3:40 Cinebench Multi Core Score
4:01 F1 2020 CPU Temps
4:37 Conclusions
5:13 Doubts about results
6:20 Bloops
Music: Perfect Ten To Me by NoMBe and Through The Atmosphere by Chris Doerkson
/ chris-doerksen-1
Does Thermal Paste Need To "Break-In" ? - วิทยาศาสตร์และเทคโนโลยี
Thanks for watching. If you need some AC5 You can Get it here -> amzn.to/3tVR0PV
Arctic Cilver 5?
Your aio won't help as that is not max temping, I feel stock cooler you may see some difference but I'm doubtful
Your ambient room temperature can cause this variable
Exactly what I was going to say - especially when the test was carried out over MONTHS, during which time the ambient temperature HAD TO CHANGE.
The result seems perfectly reasonable. It is safe to say that factors other than paste break-in (change to the ambient temperatures, changes to Windows and software used - such as updates - as you mentioned, etc.) affect temperatures more than changes to the paste over time. Assuming a proper and consistent paste application, of course. Some people are pretty OCD about paste, and as you point out, getting perfect controls over all variables is so difficult as to make conclusions arguable.
The thing I would've liked to see or know was if the CPU was boosting higher after it "broke in" according to the manufacturer which could explain the temps.
@@Adierit Wouldn’t the c23 score be higher if the cpu boosted higher. In my opinion the scores where just a margin of error.
Seems we need a test where a perfect 200 hour "broken in" thermal paste is surgically removed from a worn out system and inserted into a pristine new system... to really see it's benefits.
All modern CPUs will boost at different rates to maintain target temperatures, so you may never see a temperature change, although performance could be changing. There's no good way to test, aside from manually locking the CPU clock rates, or just use Cinebench scores as the metric (if temps are same). I applied Kryonaut over the holiday break to my 3700X and noticed zero temperature change, but then I realized my Cinebench score bumped 500 points...was obviously boosting longer, targeting the same temps with improved cooling.
thats true only if you termal throttling which is clearly not the case here with 55 degrees
@@freezerain That's how CPUs used to work. Now they boost to their thermal limits. I think Intel puts some arbitrary time constraints on their boosting, but AMD doesn't. But even then, I don't think any normal cooling system will be able to maintain max boost indefinitely.
you probably uninstalled some bloatware and don't even remember lol
I'm glad you pointed out that you didn't have a control. You have no idea how much more the temps may have risen over the course of 295 hours if you hadn't done a break-in, if at all.
wow what a gold mine of videos i found.
congratuations on your channel. more people should watch these videos
wow! this must have been a serious undertaking to gather all this info with testing! wtf. really nice job done there!
This helped out a lot. Thank you for the hard work.
I have experienced about 5C drop within 2 days with Arctic MX-4. In my case CPU is Ryzen 7 5800X and cooling is Arctic Liquid Freezer II 280 AiO cooler. Now the thing is tubes on this AiO are thick and heavy and I had a bit of trouble putting the pump on the CPU. I definitely did not have even pressure as it tilted while I was putting it on so I probably ended up with very uneven spread with much thicker paste on one side of the CPU. It was probably the worst case scenario. When I tested it shortly after putting it all together I was hitting 85c under Cinebench. When I retested it 2 days later it never gone above 78c. My assumption is that if thermal paste application was very uneven then its possible it will even itself out over some time and a few thermal cycles but if it was good from the start nothing will change.
I did a rebuild last year after 3 years of use with AS5, the paste was still fresh and sticky and could have just been reused if I was desperate.
this video explains alot,i have been building computers for years and i just now learned this about thermal paste. thank you
break-in is just the marketing excuse for snake oil, everywhere
I wonder what the performance of snake oil is.
each thermal compound is different, i've been using AC5 for many years, and have seen similar results, the best performance is seen shortly after installation but normalizes and works fine for years, my longest run with AC5 has been 5 years, and it worked fine with a CM Hyper 212, with out issues, just dusting it out about once every few months.
Thanks for the confirmation.
I want to setup a PC just to test out the longevity. Might have to built a server :o
In the past I've typically only replaced paste every time I put in a new cooler or CPU. The longest run I think was about six years and I noticed no decrease in performance. But I was using a lower-power Phenom processor (45W TDP).
Thank you. I was wondering about that long time.
4:41 Nice case!
That made me on the lookout for something else at 6:02
👍Cool video. There aren't enough thermal paste videos. The unsung hero of PC gaming - Thermal Paste.🏆
That one-frame visual pun is maximum effort
Until recently (new system build) I ran AS5 thermal paste on an AMD FX-8350 with an Antec Kuhler H20 H600 AIO. No break-in period and was using the PC heavily from the start. Rarely shut it down aside from the odd power outage or random software/windows update requiring a reboot so it was pretty much running 24/7. Temps were never below 25C and never above 70C and I live in Australia where it gets extra hot during the summer time. I never needed to reapply the paste and never removed the cooler block and the PC was fine for the 4yrs this paste has been applied. I only ever applied the paste using the dot method and not smear it all over the CPU. I've also used AS5 on a previous rig that lasted me 8yrs with the same paste and that rig had an air cooler on it and not an AIO. Same as the rig before that. :)
This new build I'm using Thermal Grizzly Kryonaut Extreme and smeared it over my 5950x as per the paste's instructions. With an NZXT Z73 AIO these chips seem to run a bit hotter but I'm yet to see temps above 77C and below 36C however I'm curious to try AS5 again as it's been my goto paste since forever. The Kryonaut is a pricey paste, you get bugger all for the price and rather sticky where as AS5 isn't and felt easier to apply. Also, I still have a bit left of the 3.5g tube of AS5 floating around in my draw. I repasted my old HD7970 with it a couple years back after replacing the fans on it. I only got one application from the Kryonaut.
TL;DR: AS5 has worked pretty damn good for me for 15+ years over 3 different rigs and a GPU. However each to their own, whatever works for you is best!
"@ least in my case" 🤣👍🏼
Since the tests were done with a Ryzen chip, was the clock and voltage the same across all the tests? I mention this as Ryzen can be inconsistent with voltage/frequencys, higher temps could be caused by the chip running a higher voltage
Thanks for the video!
My OC never holds after a repaste but after a few days of normal use i can set the exact same OC & it's rock solid again.
Your r23 scores clearly show higher avg scores after 181hrs. This tells me that the Ryzen CPU is able to sustain higher clockspeeds before reaching major thermal throttling because of extra thermal headroom. You should have also posted F1 2020 fps scores to confirm this. My guess for the higher temps in F1 is because the CPU is running faster (because of extra thermal headroom) so your GPU is also pushing more and generating more heat in your case.
I think you need to eliminate the variables on this type of test e.g. set a hard limit on cpu voltage and clockspeed that doesn't hit thermal throttling, as well as post ambient temperature differences.
Boost behaviour does make it tougher to measure thermal performance. This applies to PBO, and even GPU boosting algorithms. This is an excellent point.
I've been using AS5 for around 15yrs now and still works great.
I reached 10. High five for double digits! ✋🏻
13 years, woop
The problem is that the recommendation for Arctic Silver 5 was written in 2003 and CPUs in those days worked differently. To test the claim today, you would need to lock your CPU at a set frequency (at or under base), then do your thermal tests. The game fps and productivity test would be useless, as the frequency shifting is part of the magic modern CPUs do to get their performance today.
Please help. When I placed my new fan with already added thermal paste on it my temps were around 20 to 30s temp. Now since I cleaned and added new paste my computer is now around 50 to 60 temp. Do I need to wait it out to give the new paste time
Probably not break in. You have bad contact...maybe too much pressure. If you're using an AIO make sure the pump is plugged in.
You should capture temp delta over ambient
I notice the "new" AC5 is different color from the ones I have purchased many years ago. It was more of a dark brown color instead of the grey color. Also, the consistency is more of a sticky paste than the standard runny paste. I wonder if either they have changed the formula or there is a lot of fake ones being sold online.
@Ms Moon Boo I think they changed formula. I have switched to Kryonaut and it’s much better.
Spreading a nice thin layer of Arctic silver 5 on my 5800X3D with a PBO curve undervolt -30 all cores and a power limit on PPT, TDC, EDC of 100, 70, 100, got the exact same results as Noctua NTH1 (with all the same settings) and the AS5 spread out a lot nicer too. Other pastes are really thick, cement like affairs compared to the smoothness of Arctic silver, you only need but a little bit and you can work it out really well with a flexible plastic card. I prefer it and I find it didn’t make any difference in performance at all, so it’s up to what you like using. They’ll all do the job, just in differing ways that suit different preferences.
the one frame of the video with a computer case was interesting
Used AS5 on my Xbox, great stuff.
Ambient room temp was constant?
I have Arctic Silver 5 on a i5-6600 and an NH-D14 still stable with same temps after nearly 9 years.
for Kryonaut, I've seen increasing temperature which requires me to repaste annually.
Scar 3 running a i7 9750H locked at 3.5 ghz, max temps at 85-87c starts deteriorating to 94-95c max every year. some of my friends suspect chassis flex from having the laptop carried here in there change the way the paste touches the copper plate.
Try Noctua's paste instead of crazy over the top TIMs; probably handles flex better than Kyronaut.
I had a similar experience with my desktop CPU’s and gpu’s. Paste lasted between 9-15 months. Temps were higher by as much as 10 degrees. In one particular case, it went from the 70s, to suddenly shutting off due to overheating. I’ve since switched to mx-5. I’ve only been using it for a few months though.
thats pump out effect
@@tiiiooooooo I think it’s rare with NT-H2, my laptop is going strong on it… But it’s been just a month
5:07 We need that video if possible. Paste longevity. No one does that vid. It is as important as the paste performance.
It does vary a lot based on the brand of thermal paste, how hot the system gets and for how long, etc. It must be very hard to find a control for so many variables and likely why there are almost no videos or concrete evidence at the moment.
I had a Dell desktop running a Pentium 4 with the stock paste, and I opened it 14 years after buying it; the paste was still wet in some areas. Meanwhile my overclocked 6700k had Arctic MX-4 and was dried out in only two years.
I have a confession to make; my previous motherboard and processor served me for 9 years and I never changed the paste.
But in my defense; I never needed to. I did check up on heats and... if it ain't broken, don't fix it.
So that'd be a long test.
(Now, would a new paste applying managed to reduce the temps? I'm sure it would have, but since I was always very safely below what would have been a problem, I didn't see it necessary)
You need to record degrees over ambinet, not absolute degrees
What is the ambient temperature? How is that being controlled.
200hrs to break in? That's insanity yet we seem to think that's perfectly reasonable.
I have a heavy air cooler for my CPU. I'm installing a new CPU, I've had to re set the CPU twice because I get black screen and CPU light on Mobo. I'm about to try tightening the brackets, which seemed slightly loose. Also I'll keep my PC on it's side for a while.
I once tried a carbon based thermal paste, and it performed better with time.
Thanks for that info! I may have to try it again.
I wonder if in reality you were just getting into winter
Honestly, there are too many variables to control for with this kind of test, simple things like ambient temperature or even higher humidity effecting the performance of fans could skew any results.
I remember when conducting an extremely sensitive experiment in my lab, we had some strange results we couldn't account for until someone worked out it was probably due to a football game, during half-time the power grid fluctuated ever so slightly due to millions of people turning on their kettles during the half-time break and subsequently ruining the results of one of our experiments as we were dealing with very sensitive equipment, we now have equipment to exclude that variable, but my point is its almost impossible to account for every variable since they are practically infinite.
A change in temperature over time could be because the thermal paste is "breaking in" or redistributing itself, even with a hundred PCs to account for random things like changes in the vapour chamber, faulty fan bearings or an antivirus suddenly deciding it's time for a scan, you'd still have to account for changes in software, weather, batches of thermal paste, power configurations, and fix all frequencies and fan speeds or face the wrath of chaos theory.
My best uneducated guess is that the "break in" time is just included in case the application of thermal paste is a little uneven and causing higher temperatures, and then their "break in" theory is the thermal cycling should redistribute it over time, or waiting 200 usage hours is the average time their marketing department believes it takes for customers to forget their thermal paste isn't as effective as advertised and enough time for customers to forget there was a money-back guarantee regarding being satisfied with the product, or something more or less cynical.
Well something i will say is i got a new stereo amplifier and straight out of the box it sounded really harsh and unnatural in the treble. After about a week of leaving it playing overnight for a few days i finally sat down and played some music and OMG it sounded amazing. So people willl say my ears got used to the sound but i never listened to it while it was breaking in overnight. So take it from me i heard the change and im not crazy i even phoned up the store and spoke to the woman in the shop and i explained to her what happened and she confirmed it for me saying yes thats exactly what i heard straight out of the box to confirm my experience. So people that say burn in is not a thing are just people that havent actually experienced it themselves but still seem to have an opinion haha
I like this man's voice
Using cows as PCs BRILLIANT!
How are any of these tests valid when you didn't control the ambient temperature?
Thermal conductivity is better when the thermal interface is wetter as conductivity happens at an accelerated rate due to the moisture particles in the paste, this is also the reason why liquid metals thermal conductivity is higher, for a test you would need to try a thermal paste that has been used for some time and dried out,
As for a control test you could use a peltier cooler as this would give you far more control of temperature and or cooling.
LOL... Half truths still make a lie.
i bought ebay paste 2w/mk max temp 91 @ 4.1ghz .... changed to cooler master gel pro 8w/mk max temp @ 3.8Ghz 95 .. why can someone explain?? huawei matebook D14 intel i5 10th gen
yooo how long does the thermal paste last
Well my day 1 ryzen 1700 was just retired with AS5 on it and the temps were the same as always, so minimum of 4 years. It's reported to live for up to 10 years, beyond that maybe
Depends on the scenario. If it comes stock on a laptop or GPU I would change it within 2 years. probably 1 for my taste. Self applied like in a DIY desktop I would go for 2-3. Personally I am changing it more frequently. I always suggest changing only when you need to...when you run into thermal issues. Some paste dry out sooner than others. YMMV
My stock paste in laptop lasted 9 years. Then it became dry . I scrapped it carefully and added a drop of new 10W30 engine oil just enough to have a spreading viscosity and it's up and running 2 years now without any heating issues. But I clean my cooling fan once every 3 years of its life, so the pasted was never overcooked due to a clogged fan. When fan is clogged with dust the paste looses it's potency due to over cooking. Temp range from 40-72 deg C. Now I'm planning to repaste it with new one. Which is why I came this channel.
Thankyou "Tech literate" for ur testing . Nice job. Keep it up.
PS : I'm not a gamer. It's a calculated risk I had taken, as it's my old but main laptop.
Would you like test a Brazilian Thermal Paste ????
My thermal paste break-in 1200 years.
The advantage of as5 is that it lasts a very long time. Other pastes dry out
That's like saying paint only lasts while its wet. I mean it gains more properties after it dries. Many materials do the same.
@@ShadowManceri are you seriously comparing paint to thermal paste. Thermal paste loses efficiency when it dries out. It doesn’t completely stop working.
@@zipper978 I just use silver paint =D
So what about new servers? I mean do they break them in at the factory? We get them, install them in the rack, turn them on and they run 247365. Unless there is a power outage, then they never get this warm-up, cool-down period.
I found out that in order to get arctic silver 5 to work as it should, you first need to apply a very thin layer, spread it very well and eliminate all the air that could be in the middle, i have always got very nice results using that method
Nice shirt.
I tested the Arctic Cooling MX-5 paste it is stated to have a processor life of 8 years.The result was very poor as with all Artic pastes.Cpu and Gpu were always incredibly hot mostly abnormal.I replaced the paste with Mastergel Maker and my cpu never gets over 37 degrees again at the artic circus 75 the Gpu now has a maximum temperature of 54 degrees when playing games at Artic over 85 degrees.
My experience has shown absolutely poor results for all Artic pastes.I don’t know if I’ve come across fake traps but for everyone it would be almost impossible because I ordered in different places.
I don’t know if I’ve come across fake traps but it’s almost impossible at all because I ordered on different sites.However, it is not excluded.
arctic cooling is different with arctic silver
@@tiiiooooooo
Yes, this is possible, but now I am completely satisfied with the Mastergel paste.But I still don't know which one has longer endurance.I will see this only after a while whether it will dry or not.
@@amerikanstafford mastergel pro or mastergel pro v2? mastergel pro v2 is quite cheap in my country. just usd 5.9 about
@@tiiiooooooo
The box says Mastergel Maker.Here in Europe, the price in Euros is 19.99 EU.So far, this paste has shown itself with excellent results even when winding the graphics card.
I don't know what the result will be in the future I'll see what the tests show.I don't know why but I do the Arctic Cooling MX-5 gave a very poor result.Maybe I got the fakes themselves, I have no idea.
PS: You are right.The results for (W/m.k) MasterGel Pro-8, MasterGel-1.85 and MasterGel Maker-11 are written on the box.I used the MasterGel Maker.
@@amerikanstafford thanks
I've been building PC's for 2 1/2 decades and never seen thermal paste change significantly during use. There's alot of urban myths about thermal paste circulating out there that just aren't true. (the other is the notion that you need to change your thermal paste every few years).
Depends on the paste. It certainly can dry out after a couple years. I have seen some completely dry paste. I always tell people that if the temps are fine then there is nothing to worry about.
@@TechIlliterate How does that actually happen, though, short of using mayonnaise as thermal paste? Synthetic lubricants and even petroleum oil doesn't break down or evaporate at the temperatures that CPU's typically run at.
What if it’s the opposite. When paste application is subpar the break in helps flow it. But if you did a great job applying it then no benefit.
The claim that there is a break-in period is ridiculous on its face. Clearly this is just a line of BS they can hide behind if a customer is ever unsatisfied with the performance of their product.
If each test is done with brand new thermal paste, then it means that each test is done with the same variables...
Neat results. Yeah it just depends. Some pastes require time to cure, others do not. Also, some are better for specific tasks as well even though they may not conduct heat quite as fast-if they need to be very stable for example, or are chosen for a specific viscosity. sometimes a higher viscosity paste can work better even though its thermal conducivity is a little lower because it will be more stable and it will have better contact especially if your mounting pressure is poor. Such pastes are often chosen for laptops or GPUs for example, As they will likely not be replaced by the end-user and need to be stable for a long time at high temperatures. Simply choosing the paste with the highest thermal conductivity is not necessarily a catch-all solution like many think.
well said. People looking only at numbers and thinkink "lower temperature? it's better". No. It's only made for a different purpose. It' s like comparing family car, suv, city cars and sports car. In same cases i prefer extreme performances knowing that in one year i will open the pc again, in other cases i prefer to "loose" 3-5 degrees but apply the paste once for the entire pc/laptop lifetime. In some cases i prefer more density, other cases lower density paste...the ability is to find what better suits for your needings.
I mean, those 2 or 3 degrees dont really matter... in gaming they then make a difference of 1 or 2 fps, which is, when playing with >120 fps, not noticable I would argue. Even with a bad thermal paste spread or some bubbles in between, I would say its not that big a deal
4:00 wait. so 12k, 12,1k and 122k? what? xD
It seems like the temperature rises over time not lower
Paste all its good
Just use NT-H1, NT-H2, Arctic MX-2,4, or 5 or a dozen other carbon bases pastes. AS5 is an antique by this point.
Noctua NT-H2 FTW
Yes! I’ve used Artic Silver #5 for 20 now. Even before I realised instructions I already noticed a difference over a week. This was mostly because I was heating and cooling normally between benchmarking software. It was definitely a change of temperatures from hot to cold help settle it in. Is it noticeable for every day use?probably not. 1.5-3c . Start up idles settled in more as a max load temps dropped more noticeable in benchmarks. So if you’re going to benchmark score yes definitely break it.
i am case's medulla oblongata.
should have added ambient tho. All this is meaningless without ambient being exactly the same. 20 ambient and 75c load does not mean that with 23c ambient you get 58c load. olikely it will be 80 load as the hotter the air the lower the cooling capacity. And i have left my AS5 on for 10 years now. and its still the same. my new pc uses kryonout and that lose performance after a year already.
Am I the only person who actually spreads my thermal paste over the IHS rather than just using blob in the middle? I find that my coverage is always super bad when I just let it squish out
This exact channel did a pair of videos on that a couple of months ago
best way to do break in paste u leve the cooler slight lose runn bench mark and tight it after while chip still worm worksa 100 procent
Just put the pc in the fridge if it gets too hot. Duh
The sad thing is that people would get far better results if they stopped using Thermal Paste and instead used Thermal Pasta. Honestly, it really is amazing what a difference one letter makes.
*old video, BUT...*
He ADMITS he doesn't have a control. Even if we give him ROOM TEMPERATURE as being consistent, he almost certainly doesn't have his FANS set to a constant RPM. So even if the temperature was the SAME later on, the FANS might be spinning more. So... this test is useless as far as I can tell.
Umm.... what were the room temperatures for each of the tests? There's no point saying the absolute component temperatures without saying what the room temperatures were. Also were you running at a fixed fan speed? Some tests may have had the fans running harder as the fan controller automatically adjusted their speeds to make the temperature be within the controller's idea of the ideal range. With the information you presented here, there can be no conclusions reached, as you've left so many variables unspecified.
All tests we're done with the ambient within 1 deg of 22c
He must of tested fake artic silver 5
Bruh needs some hair dye for his beard 😂 got a lil too much arctic Siler on the one side there
So it's a bold claim, and a lie. All in order to make customer less disappointed after using the (back then) overpriced and overhyped thermal paste. I actually found their ceramic version that came after AS5 a bit better paste. Of course, any average, cheap paste today is better than AS5, especially Arctic paste (any).
Good video!
No. You started the test in summer, and ended the test in autumn, using a paste meant for the arctics. I reject these results!
Short answer . No.
Man, why even do these tests with such a weak TDP processor.
Would you like test a Brazilian Thermal Paste ????
look up yout email.. tks