YOU GUYS asked me to try this... so I did
Вставка
- Опубліковано 5 жов 2024
- Does this cooling mod for the Z690 motherboards ACTUALLY work?? Let's try it!
Learn more about the X570 FTW AMD motherboard from EVGA by heading to www.evga.com
Get your JayzTwoCents Merch Here! - www.jayztwocen...
○○○○○○ Items featured in this video available at Amazon ○○○○○○
► Amazon US - bit.ly/1meybOF
► Amazon UK - amzn.to/Zx813L
► Amazon Canada - amzn.to/1tl6vc6
••• Follow me on your favorite Social Media! •••
Facebook: / jayztwocents
Twitter: / jayztwocents
Instagram: / jayztwocents
SUBSCRIBE! bit.ly/sub2Jayz...
Now this is the type of content that people do really want to watch to build up from the limited visual information in the written articles. Keep up the good work.
@@swizzler8053 he has lots of eyes on his channel and he is providing a point into the sample size and also instructions on how to potentially fix an issue. I for one applaud this type of content and if it saves me an unnecessary RMA or buying unneeded cooking parts, I’m down.
@@swizzler8053 who is he copying? Whenever there is a break in information like this, all of the tech channels cover it. Stop crying
Agreed. I love Jay's hack and slash science.
@@mikerzisu9508 In this case he should have copied. He got it all wrong, the cooler makes no difference as its the ilm.
you gotta be pretty stupid to not understand the original article
"Linusing" has officially become an adjective, ladies and gents!
🤚🏾
I think you mean verb.
for future reference the opposite of concave is convex, although saying "domed" does get the same point across. Also for everyone saying that cpu dies have been rectangular for a while now I am fairly certain he is talking about the shape of the IHS and Substrate
why do you care to comment on it?
@@nb081199 im confused on your question. I comment on videos because I have comments to make on videos
@@nb081199 there’s a comment section for a reason, no?
@@nb081199 why do youuu care to comment on this?
I think that domed is a good word and I also think that convex is not as funny as the word domed. I care to comment on your comment, Mr. Greer - or should I say, 'Dick'?
17 minute long demonstration of the kingpin's thermal performance
Honestly this channels vids are too long.
what part is kingpin? the thermal paste?
@@suyashsingh9865 yes sir.
@@patty109109 no they are not
Havent tried the washer mod. I do have the grizzley bracket and it dropped my temps insanely, 10 to 15, per core.. im running 5.5 p 4.3 e 7300cl30 , 3090 strix all under a mora 420.
I use to work for a place that used microwaves to communicate with satellites. When we lapped the parts we use a flat stone not glass. start off with 160, 200, then 300 grit wet sandpaper you can use oil too but its kind of a mess. When you sand it you go in a FIGURE EIGHT pattern. It keeps it staying flat. If you go back and forth one side WILL wear down more than the other. Good job keep the vids coming. Thanks :)
I think a key detail was missed. It's the ILM (not the cooler) that's applying too much pressure across the middle of the CPU causing the CPU to have a slight bend. When a cooler is mounted it'll make good contact with the top and bottom of the CPU as these are the highest points. But the middle of the CPU will be lower and so the gap between CPU and cooler will be greatest resulting in less effective heat transfer and higher temperatures.
wtf's an ILM
@@peter3337 the latching mechanism.
Ya i think he completely missed the point, unless I'm confused.
The cpu is higher and now rectangle and the retention cover latches on the same spots at the center of the sides of the cpu. That's ok for a square but not so much rectangle. The washers makes the Retention top higher so it doesn't push so hard on the sides of the cpu, bending it basically
@@strongislandx ppl l lol
@@peter3337 Industrial Light & Magic
😊
This mod has been life saving for me. I'm running a 12900K with a Corsair H150 Elite Capelix AIO on an Asus Z690 Formula motherboard. In Cinebench 23 the P cores averaged around 95-97C with 2 cores reaching 100C (with original AIO thermal paste). I removed the AIO and I saw that it made contact with the CPU at the top/bottom edges and in the middle it was clean (I don't think the paste even made contact in the middle). I put new paste but the temps did not change. I thought, well, bad luck, this is it. Then I saw the article from igor's Lab. I went to the hardware store and I bought 4 washers (M5 as they did not have M4 :). I was busy this week and only today I got the chance to implement the mod. Now in Cinebench 23 the highest temp is 89C with the other P cores averaging around 85C.
one thing that might have influenced the test is that the tests in the article illustrated temperature deltas of 70+ degrees meaning that the chip was prob running in the mid to high 90s in terms of degrees celsius. At that level, chips would start to run into some current leakage, making any improvements in cooling have an amplified effect.
I am using Corsair H60i RBG PRO XT on my 12600k with slight turbo overclock to 5.2 GHz, using it with the LGA1151 screws and didn't have any issues, works fine and normal with no overheating even when stressed at max frequency
It's best to lap in a figure 8 pattern. Less margin of uneven pressure error.
It is best not to vomit what others have written without experience to back it up. I never used a figure 8 pattern, and always got my heatsinks and processors to "stick."
@@jong2359 its best not to be a dick online and assume others don't have experience.
Machinist by trade here and I agree with sirjizzalot. It is best to lap in a figure 8 pattern and to periodically rotate the object being lapped as fingertips tend to apply more pressure to one corner than to others. This has a balancing out effect and creates greater precision of the of the overall lapping process. Great video btw, doing a 12th gen build at the moment using a noctua nh-u12a chromax on a 12600k
With a "Stock" LGA 1200 mount ( which looks the same as a 12th gen LGA 1700) my i7-12700KF was pegging at 100C. With the Artic Cooling LGA1700 mount ( the ONLY difference I saw was a 1mm shorter stud ). I was at 72C. So one 1mm WASHER to make the mount go UP. Makes a LGA1200 mount the same as the Artic LGA1700 kit. BUT also gives a LGA1700 "room" to tighten. You are right with the LGA1700 kit was tighten to its at zero lash. By getting the extra 1mm you get some room to "tighten". That said, I think they may have been using a LGA1200 mount on a 12th gen. You HAVE to use a LGA1700 mount on LGA1700 board, the LGA1200 FITs.. and it FEELS right. BUT ITS LOOSE.
So what I Wana get this right. I also have a arctic freezer 2 and I am using the lga 1700 mounting kit. So do I need to do this mod or not be because I think my temps are fine. Also does the kit cause bending or is it good to
I can't see how shimming the CPU socket retention would make a difference. The tension mostly comes from the CPU cooler which only makes contact with the CPU IHS and the motherboard PCB - neither of which have had their dimensions altered. The motherboard will still bow if the CPU cooler tension is too high.
which is exactly why Igor recommended a solid steel backplate like that alphacool prototype
Go watch Buildzoids vid. He explains it very well, in his own way.
Can't see another comment saying this, but I interpreted the explanation as being the actual CPU retention mechanism itself was causing the bending of the the CPU *and* the socket backplate. I have experienced this on my own 12900K install. By adding the washers you lessen the tension holding the CPU into the socket, *not* the cooler block's contact with the CPU tension.
The other workaround that's been detailed is to ensure the cooler mounting baseplate is mounted on the motherboard *before* putting the CPU in, so that it ends up reinforcing the motherboard and CPU socket, making it more resistant to being bent by the tension applied by the CPU mounting system.
I've been experiencing cooling issues with my own 12900K under full load (Cinebench R23 multicore, thermal throttles instantly but still scores >27K) and it's very evident on my board and CPU that there is quite a lot of bending taking place around the socket, resulting in poor contact by the cooler (ARCTIC Liquid Freezer II 360). The thermal paste pattern shows very poor contact on just one side of the IHS too (the side towards the PCIe slots).
I'm working with ARCTIC support to try these workarounds and see if they help at all.
mind sharing the washer thickness that you used..?
@@fist003 I've yet to try it myself but I've ordered nylon 1mm M4 washers based on what the article said. I'm going to try mounting the cooler backplate without the CPU first and compare the bending before and after the CPU is inserted.
Did you fid a solution? I am having the same experience with my 12700KF (not overclocked) in Cinebench. 2 of the cores (cores 2 and 4) sky rocketing straight to 100 celcius, and as a result thermal throttle. The E cores running between 81 and 83C, with the P cores lowest being 87C and maxing at 100C. The cooler I am using is the MSI MAG CoreLiquid 240R with LGA 1700 mounting kit. Tried reseating and reapplying thermal paste with no improved result. Short of trying a different cooler, not sure what else to try. Would expect better cooling even from a 240 AIO.
@@PeteyPete173 I've built another 12900K system (for someone else) & seen the exact same behavior. I tried the washer workaround on my own system and it made no difference. On the 2nd system I didn't see any of the socket bending but the CPU still thermal throttles instantly under Cinebench R23 multi-core. The scores are good though, as the CPU seems to immediately clock down slightly to find its equilibrium. I also built a 12700K system but didn't see this issue. That CPU maxxed at around 90 deg C under same load. My best guess so far is that this issue a combination of silicon lottery and motherboard defaults/settings for 12th gen. With the MSI board it asks for the type of CPU cooling system being used on first entry to the BIOS, and if you select water cooling it maxes power delivery out. The CPU then draws huge amounts of power once under full load until it clocks back slightly to prevent thermal throttling.
@@DamianPEdwards Interesting findings. I have a Gigabyte board and it didn't ask me what cooling I would be using. Under normal use the temps are relatively in check, haven't really put it to the test with any sort of demanding game yet. I'm not a big gamer, but it's nice to have the power there when needed. I may end up trying the suggestion Igor's Lab mentioned, mounting all bracketry before installing the CPU, as the article mentioned it could be an issue in regards to the rigidity of the board and the cooler brackets helping to stop any flex.
Awesome video! Would have been interested in the results of the lapped cpu without spacers. Keep these kind of videos coming! Can’t learn enough from your content!!
me to
I don't see the issue with my 12700k / Lian-Li Galahad 360. Even with the pump running at 50%. Then again, I, like most people, don't run Prime-95 or Cinebench all day...
this 12gen is perfect for winter. need to get one
Also don't forget a 3090ti! My 3090 aorus extreme has AIO surely transfers heat from gpu to the room creating a crack oven
Great video; however the concave IHS has nothing to do with AIO/Cooler bracket pressure, but ILM central pressure on the two central flaps of the IHS. The flat cold plate of the AIO should actually counter the concavity a little, not cause it.
When I was a teenager I loved lapping my CPU for the performance gain. I'd run my rig without thermal paste on those lap jobs. I had a jig I built to lower the cpu or cooler down evenly, and you could mount cpu coolers into it as well with some additional parts. Mirror finish on the CPU and the cooler and no thermal paste netted me better temps than i've ever had, but the amount of work isn't worth it anymore IMO, especially with the gains from overclocking being so much lower than back then. At the time we were taking 2.8-3.3ghz CPUs to clock speeds similar to what you see today. Took an e8700 to 5.2ghz in 2010, lapped mirror flat along with the water block that was running it. Same CPU ran 5ghz all day 24/7 stable as well for years before it got replaced with a 3770k if memory serves, dude was broke AF but had plenty of sanding supplies lying around since his dad did auto detailing lol
I wonder if liquid metal between the IHS and cold plate would've helped or hurt your temps.
@@pronewbofficial hard to say, the chip and cooler tended to vacuum seal together, it could get really difficult to pry the CPU off if you pressed it against the cold plate when it was out of the socket. My bet is it wouldn't have any significant effect, or would just make it worse. One of the installations I did this way actually cold vacuum welded together because of how perfectly flat it was. It was installed for a long time and at some point part of the IHS copper and the cold plate copper just fused together lol. Normally this would never happen since copper oxidizes super fast, but we think the friction of the 2 surfaces managed to disrupt enough oxide for it to be a non issue. We also basically did a finishing lap on both immediately before applying them together, so it only had seconds exposed to the air before getting installed for several years. It's a LOT of work though for very little gain at this point. Delidding the CPU and designing a cooling block that goes directly on die gives you far greater gains for less overall worki
We're also talking 150w less TDP. I had an I7-3770k that ran for 9 years @4.3GHz on stock voltage with a Corsair H100. Never got above 65C at full load. I never touched the IHS. 77w vs 125w TDP and drawing far less current.
@@dragonthc
I had a 4770 and I just stuck a tower cooler on it, nothing posh just big 'n' cheap with I think 5 or 6 pipes in a cheap case and that never went above 70C. Times have changed I suppose, Intel cranking up the electrons for performance more and more.
@@dragonthc Actually, at these OC levels you could expect 300w+ from the quad cores and up to 200-250w easily from the high end dual cores. Power levels were basically the same as today with intel's high power draw on the last few gens. Remember, we're talking about a bump of almost 2ghz and significant voltage increases from baseline. Like from 1.05-1.2v to as much as 1.45v, so more volts and more amps both (as a result of the clocks). The situation was similar on my 2600k at 4.7ghz, I had a VRM designed for 400w+ just to hit that as a daily stable OC under water. This was before intel started using thermal paste under the IHS, I basically stopped OCing around then because I wasn't comfortable with delidding at the time, and the boost speeds of the 4790k were good enough for the time. With my Q9650 at 4.7ghz daily driving, and crossfire HD5870s also overclocked, I actually killed a 1050w Enermax PSU (Revolution 85+, came highly rated by jonnyguru himself), should have gone with the 1250w model lol. This was back when I ran 6 hard drives, tons of LED strips, 2 pumps @ 30w each, and 15 fans, so even with only ~900w going to the 2 GPUs and CPU, that only left 150w of overhead for all of that. You do the math xD
I highly recommend this mod! Not only because of better temperatures, it also fixed my errorcode 55. The pesure on the socket was so high, that my RAM-slot A2 wasnt detected and caused thet Motherboard-error. I used 0,8mm plates. now everything runs fine.
17:23 looks like a Mitutoyo digital caliper was used as a straight edge but it is the old version not the ip66/67 version.
I just did this mod. My CPU clamp bracket was putting down a ton of pressure before. Lapped it flat and used kryonaut at the same time. Dropped 20c max temp during one cinebench run... Insane. 93 to 73. 12700k, 4.9 & 3.9. 1.3v. My lapping pattern midway was pretty crazy.
Do i see the star wars build and a airbrush gun in the background! Excited
Back in the x58 days i used the same mod on my xeon x5660 to keep ist below 95c at 5+ GHz.
By the way.. you have to lap the CPU while ist mounted in the socket, to have a flat ihs when the CPU is mounted. The easyest way to do that is to cut a socket out of a broken mainboard.
There’s a technique to lapping something like that. You put that glass setup on a barrel. You sand say 10 circles. Then you turn the cpu 90° clockwise while you walk 90° counter clockwise around the barrel. Repeat till you’re back where you started
Or you know, you don't and it makes almost no difference to temps.
True. But it could be useful for anyone lapping any sort of heat sink. Good thermal paste and ever more efficient cpus help mitigate the effects of them not being perfectly flat. And if you’re going to lap the cpu you should do the same to the heat sink. For many people including myself lapping the surfaces isn’t necessary. But if someone wants to give it a try this could be useful so more problems aren’t created.
Or a figure 8 pattern
@@jaceyrector9320 I've been wondering how many people actually see a temp issue with their Intel CPUs that they would need to consider doing much to them. I'm running a 10900k with a Corsair H100i, and I literally just pulled everything out of the boxes and put them in my system. I've never messed with after market thermal paste and the highest I've ever seen my CPU temps get is the high 70s, maybe 80 at a spike. My old 9590fx AMD CPU with the same model of cooler used to get 85c+ easily, and it had half the cores with the same clock speeds. I guess maybe if I was into over clocking I might need to do something, but as I sit now with the base clocks and base boosting clocks everything runs cool in my 7 year old NZXT Phantom case. I even have a 3090 sitting in the case air cooling and it peaks at about 80c when I am putting a full load on it, so the air in the case is warm and goes up through my CPU rad, but the temps still never get high.
@@xero6912 Ya my pc is “stock”. I’ve never bothered lapping a CPU. Even back in the single core days when I had an alcohol cooled monster. Back then 70° would damage a cpu over time. But on my pc now I’ve never seen a need to overclock. My computer is a tool so I can work on my other hobbies. Gaming, 3D cad, etc. I assume I’m similar to the average pc user and My pc is more than enough without doing anything to it. But for some the hobby is getting the most you can out of a pc. And if that’s the case you do anything you can to eek out a bit of performance.
haven't lapped a cpu in years, when I do my OCD kicks in and i make a whole thing out of it, just checked my 12900k and it it was as flat as can be, no light using a ruler and a precision speed square.
I sit around 63 degrees Celcius on a full load nonstop. When it's idle, I get about 25-27 degrees Celcius. I have no issues with the 12900k.
My 12900K is on the way, which cooler you got? I haven't bought one yet
Sounds like youve got a monster cooler
@@monke2361, I purchased the iCUE H170i ELITE LCD Display Liquid CPU Cooler. Very easy to install and use AIO.
@@are3287, I could have purchased the iCUE H150i ELITE and obtained temperatures around 66-69 degrees Celcius on a full load nonstop.
@@evocybernation8839 with the retrofit kit right?
It's concave and convex. Or a dip and a dome shape. It's all good. Very informative and very helpful.
I want to know why we pay the crazy prices we pay for high end chip and they are not 100 percent flat to begin with like wtf. That is something that should have been fixed by now.
I was thinking the same thing.. Would it be really that hard to mass produce the ihs at some reasonable degree of "flatness"?
its a thin stamped piece of metal, machining that well is way more expensive than youd think
@@AnticipatedHedgehog itd be a machining step, and a far more precise plating process. or they could skip the IHS entirely and sell it with a disclaimer
Guess it don't mater they ever address that kind of issue. They don't need to be flat in real life when you can't buy them anyway lol just flat in theory. By the time you get one the next gen of a not so flat chip will be out.
I am using my Noctua DH 15 air cooler using the 1200 mounting holes on my Asus Prime Z690-A and am getting great thermal readouts of 35C to 45C max so I do not have any problem with my thermals. Lucky I guess.
What about when you run an all-core Cinebench load ? What are the temps?
Mostly likely running stock or lowered vcore
@@j340_official if Cinebench is the only thing that gets the cpu hot why does it matter? If all you do for 8 hours a day is run that benchmark then Ok. But for gaming ?
@@NorthsideChamps16 that's not my point. The original person said he or she has 35C to 45C max temps at idle. That's great. At idle those temps are normal. But the original person also claims to that the lga 1200 mounting mechanism for his/her cooler has no issues. But to investigate that claim we must ask what are the temps under load? If temps are normal (75-80 C) then that claim probably is true. But if temps ramp to 100C then there could be a problem with the mounting mechanism...
You might think all you do is game so therefore you never use all cores, but sometimes windows update or some other program may temporarily need to use all cores and if temps are spiking to 100C+ that could be a problem. Especially if overclocked. So don't just use data at idle and think you're fine. Instead, I suggest to be an empiricist and test various use cases to see if your theory (that lga1200 mounting is fine) holds up under various scenarios.
I wonder if its an issue that will start to show up more after a longer time. As in CPUs will get bent more after several months under pressure... or maybe some motherboards have tighter mounts than others?
The type of backplate you are using on your aio may make a difference too
Mine is so tight that I was affraid that cpu or motherboard will break if I will lock it, thankfully nothing break but the pressure on cpu is huuugeeee.... MB is gigabate aorus z690 ud ddr4. Im thinking about that mod when i will be changing to new cooling in a few days
I think the stability of the MB and the Socket is make a huge difference from cheap to high end. Igor says that too
@@seriousplace7756 Anandtech used GIGABYTE Z690 Aorus Master and it bent like crazy :(
@@1starfish11 bro do you know any mobo which is not too tight and works well?
The z-height is *reduced* on lga1700, so old mounting kits don't have enough mounting pressure. For example, my alphacool water uses washers instead of just spring tension, like on lga1200
Exactly why I bought an Asus board. My cooler doesn't yet have a conversion kit, but the Asus mobo comes with LGA 1200 mounting holes, and my cooler uses plastic spacers instead of standoffs. So I just didn't use the spacers and was able to screw it down tight onto my 12600k.
what about just straight lapping instead of adding spacers to socket retention brackets
Yeah, not using springs for mounting is always going to cause issues with pressure. Either under or over pressure.
I wonder how much difference the kingpin thermal paste makes, if it's too much it might be worth swapping back to something more common.
The standoff height for the Corsair LGA1700 upgrade kit is 1.05mm shorter than the LGA1200 standoffs they replace so it seems more likely that the socket, or the cpus, are getting shorter, not higher.
I can confirm it worked for me, on gigabyte z690 ud and arctic freezer 34 esports duo temp drops from 84 to mosty 77, sometimes 80 but mosty is 77C. This is 7C drop!
you didnt say what cpu you had and where do you get the washers or a part number for them? thanks
@@tdb0ne1 i9-12900kf, 8mm plastic washers from hardware store
You need to do this mod BEFORE your CPU already turned concave. It's a mod to avoid it bending overtime and thus losing contact with your cooler (unless you wanna repaste every week) or having worst thermal 'cause more thermal past needed.
It's still good to do even once it turned concave, so it doesnt keep bending overtime (and thus turn temps bad again).
Either way, that's not something you can notice on new builds or builds that doesnt last more than few days.
so you're saying you need to add the 4 total washers before you even put a new 12th gen intel cpu in the socket?
@@VintageCR ideally, yes.
Idk about other manufacturers but I can confirm that Arctic liquid is sending new standoffs with a new backplate for lga 1700. They stated it in their instructions but I measured them both with a caliper and they are in fact a different size (I cant remember the size off the top of my head but I believe they’re shorter, either 13mm>12mm or 14mm>13mm).
If you find acquiring tempered glass, to lap, hard to find, try a granite or marble food chopping board. Cheap and easy to find in a supermarket.
Bought Corsair's LGA1700 scews for my 12700k and used it with an H100i Pro XT, along with Kryonaut thermal paste. Idles in the low 20s and in gaming I haven't seen above 50-55C. Many games run at 40C in 1440p.
I think the Corsair LGA 1700 Retrofit Kit may fix this for their coolers at least, they sell them on their website for like $3.
i have a h00i cappelix with the retrofit kit and my temps are max 65 even on stress test
Thanks for testing as I was wondering what the result would be. On my 12700K I have no issues and even overclocked with max temps on CBR23 multicore,I hit 80 degrees on the hottest core and that is with a 5.3GHz single core and 5GHz all core, though I am using a 360mm AIO.
Your cooler manufacturer must have taken the z height into account, just like noctua's mounting mechanism. Thus, good thermals
Perhaps you having a concave 12900K, when you apply too much pressure on it with the unmodded retention system, you actually flatten the IHS out and not make it bow, like it would do with an IHS that was flat/bowed to begin with?
seeing the Corsair One in the back without the panels on. The swap has begun!
I forgot where, but was reading different boards were causing the cpu to go from flat on its own to convex in the socket. The suggested fix was a good stiff backplate to prevent flexing/morphing around the socket, claim was it kept the cpu from doming and helped a bit with thermals (but as you said temps high because power draw high)
Right, Igor has mentioned, that boards with a strong metallic backplate - as Jayz had used - seems not beeing effected.
@5:07 I legitimately paused the video, said to my wife "Jay and Linus, seriously, lol." Then resumed to hear Jay say the same thing 😆. No hate, I just thought the whole situation was funny.
Well, you did get a 5 degree drop once you added the washers. So, the article is correct as far as dropping 5 degrees. But it still depends a lot on the configuration of your system a far as what your overall temps will be.
how much voltage did he use for the test tho ? He never said Vcore used
@@Qersoras At the beginning, he said he was keeping all the setting stock. And didn't adjust any bios setting other than to try and lock voltages and fan curves. But you are right he didn't actually mention the Vcore voltage if that was stock or adjusted.
I have a decent idea for a video, I thought about this while building my engine for my car. Copper gasket spray as thermal paste. The spray would adhere well to the cpu and could maybe work. I'm thinking not very well, but would be interesting to see. If it worked, one can would last you a lifetime. Also it's not permanent.
Great idea, would be a fun watch!
Just got to be careful that it dosent get in the pins or it will most likely short out the CPU or other parts on the motherboard. If you applied it specifically to the surface of the CPU it would be fine as long as it's not corrosive.
Try some nickel and/or copper based anti seize compound while he is at it. They look like they would make awesome "heat sink goo"
Would this be any diferent in performance to existing liquid metal products (Thermal Grizzly)?
I wonder if all the folks calling it hot are using pre applied paste on their coolers.
Although running hot may just be different to what they're used to, if you went from 35 to 45 idle temp, sure it's fine, but it would seem hot by comparison. especially if they overlooked something like power draw or voltages.
The stock paste lasts a lifetime, usually good performance for stock. It's probably ignoring proper power ratings like you said.
No applied it myself with a x shape and also tried to spread it out with some spreader thing once. Doesn't make much of a difference either way and my temps in prime 95 are easily 70c stock settings with a 12600k and one of the best aios you can get.
Overclocked it easily goes over 100c if it wasn't throttling.
And I am used to much cooler cpus(funny thing is I had a 300w cpu heavily overclocked with worse cooling that ran much cooler). But these temps are just literal fire. Even laptops are easily 30 degrees celcius cooler if not much more. The hottest cpus ever prior to these few years were easily that much cooler.
@@jc.1191 yeah it's amazing how many people talked about AMDs 5800X thermals, but overlooked the part where most boards are running default it at around 140W
admittedly i haven't had an intel chip in a long time, but I feel like that would be the first place to look if it ran hotter than I had anticipated
@@tilburg8683 yeah I tend to find as long as you get full coverage the application of paste makes little difference, I was more thinking at least with some AIOs I've seen in the past pre applied thermal paste is insufficient to fully cover the IHS, probably more noticeable if they did have a concave/convex IHS and were conservative with the amount of pressure they put on the CPU socket, but having used AMD for so long now I feel like I would just instinctively feel like most intel chips run warm by comparison even though they're fine.
LGA 1700 kinda needs 4 pressure points on the latch then, but that also requires a CPU redesign maybe... not sure. It would just be the IHS needing 4 pressure points (2 each side instead of 1).
That will be interesting to see on Intel's next socket design if they use a 4 point pressure point mounting system.
Or if later CPUs get bigger but they make it wider so it is a square again.
I'm pretty impressed with intel here really. The delta between the hottest and coolest cores are flippin great. I love new tech
So uneven core temps is normal for them? Mine sometimes makes fans increase and decrease in bursts, sounds really odd when it's doing it.
@@Nocure92 yep uneven core temps are the norm. But the main thing is that you’d like to see the difference in temperature as little as possible which is borderline impossible because it’s a never ending physics problem. Some cores may be in use more than another and that’s really all it takes to create a difference in temperature within the CPU itself.
But what your describing sounds like an issue with your fan curves. If idle it shouldn’t not be constantly jumping through bursts of fan speeds. Take a look in your bios menu and adjust fan curves in there
@@Nocure92 no lol they are not. Intel makes so many problems its not funny. I went AMD all the way after only using Intel and NVIDIA no more.
@@Correa24bori I used armory crate to improve it, but it still sometimes does it when the p cores get hot. Then it just bursts for a second or so then goes back down.
@@FenrirYT13 I mean for 12th gen. I switched from a 4770k after black friday.
I've just canceled my order of LGA 1700 parts. I can't be bothered dealing with another socket issue like I had to during the Skylake days.
I'm pretty sure my EK water block came with with washers and springs as well for 12900k
It's not about the older cooler mounting, which, if anything, should play less role in CPU deformation than on LGA11xx because on LGA1700 the CPU IHS is lower than on LGA11xx. It's about the retention bracket design and the fact that the CPU is now rectangular. The retention plate applies pressure in the middle of the CPU and the socket pins push back on the sides, which makes the CPU bend. Buildzoid explains this very well in his video.
Noctua released a socket LGA1700 conversion kit for the NH-D15 and other coolers and its not a bodge, its a backplate and hardware that is perfect fit for the 1700 socket only.
Try again with Prime95 Small FFT. Some may argue this is "unrealistic" work load, but if you want to test the cooling setup to its limit, this is the way to go.
Jay, I feel like a Physics 101 student sitting in on a symposium with Einstein presenting relativity
I felt so incredibly out of my element, I didn't understand any of the lingo haha :D
I saw an article about this on my news feed and clicked passed it figuring it was probably either wrong, unreliable or some kind of scam. Either way I have been around too long to take articles with titles like 'Use this simple trick to reduce Intel CPU's by up to 5 degrees for free' seriously.
i laugh too hard when jay said "linus-ing everything" haha
I'm no mechanical engineer but I would think scuffing it a bit and providing more surface area for the thermal compound would actually improve thermal conductivity (assuming a flat plane). Isn't that the idea behind leaving the machining pattern on the water-block?
I just ordered a Noctua 78mm bracket kit for $9 and my 12700K is idle at 20c/76c under load. Just make sure you order the right one. Noctua has 78mm and 83mm.
I'm sure you already know this by now but if anyone is curious lapping a 12th gen CPU doesn't really do anything unless you lap it while it's clamped in the mounting bracket under pressure. You can lap the CPU all you want but as soon as you mount it back in the socket the mounting points clamp down and just bend your CPU again causing the little concave in the IHS. LGA 1700 dropped the ball on that one. Washer mod helps but still not worth lapping. You would literally need to sacrifice a mother board and cut out the mounting bracket while your CPU is clamped in it and then lap it while it's mounted so it will be lapped flat while under pressure. Either that or get the thermal grizzly mounting frame and ditch the stock mounting entirely.
"I'm Linusing everything today" is my favorite PC channel reference to date.
Jay over here looking like Sabortooth Victor Creed
Wolverine would be jealous
Think you jumped the gun a little on this one Jay.
It actually has to do with the ILM pressure, not the cooling retention mechanism at all, bending the entire CPU not just the IHS. The washers are used to reduce the ILM pressure.
Your quick stress/heat test won't actually prove much in this case. The CPU used and discussed in the article had hundreds of hours of use. Which under the ILM pressure and combined heat of the CPU, has actually deformed it to a "U" shape, thus pulling the center axis away from the cooling block.
Also, the LGA1700 is thinner and reduces the Z height compared to previous sockets.
5:06
"I'm Linusing everything today... Ya, that's flat"
Well, it's the E/P core first generation. I have no doubt it has one or the other issue. But i guess it gets better with 13th and 14th gen. 🤔
Your 12900 is concave so it may be working to your advantage when using the old coolers putting more pressure on the screws may just be putting enough pressure to flatten out the 12900.
Better add another ten degrees for a closed case Jay. Do it with a cheaper cooler. We don't have $300.00 for a cooler.
Alder Lake CPUs are bending under Independent Loading Mechanism (ILM) pressure, and that is an ISSUE!
But instead of this being presented as an issue, we get this type of shilling and is spun as a positive, as a mod... impressive. The same **** Tom's Hardware did with this issue, made it a cooling mod topic, with zero statements about how this is an actual issue...
How is bending your CPU not an issue? It's intel's and MB makers fault here and they should be held accountable, not ignored or given a free pass. So many ignorant people...
Your results suggest to me the limiting factor for thermals are either the bond between the IHS and the silicon of the CPU, or the performance of that waterblock, it simply can't remove anymore heat than it was in the first test.
A lot of the talk about the 12900K being hot is coming from aircooling fans. Lots of people are uncomfortable with the idea of a CPU that NEEDS a 360mm water-coling setup in order to function properly.
This was my resolve with my 5950X. It's just standard with these higher CPUs now
"I'm Linusing everything today" Shots fired! XD
I modified the standard Lian Li Galahad 240 LGA 1200 backplate Bracket to fit the 1700 and works perfectly, the bolts already have springs and that particular AIO already comes with washers, of course, normally you would be able to just order an LGA 1700 compatible kit for this AIO from Lian Li, but Im not in the US and I can't order stuff internationally.
Would love to see a test with liquid metal on 12th gen. Had to do that and lap an 11th gen to get it under control. My OC'd 12th runs hot even with a 420mm AIO.
Yep Artic Freezer Liquid II 420mm AIO.
Crazy it still runs hot with a 420. I was surprised that mine ran hot with a 360.
Oh wow I just learned something new thank you Jay it has been a while since I have learned something new.
My 12900k runs hot on an Optimus block . 84C when running cinebench on a custom GPU and CPU with 3 360mm radiators. I am ordering an EK block to see the difference and will be keeping the better one
You can't really beat an optimus block. They are one of the best out there, much better than EK. You IHS could definitely use some lapping or you may want to try jays trick. I'm running optimus v2 and I don't go above 70C.
@@Return_To_Sender are you using it on the 12900k?
My 12900k with a NZXT Kraken 73 instantaneously hits 100c running cinebench. Gotta take this PC back to the shop!
No heating issues with mine 12700K around 60° using a Thermalright Macho Rev.B cooler
I notice you tested with the motherboard laying on a box. That doesn't seem realistic of a motherboard fitted in a case. Surely under high tension the motherboard can absorb more of the tension by flexing itself. If the motherboard had been mounted it would not be able to flex as much, putting the extra tension on the cpu.
This a good example of how a research, findings and results should be attempted and repeated by various identities to show the possible validity of the findings instead of naming other people results as wrong based on just opinions or credibility claims. Love the end when Jay says: it didn't work for me but I'm Not saying the articles is wrong
Just so people are aware technically speaking lapping should always be done in a figure 8 motion to keep consistent varying angles of pressure. I'm sure the way jay is doing it works just fine, however it's my understanding that there is a better way.
Oh, I want a t-shirt with a silluette of Jays face on @11:41 and a quote below saying "I hate when smart people are right"
Thanks for doing this for us.
Good to know.
It's a pity that it's a bit harder for me to get a history graph of the CPU temps in a terminal, in Linux without having to write the data out to like a CSV file and then post-processing the data to make the graphs/plots.
Me dying watching jay sanding the "expensive" cpu.
I actually thought this was the opposite, the coolers from last gen doesn't work as well because they don't make as good of contact because 12th gen isn't as tall so the cooler cannot make as good of contact. In fact, the LGA1700 kits you get from Cooler Master come with shorter studs, not taller....
Enjoyed your break down of the issue.
Reactor Core Temp: Unchanged
Passenger count: Unchanged
Regenerative food buffet: Unchanged
Great video, thanks Jay!
Am I being thick, but how does spacing up the retention bracket affect the clamping force of the cooler? Shouldn't it be the cooler mounting mechanism that needs spacing up?
The people are stating that it's allegedly putting too much pressure on the cpu in the mount, causing them to warp under pressure basically. By bringing the hold down up 1mm. You are reducing the pressure on the cpu in the mount. Thus, reducing the amount of hold down pressure on the cpu and making it so that it doesn't warp.
Because retention bracket gives little bit of resting area for the cooler when tighted hard, preventing bowing of the cpu/motherboard...
Yeah Me too, As the Socket is Soldered to the Motherboard, the only thing I see that mod changing is the pressure between the CPU Pads & The Socket Pins,
which is then increased when you apply pressure from the Cooler anyway, So... 🤷♂️???
If the problem is people Tightening the Cooler Mounts too Much Bending the Cooler Surface & Decreasing Surface Contact, then to me the Solution would be to add the Washers to Raise the Cooler Mounts or not tighten the Cooler Mounts so much
Because its Jay you should think twice. The cooler has nothing to do with it as those spacers just increase the pressure for that mount. Its the socket locking arm sitting too close thats causing the bending and raising it up releases pressure.
@@jepulis6674 Gotcha, it's the two central lugs on the CPU bracket pushing across the centre line of the CPU and bending it. Must be a hell of a force.
I have the i7 12700k cpu, and the specs show that this cpu runs at 190w max and 100 degrees Celsius. So I limited my bios to 180w, and the temperature improved dramatically. The wattage doesn't go over 180w anymore. I use the Gigabyte z690 Game X, which was running high wattage on auto. Now, it is controlled to 10 watts less than the max specs of my cpu.
I havent been having any temperature problems on my alder lake build either, maybe I just got lucky. I've been scratching my head alittle bit honestly.
Mine is just literal fire, and I thought ryzen 5000 was hot..(which it really is, but this is just so much worse).
@@tilburg8683 How hot does it get?
"You don't need it to be polished like a mirror"
In fact, I would argue this could make things worse, but I would have to illustrate it to explain myself.
But basically, for a given average amount and thickness of air gap, a slightly rough surface should have more surface area, improving the effectiveness of the thermal paste.
All I know is I have a 5800x and found out the BRAND NEW MSI core liquid 240 failed and was hitting 90°c now I got a Corsair 360 and I'm sitting 60°c max oh and KPX is AMAZING
As a diesel tech, I lap with figure 8 motions. Not that what jay did won't work, but to me you get a flatter surface faster 🤘
Would love to see preformance tests on a cpu that you lap from og to super polished to the extreme!
From my experience with aftermarket coolers, not just liquid coolers, their mounting system is very jank and I have ALWAYS had to jury rig it to sit properly.
I just wish my 1st gen Corsair H60se (from like 10yrs ago) could manage the temps of a 12700KF so I know if I installed it the best I could or adjust it better.
Idk man I’m sitting at 22 at idle and at 58 on tarkov with my 12900k….. if you know tarkov that game runs off carrier pigeon for its optimization
i liked how you called the opposite of concave, domed. Who knows what convex curves are anymore?
You should always lap in a figure 8 motion as this will give you the most level/even surface
This can be disputed. You as a human apply uneven pressure as you make the turns in that pattern.
If the cooler bracket is bowing the IHS then I'd think a concave housing would get flatter. Maybe this solution works better for convex housings.
For real...
The physics of how this happens are being completely ignored.
I'd bet the IHS is concave because the mounting pressure + a domed heatsink was used (common)
If the pressure is high it's not going to lift the CPU up from the sides... The motherboard is clamping it down from the sides and nothing is lifting the sides.