For TLDW people: This is NOT a recommendation for everyone to do our of the box, BZ only did it because he had a custom lapped/flat heatsink which didn't have good contact, unlike his factory curved one. This should be a backup solution if you have thermal issues that aren't fixed by reseating/reapplying TIM/tightening.
But... the curvature of cold plates is radial while the IHS curvature due to bending is rather linear/one dimensional, right? Should not be a perfect match, either.
@@japaneserequired6314 Der8auer recently pointed out that part of the IHS curvature seen in modern CPUs might be due to thermal effects during/after the manufacturing process (e.g. the contraction of solder after the IHS was mounted). This may not be by design but inherent to the product, still. I personally would not be surprised if cooler manufacturers would compensate for such effects. And BZ is among a good number of tech reviewers discussion cold plate curvature and also cross fit between Intel and AMD. See BZ's own experience with lapping the discussed water block.
I'm no expert, but if you don't do this mod, and the socket causes the CPU to bend, that's it. If you add washers afterwards, the CPU will still remain bent. If the socket lever feels very tight, I'd take that as indication that washers are necessary.
Z690 hero, brand new, no cpu ever installed, the socket back plate was actually slightly bent out of the box. Removed the socket backplate from motherboard, it was straight. That means, out of the box, without any cpu installed, the pressure is too high on the socket backplate. I think it's safe to say that the washers are necessary no matter what. Am just not sure if we should be put on 2 or all 4 holes of the socket backplate
thank you for the detailed walkthrough of the problem statement, solution, and justification. can make an informed decision as to when and how this may help
The springs don't plastically deform, and the CPU bottoms out on the plastic piece. However the springs could creep due to temperature effects. Search up the effect on springs in loaded magazines and you'll see.
Looking at some viewers' comments, it seems either the LGA pins aren't performing very well as springs in the long term and they do plastically deform (memory channels fail when a different cooler is mounted, and works again when swapped back) or it could be that the CPU substrate, IHS, motherboard, or cooler base (or any combination) has plastically deformed to cause a poor pin contact issue. Regardless whether it's spring, IHS, cooler base, or fiberglass, it explains how and when the washer mod will or will not work.
So you are saying you want a corporation to spend 50 more cents for the benefit of end user and having a good long term investment of their product instead of saving it and potentially have them buy another CPU/Motherboard in the future? how vulgar
i feel that the pin do act as a balancing for pressure distribution personally i feel like setting in the case mounting every screw could potentially help avoid that socket to over ben was quite scary and not great looking and was whit the bare bone small included cooler which i got serious doubth about the amount pressure put on the his finger crossed when get my gpu i intend to use whit my main air heatsink of 2 pound get the job done wish me good luck.At least if it work will be happy if it is broken dam will regret changing board.
Thanks for the info! Been following this since I read it on Igor's Lab. Glad to see your input! So, I'm building a new system and have an MSI Z690 Unify-X. When I pulled the ILM there was a 1mm spacer under it already from the factory. Not sure if this is normal to protect the PCB or if MSI knew of this prior.
Huh, my Asrock z690m phantom gaming 4 also has a non-metal spacer 1mm thick under the bracket. I wonder if I want to add a second spacer there, though.
Actually, just noticed there is ANOTHER pair of 1mm plastic spacers on the bracket itself, right next to the tabs that push the cpu down. They push against the already plastic edge of the socket when closed, so they aren't there to prevent scratching, but they guarantee that the tabs are 1mm higher than they would have been otherwise. This sure looks like a quick&dirty hotfix to me, just like the washer mod but right from the factory.
Received the 1700 adapter for my Galahad 360 and it mounts properly on my 12700K. Thermals are pretty good - 5.0Ghz all core is stable for everything and 5.1Ghz runs fine for CB20 or CB23. 5.4 and 5.3 GHz single/dual core is also stable.
Thank you... this is the kind of post I was looking for. I am about to build using the galahad and the 1700 adapter but am hesitant about all of these reports of bowing the mobo and cpu. Do you know if yours bowed or is bowing?
I did the washer mod on my ITX system. I observed a similar spread of thermal paste that was thicker in the centre. I went from an undervolted 12900K, pulling an average 195W peaking at 92C after 60 mins of r23, to a 12900KS with an undervolt pulling 235W average peaking at 93C after 60 mins. 6 - 10% net performance improvement, putting out 20% more wattage for the same temperatures as previous. I'm using a cooler Master Nrp200 Max, and the included asetek 280mm AIO, in ambient temps of about 24C. The washer mod works, but that is because the asetek water block was surprisingly flat, and I made this deduction because it says it supports all Intel and all AMD cpu, and they both have different curves in the IHS. I used 1mm metal washers for the mod, and torqued the ILM close to the same tension as when removed. No memory issues so far.
If anyone just wants to test this and doesn't have any thin washers on hand, Post IT Notes or just regular paper can be used to make shim washers out of. Its non-conductive and fairly thin. Paper is about .08 ish millimeters thick (about 3.5 thousandths) and Post ITs are .1 mm thick exactly. Yes it does compress a little with pressure down to maybe half its thickness but the difference in thermal paste we're seeing here is likely 2-8 thousandths of an inch (.05 - .2 mm) or 2 layers of Post ITs stacked (maybe 3 compressed). This would also allow you to try 1, 2, 3 layers and commensurate mounting pressures. Regular paper is fine too, just a tad thinner so it would maybe have less of a delta if you're testing for these temperature issues and want to dial your mounting pressure in. This is also the reason for AMD with their chiplet approach that some people see high temperatures with 5800X. Lapping or mount alterations or some combination of can also successfully address this. Because 5800X has only 1 chiplet, the support under the IHS isn't uniform when its heated and soldered on and can deform it slighty. Mine was out by about 3 thousandths of an inch (.075 mm) which lapping took care of. It does void the warranty but if you've established you have a good clocking functional chip, it should stay that way. YMMV
1700 scared the crap out of me when I opened the retention arm for the first time, and it launched the cpu out when I opened it with it in the socket. don't mess with this socket unless u have to lol
i usually make a hammer and sickle pattern with my TIM, in the hopes that all cores will have equal temperatures unfortunately core 3 needs to go to the gulag with its +10c delta
I had a very strange issue with an old 8700k just recently. I've had a D15 on it for 3 years straight. I was curious how it would perform with a liquid freezer II on it, so I swapped. After that, it would not post with all DIMM slots filled. It would only post if I had a stick in A2. All other slots were dead. When I swapped back to the D15 it was working fine again. I had no clue what was causing this but from what I can tell from this video, it could have somethign to do with the socket pins? Can a large, heavy air cooler contribute to socket pins losing contact after years of use?
@@ActuallyHardcoreOverclocking This is scary. Imagine I had an 8700K right now in my primary PC and in another generation or two I want to upgrade and put the 8700K into an HTPC with a smaller form factor, lower clocks/voltages/power all that. And it would appear to be dead just because a large heatsink was on it in an ATX tower? That’s a medium oof, and they made the IHS for 10th gen and onwards thinner? Yikes
@@GravitySandwich Yes I agree with you. The reason the pins closer to the center of the socket bend more out of shape is because the entire CPU is flexing. The IHS is what provides the most structural rigidity. The bottom line is this: Intel should bite the $0.50 bullet on the motherboard retention mechanism.
This is similar to the issue that you have in building engines. When you clamp the head onto the block you deform the bore. This 'can' lead to the piston rings not sealing as well as you might like. Thus a loss of power. The solution is to bolt a 'torque plate' to the cylinder block to mimic the force that the head applies. Then bore and hone the cylinders. This ensures that the cylinders are round when the engine is assembled. Despite the fact they will be slightly out-of-round during assembly when the head isn't bolted on. Where I'm going with this. If you want to successfully lap an LGA1700 CPU flat. For best results, it would possibly be required to sacrifice a motherboard to cut out the socket. Mount the CPU and then lap the CPU with it mounted in the socket. This will ensure a flat surface with the CPU is mounted. As you're flattening the CPU IHS after it has already been deformed. Great content. Jordan
I've always wondered why more folks haven't tried sanding down the cold plates to try and match flatness to the IHS. I hadn't considered how much deflection is caused purely by the mounting pressures. Neat!
I have a Monster Labo beast, I believe it uses EK blocks. I had this issue when I first got the PC. I changed the mech to a cheap contact frame and it solved the problem.
The really crazy thing about this is that according to this video these springs are so powerful that they can warp and maintain that warp against the enormous pressure created by the heat sink. For some reason, according to the video, only intel's mounting system has enough pressure to counteract the pressure created by the very tiny springs. What space-age material are these springs made of?
I wonder if the strangely shaped Zen 4 IHS has just been explained? That "jig saw puzzle piece" shaped IHS should have a *significant* amount of rigidity by design.
Had I not got a waterblock with an RA of 0.2um (extremely flat) I'd not mess with this, now I've ordered 1mm nylon m4 washers, and will see how it goes, rockit cooling put out their delid kit, and will soon have a copper ihs so I'll add some extra thermal paste & washers in the meantime! thank you for the overly informative video, that educated me, and helped with understanding the purchase decisions I'll make accordingly.
Yes this video is defiantly longer than it needed to be, BUT this is the only video I have seen that actually explained the reasoning behind the washer mod!
Jay is almost never explaining well... :D In my opinion he's just a clicky baity hyped tuber. Unfortunately I always get a notification about his new videos...
Seems to me the mounting pressure of the cooler should ideally be higher than that of the loading mechanism in order to assure good IHS contact. One could - via washers - reduce the role of the loading mechanism to a mere mounting aid holding the CPU in place.
There is *another* advantage of the "washer mod", when using a block designed for LGA15xx/LGA20xx: because the Z height of LGA1700 is *less* than that of these prior packages, you may not get the same level of contact between the cooler coldplate and the IHS when the ILM is "pulling down" the CPU into docket with full force. So, by *reducing* the amount of "pulldown" into the socket, the CPU sits (slightly) higher, and therefore, makes better contact with a cooler/block designed for a thicker CPU.
Not only lapping with remove LM residue (which I don't like, you should use fresh LM, liquify it, wipe it and use steel wool to keep the nickel stain) but having the coldplate AND the IHS as flat as possible is better for LM since you need as good contact as possible since you use very thin layers of LM. I haven't found many coldplates that aren't flat (they do a good job, at least on AIOs) but my 1800X and my 3900X had both very badly curved IHSs (I guess happens on the soldering phase)
It seems that some boards come with this mod from the factory now. My Asrock z690m phantom gaming 4 has 1mm plastic spacers between the braket and the board and ALSO on the bracket right next to the tabs that push the cpu down. The latter push against the edge of the socket when closed, guarantee that the tabs are 1mm higher than they would have been otherwise, and honestly look more like a hacky hotfix than intentional design.
@@ActuallyHardcoreOverclocking my gigabyte arous elite ax also comes with plastic washers it seems. I tested the mounted CPU with my credit card and it seemed to be very flat as well
Just opened a MSI Z690 Force wifi and I was ready to install thermalrite bracket but there was a black rubber material beneath the ILM already. Seems not needed at all so far.
Lithography engineers: We need more money for... Intel: Say no more. Here's $10 billion. Loading mechanism engineers: We're having QA issues with... Intel: GET OUT OF MY OFFICE!
They blew their budget on super tiny springs that are powerful enough to bend metal and maintain that bend against the enormous pressures of a well-mounted heat sink. When all they needed was springs strong enough to make good contact. Do you really think they deserve more money to waste... Seriously, these heat sinks have enormous downward force, and according to the video, the springs are more powerful.
This is interesting. Wondering if this could end up being a widespread issue for the 1700 socket clamping mechanism over time. Seems to me they perhaps need to redesign the mechanism to spread out the pressure wider lengthwise on the CPU vs too much in the center. AMD/AM5 take note!
Amd doesnt have to take notes they already have a rock solid socket for Threadripper, i’m pretty sure they got it covered for Zen 4 unlike these idiots at Intel :D
The principle applies when direct-die cooling, where the IHS and ILM are both dispensed with. The CPU is pressed into the socket by the cooler plate _only_ via the die and the result will be great cooling but often a warped CPU PCB (up at the edges/corners) which can lead to inconsistent pin contact. Hence the often reported memory channel issues and such, but I'd also be concerned with the VCC (power) areas - they can be transmitting well over 100 amps (in fact over 200 amps in some overclocked systems of recent generations) and I wouldn't want any of the VCC pins barely tickling the lands. ETA >> the solution for DD is a shim around the die to distribute the force from the cooler plate, better than 'direct die frames' IMO.
I got a 12700k on sale and this makes me really concerned about the long term usage of alder lake. I keep my computers for years and seeing some of the images of bent alder lake CPUs is scary.
I think if the flatness of thhe cpu is as sensitive as you say, you ought to use arbor shims instead of cheap stamped flat washers. The shims will give you much better thickness tollerance whereas the plain washers can vary in thickness quite a bit.
From the pictures I’ve seen though the mounting pressure of the loading mechanism seems to sometimes bend the whole motherboard PCB and not just the integrated heat spreader.. especially on low end boards that aren’t as stiff
You made my day, by doing this and sanding 12900K ihs ( it was soooo concave ) got temps down by 15-20 C ( !!! ) using Optimus Foundation waterblock. Now 300 W is not a problem, CB23 30280 at heathy 92 C :)
Totally agree the loading mechanism feels overly tight. First time putting in a CPU, had it in the correct orientation but still took it out again to check because it felt the lever would bend from the strain. Turns out that is ‘normal’.
If you're using stamped washers give them a quick sanding to take the edge off. OR use a non metallic washer made of delrin or hard fiber. Not something too soft like nylon or plastic, where it could squish.
I've watched that video, and I agree with Buildzoid too. It seems that Jay's configuration maxed out the AIO radiator relative to the CPU power. Notice how after the washer mod was applied, there is that advertised temperature drop before the radiator becomes saturated. This mod makes way more sense when the builder is encountering a situation where CPU heat is still an issue, even though the cooler is overkill for the power demand. An example is if a cooling loop is relatively cold, despite a CPU overheating.
But jayztwocents claimed it had to do with mounting pressure of the cooler being too much. Adding those spacers will in theory increase the coolers mounting pressure as now the cpu is not held as tightly against the socket without it. And if indeed the socket is bending the cpus, you should check for curvature while the cpu is installed. I would guess the cpu wont stay as bent when pressure is not applied, much like evga shills :D
I did this mod on my TUF D4 board and it made the CPU not make contact with some pins, as aparently, I lost the two innermost dimm slots. Going back to stock solve the issue. Just making people aware than contact issues may arise when performing this mod, so don't panic if it happens to you.
I just did this on my Tuf D4 and did not loose any of my 4 dimm slots, i used .75mm plastic washers (thats just what i had laying around), this mod improved my temps at load by 3.5c.
Bro, you can also try to bend the tab on top of the socket with pliers a little, so you can also reduce the clamping force in the middle of the processor.
I can confirm that, done that, 0.8 plastic washer, temperature drops from 84 to mostly 77, sometimes 80, but mosty is 77C on artcic freezer 34 esports duo, incredibile!!!
I put PTFE washers (4x1mm, measured precisely). Hard to say it's a coincidence or not: cpu temp 70 at high load, benchmark max 88 (magic treshold lol it never surpasses). I'm happy with the results considering I have just aio not custom cooling, my room temp is high all year round and the gpu ocd.
eso de las arandelas es un riesgo, se puede dañar el pcb con el filo de corte de las mismas, lo mejor es ocupar el bracket fabricado por thermalright o grizzly que ya están diseñados y mecanizados con las medidas necesarias para que apliquen la presión uniforme en todo el borde del IHS del procesador sin dañar nada
I recently came across this mod, I got a Supercool Direct die installed, in case you don’t know it , it’s a direct die waterblock that uses an alternative IHS as the contact plate to the die but on its other side it acts as a waterblock since Coolant passes on it. Would that need the washer mod ?
I'm seeing 77 degrees on a 12700k and a Noctua d15 stock settings with Cenebench r23. I think this comes down to manufacturing tolerances with the cpu bracket. I don't think I would achieve anything doing this mod. AIO users maybe.
Great now I have a reason to skip an entire socket for enthusiast level builds. Budget is the way with alder lake and anything else they release on lga1700
For anyone wanting info on the LGA 1700 socket: Specification Intel socket LGA 1700: IHS to MB Height (Z-Stack, validated range) 6.529 - 7,532 mm Socket Seating Plane Height 2.7 mm Thermal Solution Hole Pattern 78 x 78 mm Maximum Thermal Solution Center of Gravity Height from IHS 25.4 mm Static Total Compressive Minimum 534N (120 lbf), Beginning of Life 356 N (80 lbf) End of life maximum 1068 N (240 lbf) Socket Loading 80-240 lbf Maximum Thermal Solution Mass 950 gram
I just realized my old delidded 3570K might not be dead (no visible damage). It may have lost contact to some pins after shaving off the height. Would you say that Zen4 is going to have the same issue as you just described because the AM5 LGA loading mechanism pushes only in the middle exactly like LGA 1700? Or is it too short? Also the Zen4 IHS has cutouts which will probably allow for more curvature. How about putting something (washers?) between the IHS and the loading mechanism?
i have the same waterblock. have you checked if the coldplate is 100% flat (i didn't before mounting, unfortunately)? what are your experiences in terms of temperature delta to water? how are your temps @stock CB23 in relation to water? i'm peaking at 79 °C cpu package temp while water is at around 27 °C => Delta Water / CPU = 51°C. Delta coolest / hottest core = 67 °C vs 79 °C = 12 °C.
Okay so what if we put the cpu inside a "fake" loading mechanism while laping, to simulate the pressure of when it is actually inside a socket. In my head the cpu would now be more or less flat when it is inside a socket and be curved when it is outside of one. Idk if this would help, if any, but could be fun trying it out.
What about threadripper with all it's pins, does it bend too ? hmmmm... I think I have seen a very sturdy frame with multiple pressure points... maybe not eh ? hmmm...
I didn't even get that far. After no posting with the first cpu I tried, I took the cpu out of the MSI Z690 Edge wifi DDR4 mobo and saw that the top left corner of the substrate was bent up towards the IHS. I'm guessing since the substrate was bent it wasn't making contact with the pins, so I ordered a second 12700k and a different mobo and I got a noctua air cooler instead of the NZXT X72 that I was using. Still waiting on the air cooler to be delivered so I can eliminate what parts are causing this. Very frustrating to go through all of this. At this point it's almost worth not upgrading. Any tips to make sure I can get a post and maybe avoid bending the freakin substrate would be appreciated.
@@ActuallyHardcoreOverclocking I see, yea I saw that headline earlier and watched Jay's video first out curiosity. Made no sense to me to be placing washers on a CPU retention bracket if the issue is some coolers applying too much mounting pressure, or at least the impression I got from his video. Your video cleared that up with something that makes sense.
12:00 and thus lapping your CPU (if you like to void your warranty) is hereby debunked for normal/gaming use cases. because if you still lap the CPU, you might as well void your CPU Cooler warranty as well and lap it since that wasn't straight to begin with.. The ONLY effect lapping has or CAN have is better temps when you use LN2 paired with hardcore overclocking which.. again, in normal/gaming use cases is NOT the case. READ ON: the washer MOD everyone is hyped about, MAY or may NOT work for you.. but you don't void warranties by adding/removing the washers.. However, the actual trick to this MOD is to do this AT THE START -when you actually bought a fresh and new system/parts and still need to build it.. if you do the washer mod AFTER the fact, the chances are that the CPU is already bend and wont bend back. so.. your luck may vary. If you DID do this mod from the start and your temps are.. off the charts high say 80+ when stressing, remove the washers.. add thinner washers.. and test again. test until you either are satisfied with the temps and/or have no washers installed (so no mod) which means the waterblock/cooling was already matched with the IHS.
Older generations produce some heat and I've got the disassembly figured out by now. A lot of patience is needed. The issue is to find the right TIM. If to thin it often won't make any contact with the heat spreader.
nice vid, and there is 7700k , delided, that only works on water cooling block, beacause if you tighten it too much, it refuse to boot with both mem sticks :) i applied silicon under IHS like factory did, but, didnt help. well, its water coooled, screws are loose, buuut, it works, if you touch it (repaste), you gonna curse a lot. 4,8 with normal voltage
I use an nh-d15 which is supposed to have a curved cold plate and I still see spread like you show. Hit about 90 in cinebench and 60s to low 70s while gaming. I'm sure I'd have gains from doing this but not sure it's worth it for my use case.
Make sure the fan speed is able to hit max. (1400-1500RPM) If you use the lownoise adapter it will be 200-300 RPM lower and that could be the reason for the 90 degrees. 90 is still ok though, i think 94-96 is the throttle temp.
This is very good info thanks bz, I didn't knew the LGA1700 is going to have this much pressure than the previous sockets. Well I guess I have to find washers(hope the plastic kinds don't crack) for my already lapped raystorm waterblock and if I'm going to have somebody asking me to help them build their new pc with this LGA. Also quick question for my clutch electric screwdriver: how much torque(N•m or ft•lbs) is it supposed to have to safely rescrew the loading mechanism back?
@@ActuallyHardcoreOverclocking oh thanks bz! I guess I have set the clutch at around 0.3 - 1Nm(0.2 - 0.73ft.lbs) hopefully the lowest torque on the screws could hold it. Also my driver is the makita DF012DSE its an awesome pen driver around computers.
Make sure you're accounting for the lower z-height of LGA1700 vs earlier sockets, presumably with washers adding to the spring pressure of the Raystorm mount. Nylon washers won't crack, but may squish over time. Never use an electric screwdriver without having a torque spec. It saves like 10s over hand tools. Don't use a wrench, either.
@@concinnus I know about the lga1700 z-height situation it's around 0.717-0.783mm lower than lga1200, that's why this washer mod is an option just in case if I upgrade my brother watercooling PC with the lapped raystorm cuz those new lga1700 spring pins are quite tough & I don't mind the issues of this mod to made it fit. I don't have 12th gen, I'm still on 10th gen & ryzen zen3 with a itx. When i was replacing the tubes for my bro's PC, I noticed the raystorm mount springs with the nylon washers was squeezing the solder mount of my bro's z490 godlike marking rings on them, thankfully nothing bad happened and I loosen them a bit after that with nearly the same temps. The Makita driver I mention in my comment before does have a auto stop clutch and torque specs in the manual which is amazing cuz this tool is very gentle on sensitive electronics like motherboards and tough enough to drill fan screws and other stuff, big time & life saver around computers, but it need a set of driver bits like Hexs & other Phillips bits especially the taller bits for tall cpu coolers.
@@emini6 To be clear, the washer mod Buildzoid is talking about is to raise the z-height of the retention bracket, whereas the washers I was talking about were to lower the z-height of the waterblock.
I am kinda wondering what’s the deal on the itx boards? For instance I have a z690i aorus ultra itx ddr4, I’m using a heatkiller iv waterblock, my temps are not outrageous, but I am sure it can’t be better. I’m using the update 1700 backplate, springs and washers from heatkiller as well. Idk if my thinking is correct, but the smaller itx size would mean the board has less flex, also my waterblock it’s got a really good mounting system, it’s almost impossible to get a bad mount on heatkiller iv block. Can anyone chime in on this? Forgot to mention the z690i aorus has a robust amount of solid heatsinks and a full aluminum backplate so, I’m not sure if this mod would work for me, I think my mount might be as good as it will get.
So for AMD cpu with flat (either factory or sanded) heat-spreader the optimal water-block should actually be also flat then because AMD doesn't use LGA socket. Correct me if i am wrong please.
For TLDW people: This is NOT a recommendation for everyone to do our of the box, BZ only did it because he had a custom lapped/flat heatsink which didn't have good contact, unlike his factory curved one. This should be a backup solution if you have thermal issues that aren't fixed by reseating/reapplying TIM/tightening.
@
aceofspades - Thanks for additional topic coverage...
But... the curvature of cold plates is radial while the IHS curvature due to bending is rather linear/one dimensional, right? Should not be a perfect match, either.
@@japaneserequired6314 Der8auer recently pointed out that part of the IHS curvature seen in modern CPUs might be due to thermal effects during/after the manufacturing process (e.g. the contraction of solder after the IHS was mounted). This may not be by design but inherent to the product, still.
I personally would not be surprised if cooler manufacturers would compensate for such effects. And BZ is among a good number of tech reviewers discussion cold plate curvature and also cross fit between Intel and AMD. See BZ's own experience with lapping the discussed water block.
I'm no expert, but if you don't do this mod, and the socket causes the CPU to bend, that's it. If you add washers afterwards, the CPU will still remain bent. If the socket lever feels very tight, I'd take that as indication that washers are necessary.
Z690 hero, brand new, no cpu ever installed, the socket back plate was actually slightly bent out of the box. Removed the socket backplate from motherboard, it was straight. That means, out of the box, without any cpu installed, the pressure is too high on the socket backplate. I think it's safe to say that the washers are necessary no matter what. Am just not sure if we should be put on 2 or all 4 holes of the socket backplate
"This video is longer than it needed to be" -- title of the upcoming Buildzoid autobiography :D
Subtitle: But it has a postcode!
😁
Would buy it
every buildzoid video ever
Had my CPU in my board since launch and it worked for me. Arctic LF II and Gigabyte Aorus Pro. -5°C average in CB23.
thank you for the detailed walkthrough of the problem statement, solution, and justification. can make an informed decision as to when and how this may help
The springs don't plastically deform, and the CPU bottoms out on the plastic piece. However the springs could creep due to temperature effects.
Search up the effect on springs in loaded magazines and you'll see.
Looking at some viewers' comments, it seems either the LGA pins aren't performing very well as springs in the long term and they do plastically deform (memory channels fail when a different cooler is mounted, and works again when swapped back) or it could be that the CPU substrate, IHS, motherboard, or cooler base (or any combination) has plastically deformed to cause a poor pin contact issue. Regardless whether it's spring, IHS, cooler base, or fiberglass, it explains how and when the washer mod will or will not work.
They should have added additional holding lip at top and bottom on the IHS and loading mechanism to even the distribution pressure.
So you are saying you want a corporation to spend 50 more cents for the benefit of end user and having a good long term investment of their product instead of saving it and potentially have them buy another CPU/Motherboard in the future? how vulgar
i feel that the pin do act as a balancing for pressure distribution personally i feel like setting in the case mounting every screw could potentially help avoid that socket to over ben was quite scary and not great looking and was whit the bare bone small included cooler which i got serious doubth about the amount pressure put on the his finger crossed when get my gpu i intend to use whit my main air heatsink of 2 pound get the job done wish me good luck.At least if it work will be happy if it is broken dam will regret changing board.
The IHS is longer now, right? So one additional issue is that the same curvature to the IHS results in a larger z gap at the center.
I guess this is the real cause of the issue, because the mounting pressure is exactly the same as last gen LGA1200.
Thanks for the info! Been following this since I read it on Igor's Lab. Glad to see your input! So, I'm building a new system and have an MSI Z690 Unify-X. When I pulled the ILM there was a 1mm spacer under it already from the factory. Not sure if this is normal to protect the PCB or if MSI knew of this prior.
Huh, my Asrock z690m phantom gaming 4 also has a non-metal spacer 1mm thick under the bracket. I wonder if I want to add a second spacer there, though.
Actually, just noticed there is ANOTHER pair of 1mm plastic spacers on the bracket itself, right next to the tabs that push the cpu down. They push against the already plastic edge of the socket when closed, so they aren't there to prevent scratching, but they guarantee that the tabs are 1mm higher than they would have been otherwise. This sure looks like a quick&dirty hotfix to me, just like the washer mod but right from the factory.
Received the 1700 adapter for my Galahad 360 and it mounts properly on my 12700K. Thermals are pretty good - 5.0Ghz all core is stable for everything and 5.1Ghz runs fine for CB20 or CB23. 5.4 and 5.3 GHz single/dual core is also stable.
Thank you... this is the kind of post I was looking for. I am about to build using the galahad and the 1700 adapter but am hesitant about all of these reports of bowing the mobo and cpu. Do you know if yours bowed or is bowing?
@@mikerzisu9508 I've only mounted the heat sink once and not taken it off so I don't know if there is any bowing.
I would suggest finding washers made from non-conductive material.
Nylon would be good. Machined aluminum also good and guaranteed not to yield (creep) over time.
@@MrKentaroMotoPI do metallic non-stamped washers exist?
@@salmiakki5638 Yes, there are machined aluminum, stainless, and plated steel. I would avoid the plated as they may have tiny burrs.
I found 0.5-1mm washers made from PCB
You can get washers of pretty much any size off Aliexpress.
I did the washer mod on my ITX system. I observed a similar spread of thermal paste that was thicker in the centre.
I went from an undervolted 12900K, pulling an average 195W peaking at 92C after 60 mins of r23, to a 12900KS with an undervolt pulling 235W average peaking at 93C after 60 mins.
6 - 10% net performance improvement, putting out 20% more wattage for the same temperatures as previous.
I'm using a cooler Master Nrp200 Max, and the included asetek 280mm AIO, in ambient temps of about 24C.
The washer mod works, but that is because the asetek water block was surprisingly flat, and I made this deduction because it says it supports all Intel and all AMD cpu, and they both have different curves in the IHS.
I used 1mm metal washers for the mod, and torqued the ILM close to the same tension as when removed. No memory issues so far.
This video helped me understand what could make the difference in cooling with LGA 1700
Thank you !
If anyone just wants to test this and doesn't have any thin washers on hand, Post IT Notes or just regular paper can be used to make shim washers out of. Its non-conductive and fairly thin.
Paper is about .08 ish millimeters thick (about 3.5 thousandths) and Post ITs are .1 mm thick exactly. Yes it does compress a little with pressure down to maybe half its thickness but the difference in thermal paste we're seeing here is likely 2-8 thousandths of an inch (.05 - .2 mm) or 2 layers of Post ITs stacked (maybe 3 compressed). This would also allow you to try 1, 2, 3 layers and commensurate mounting pressures.
Regular paper is fine too, just a tad thinner so it would maybe have less of a delta if you're testing for these temperature issues and want to dial your mounting pressure in.
This is also the reason for AMD with their chiplet approach that some people see high temperatures with 5800X. Lapping or mount alterations or some combination of can also successfully address this. Because 5800X has only 1 chiplet, the support under the IHS isn't uniform when its heated and soldered on and can deform it slighty. Mine was out by about 3 thousandths of an inch (.075 mm) which lapping took care of. It does void the warranty but if you've established you have a good clocking functional chip, it should stay that way. YMMV
1700 scared the crap out of me when I opened the retention arm for the first time, and it launched the cpu out when I opened it with it in the socket. don't mess with this socket unless u have to lol
its pretty safe just take the cpu out first lmao
Thanks BZ. Well put together here.
Good clarification video on the topic!
Thank you mr buildzoid, perfect video.
i usually make a hammer and sickle pattern with my TIM, in the hopes that all cores will have equal temperatures
unfortunately core 3 needs to go to the gulag with its +10c delta
He's whipped! He loves that board so much he put a ring on it! Several infact 🤣
I had a very strange issue with an old 8700k just recently. I've had a D15 on it for 3 years straight. I was curious how it would perform with a liquid freezer II on it, so I swapped. After that, it would not post with all DIMM slots filled. It would only post if I had a stick in A2. All other slots were dead. When I swapped back to the D15 it was working fine again.
I had no clue what was causing this but from what I can tell from this video, it could have somethign to do with the socket pins? Can a large, heavy air cooler contribute to socket pins losing contact after years of use?
if it's leaning it might've flattened one side of the socket.
@@ActuallyHardcoreOverclocking This is scary. Imagine I had an 8700K right now in my primary PC and in another generation or two I want to upgrade and put the 8700K into an HTPC with a smaller form factor, lower clocks/voltages/power all that. And it would appear to be dead just because a large heatsink was on it in an ATX tower? That’s a medium oof, and they made the IHS for 10th gen and onwards thinner? Yikes
@@depth386 He is talking about the socket pins. Not the IHS nor CPU thickness in this situation.
@@GravitySandwich Yes I agree with you. The reason the pins closer to the center of the socket bend more out of shape is because the entire CPU is flexing. The IHS is what provides the most structural rigidity. The bottom line is this: Intel should bite the $0.50 bullet on the motherboard retention mechanism.
@@depth386 Unfortunately the consumers will end up eating that cost anyway.
This is similar to the issue that you have in building engines. When you clamp the head onto the block you deform the bore. This 'can' lead to the piston rings not sealing as well as you might like. Thus a loss of power. The solution is to bolt a 'torque plate' to the cylinder block to mimic the force that the head applies. Then bore and hone the cylinders. This ensures that the cylinders are round when the engine is assembled. Despite the fact they will be slightly out-of-round during assembly when the head isn't bolted on.
Where I'm going with this. If you want to successfully lap an LGA1700 CPU flat. For best results, it would possibly be required to sacrifice a motherboard to cut out the socket. Mount the CPU and then lap the CPU with it mounted in the socket. This will ensure a flat surface with the CPU is mounted. As you're flattening the CPU IHS after it has already been deformed.
Great content.
Jordan
I've always wondered why more folks haven't tried sanding down the cold plates to try and match flatness to the IHS. I hadn't considered how much deflection is caused purely by the mounting pressures. Neat!
Modding IHS and heatsinks by lapping them tonmirror finish have been done since ancient times. IHS aren't perfectly flat either.
I have a Monster Labo beast, I believe it uses EK blocks. I had this issue when I first got the PC. I changed the mech to a cheap contact frame and it solved the problem.
For me, I tightened down the EK velocity 2 with the washer mod, and everything worked out perfectly. I lost 7 degrees Celsius.
I am about to install my EK velocity2 on my 12900k and am looking for this exact information.
Lost as in lowered temps or increased?
The really crazy thing about this is that according to this video these springs are so powerful that they can warp and maintain that warp against the enormous pressure created by the heat sink. For some reason, according to the video, only intel's mounting system has enough pressure to counteract the pressure created by the very tiny springs. What space-age material are these springs made of?
Yes, this video is absolute nonsense.
I wonder if the strangely shaped Zen 4 IHS has just been explained? That "jig saw puzzle piece" shaped IHS should have a *significant* amount of rigidity by design.
isnt it coz they wanted to make am4 coolers compatible ?
Had I not got a waterblock with an RA of 0.2um (extremely flat) I'd not mess with this, now I've ordered 1mm nylon m4 washers, and will see how it goes, rockit cooling put out their delid kit, and will soon have a copper ihs so I'll add some extra thermal paste & washers in the meantime! thank you for the overly informative video, that educated me, and helped with understanding the purchase decisions I'll make accordingly.
Yes this video is defiantly longer than it needed to be, BUT this is the only video I have seen that actually explained the reasoning behind the washer mod!
This is much better explained than Jay's video
Jay is almost never explaining well... :D
In my opinion he's just a clicky baity hyped tuber.
Unfortunately I always get a notification about his new videos...
Seems to me the mounting pressure of the cooler should ideally be higher than that of the loading mechanism in order to assure good IHS contact. One could - via washers - reduce the role of the loading mechanism to a mere mounting aid holding the CPU in place.
As always a very informative and detailed video.
If you have a ek velocity2 Waterblock for lga1700 and torqued properly to 0.6nm, temps are the exact same.
isn't that block specifically designed for LGA1700? EK probably designed it to compensate for the curve.
Thank you
did the washer mod
Z690 msi edge + artic freezer II 360mm
Newest member of the -5°C club
I've bought the same mainboard and a 12700K. Do you have a "Lotes" Socket? I have one but I didn't measure the ILM yet, nor the 'used' 12700K yet...
Thanks, this helped me to wrap my head around washer mods.
There is *another* advantage of the "washer mod", when using a block designed for LGA15xx/LGA20xx: because the Z height of LGA1700 is *less* than that of these prior packages, you may not get the same level of contact between the cooler coldplate and the IHS when the ILM is "pulling down" the CPU into docket with full force. So, by *reducing* the amount of "pulldown" into the socket, the CPU sits (slightly) higher, and therefore, makes better contact with a cooler/block designed for a thicker CPU.
Is this theory or can you confirm this with your rig??
Not only lapping with remove LM residue (which I don't like, you should use fresh LM, liquify it, wipe it and use steel wool to keep the nickel stain) but having the coldplate AND the IHS as flat as possible is better for LM since you need as good contact as possible since you use very thin layers of LM. I haven't found many coldplates that aren't flat (they do a good job, at least on AIOs) but my 1800X and my 3900X had both very badly curved IHSs (I guess happens on the soldering phase)
commenting for the algorithm. I love this video!
Great video, fully explanatory. I'll be mounting a 12700k on Asus strix D4 with Corsair AIO 150 LCD, Let's see how it goes.
It seems that some boards come with this mod from the factory now. My Asrock z690m phantom gaming 4 has 1mm plastic spacers between the braket and the board and ALSO on the bracket right next to the tabs that push the cpu down. The latter push against the edge of the socket when closed, guarantee that the tabs are 1mm higher than they would have been otherwise, and honestly look more like a hacky hotfix than intentional design.
Do you have a picture of this? because I'd love to see it. My email is buildzoid@gmail.com
@@ActuallyHardcoreOverclocking my gigabyte arous elite ax also comes with plastic washers it seems. I tested the mounted CPU with my credit card and it seemed to be very flat as well
Just opened a MSI Z690 Force wifi and I was ready to install thermalrite bracket but there was a black rubber material beneath the ILM already. Seems not needed at all so far.
the bracket should be designed to apply pressure all the way around and the substrate should bottom out at the ideal spring load.
thank you for your videos bz, you are the best!
Lithography engineers: We need more money for...
Intel: Say no more. Here's $10 billion.
Loading mechanism engineers: We're having QA issues with...
Intel: GET OUT OF MY OFFICE!
To Intel, as long as it doesn't thermal throttle at PL1 whether it runs at 50 degree or 90 degree matters little to them.
They blew their budget on super tiny springs that are powerful enough to bend metal and maintain that bend against the enormous pressures of a well-mounted heat sink. When all they needed was springs strong enough to make good contact. Do you really think they deserve more money to waste...
Seriously, these heat sinks have enormous downward force, and according to the video, the springs are more powerful.
Do you own a 12900k ? I've had one since release. I haven't experienced any problems or out of control tempatures
This is interesting. Wondering if this could end up being a widespread issue for the 1700 socket clamping mechanism over time. Seems to me they perhaps need to redesign the mechanism to spread out the pressure wider lengthwise on the CPU vs too much in the center.
AMD/AM5 take note!
Amd doesnt have to take notes they already have a rock solid socket for Threadripper, i’m pretty sure they got it covered for Zen 4 unlike these idiots at Intel :D
The principle applies when direct-die cooling, where the IHS and ILM are both dispensed with. The CPU is pressed into the socket by the cooler plate _only_ via the die and the result will be great cooling but often a warped CPU PCB (up at the edges/corners) which can lead to inconsistent pin contact. Hence the often reported memory channel issues and such, but I'd also be concerned with the VCC (power) areas - they can be transmitting well over 100 amps (in fact over 200 amps in some overclocked systems of recent generations) and I wouldn't want any of the VCC pins barely tickling the lands. ETA >> the solution for DD is a shim around the die to distribute the force from the cooler plate, better than 'direct die frames' IMO.
I got a 12700k on sale and this makes me really concerned about the long term usage of alder lake. I keep my computers for years and seeing some of the images of bent alder lake CPUs is scary.
Same! The bent IHS and substrate scare me too. I haven't installed my 12700K, and I'm wondering if I should return it for a 5800X or 5900X.
That's why when I'm sanding the IHS I'm also sanding the cooler base.
Basic statics with Buildzoid. I like.
Very good explanation!! Thanks
I think if the flatness of thhe cpu is as sensitive as you say, you ought to use arbor shims instead of cheap stamped flat washers. The shims will give you much better thickness tollerance whereas the plain washers can vary in thickness quite a bit.
It doesn't really matter. All these washers do is lessen the pressure onto the CPU.
From the pictures I’ve seen though the mounting pressure of the loading mechanism seems to sometimes bend the whole motherboard PCB and not just the integrated heat spreader.. especially on low end boards that aren’t as stiff
Wouldn't a MOBO backplate help with that?
@@BAJF93 yes, that’s why it’s less of an issue with enthusiast boards that have metal backplates
My asus apex is deformed around the socket area, I can see it from the top of the board, perhaps for this reason
You made my day, by doing this and sanding 12900K ihs ( it was soooo concave ) got temps down by 15-20 C ( !!! ) using Optimus Foundation waterblock. Now 300 W is not a problem, CB23 30280 at heathy 92 C :)
Totally agree the loading mechanism feels overly tight. First time putting in a CPU, had it in the correct orientation but still took it out again to check because it felt the lever would bend from the strain. Turns out that is ‘normal’.
Do you think long term all the bending could damage either the motherboard or CPU?
Wonder if it could damage the socket itself, though?
If you're using stamped washers give them a quick sanding to take the edge off. OR use a non metallic washer made of delrin or hard fiber. Not something too soft like nylon or plastic, where it could squish.
This seems to match what jayztwocents got when he tested the mod, where there wasn't much of a difference between pre-mod and post-mod.
yeah and he tested with an AIO.
I've watched that video, and I agree with Buildzoid too. It seems that Jay's configuration maxed out the AIO radiator relative to the CPU power. Notice how after the washer mod was applied, there is that advertised temperature drop before the radiator becomes saturated.
This mod makes way more sense when the builder is encountering a situation where CPU heat is still an issue, even though the cooler is overkill for the power demand. An example is if a cooling loop is relatively cold, despite a CPU overheating.
But jayztwocents claimed it had to do with mounting pressure of the cooler being too much. Adding those spacers will in theory increase the coolers mounting pressure as now the cpu is not held as tightly against the socket without it.
And if indeed the socket is bending the cpus, you should check for curvature while the cpu is installed. I would guess the cpu wont stay as bent when pressure is not applied, much like evga shills :D
Jay is great, but I think he missed the point of the other washer mod entirely.
I did this mod on my TUF D4 board and it made the CPU not make contact with some pins, as aparently, I lost the two innermost dimm slots. Going back to stock solve the issue. Just making people aware than contact issues may arise when performing this mod, so don't panic if it happens to you.
I just did this on my Tuf D4 and did not loose any of my 4 dimm slots, i used .75mm plastic washers (thats just what i had laying around), this mod improved my temps at load by 3.5c.
This is very well explained, as always.
I always suspected you were the Bob Ross of overclocking.
Bro, you can also try to bend the tab on top of the socket with pliers a little, so you can also reduce the clamping force in the middle of the processor.
24:37 TLDR: if you do that please let the CPU sit on the pins then do the Waser mods. if you drop stuff it protects your pins.
Very important. Those LGA socket pins bend like crazy...
An electrical engineer showed me the same spread, I love it.
I can confirm that, done that, 0.8 plastic washer, temperature drops from 84 to mostly 77, sometimes 80, but mosty is 77C on artcic freezer 34 esports duo, incredibile!!!
I used high grade Automotive rubber washers and they worked just fine. Running on average 27°c while running various benchmarks
Great work Pinocchio!
Intel did occasionally have flat CPU heat spreader back in the Pentium4 days but after they went LGA they never had flat heat spreaders again.
Probably the reason I can run my P4P800SE testing board without paste lol
Yay builzoid (our best friend) is drawing again
I put PTFE washers (4x1mm, measured precisely). Hard to say it's a coincidence or not: cpu temp 70 at high load, benchmark max 88 (magic treshold lol it never surpasses). I'm happy with the results considering I have just aio not custom cooling, my room temp is high all year round and the gpu ocd.
Thank you very much for explaining :)
great video
eso de las arandelas es un riesgo, se puede dañar el pcb con el filo de corte de las mismas, lo mejor es ocupar el bracket fabricado por thermalright o grizzly que ya están diseñados y mecanizados con las medidas necesarias para que apliquen la presión uniforme en todo el borde del IHS del procesador sin dañar nada
firtst
I recently came across this mod,
I got a Supercool Direct die installed, in case you don’t know it , it’s a direct die waterblock that uses an alternative IHS as the contact plate to the die but on its other side it acts as a waterblock since Coolant passes on it.
Would that need the washer mod ?
Very well explained
I'm seeing 77 degrees on a 12700k and a Noctua d15 stock settings with Cenebench r23. I think this comes down to manufacturing tolerances with the cpu bracket. I don't think I would achieve anything doing this mod. AIO users maybe.
im running a 12700kf and a lga 1200 AIO, im considering it. im running my 12700kf OC to the 5ghz all core. 280W.
@@Hammercannon did you do it?
@@mikerzisu9508 this weekend or tonight if I have time.
@@Hammercannon what aio are you running?
@@mikerzisu9508 Coolermaster ML360ARGB LGA1200 version. Side mounted in a Thermaltake View 71, with push/pull fans
Great now I have a reason to skip an entire socket for enthusiast level builds. Budget is the way with alder lake and anything else they release on lga1700
For anyone wanting info on the LGA 1700 socket:
Specification Intel socket LGA 1700:
IHS to MB Height (Z-Stack, validated range) 6.529 - 7,532 mm
Socket Seating Plane Height 2.7 mm
Thermal Solution Hole Pattern 78 x 78 mm
Maximum Thermal Solution Center of Gravity Height from IHS 25.4 mm
Static Total Compressive Minimum 534N (120 lbf), Beginning of Life 356 N (80 lbf)
End of life maximum 1068 N (240 lbf)
Socket Loading 80-240 lbf
Maximum Thermal Solution Mass 950 gram
I just realized my old delidded 3570K might not be dead (no visible damage). It may have lost contact to some pins after shaving off the height.
Would you say that Zen4 is going to have the same issue as you just described because the AM5 LGA loading mechanism pushes only in the middle exactly like LGA 1700? Or is it too short? Also the Zen4 IHS has cutouts which will probably allow for more curvature. How about putting something (washers?) between the IHS and the loading mechanism?
Zen 4 AM5 won't have this issue, AMD uses extra screws to balance out the ILM with better retention mechanism.
With EK's EK-Quantum Velocity² D-RGB - 1700 and 0.6 Nm torque doesn't do nothing on cpu temp , it's aleady set .
i have the same waterblock. have you checked if the coldplate is 100% flat (i didn't before mounting, unfortunately)? what are your experiences in terms of temperature delta to water? how are your temps @stock CB23 in relation to water? i'm peaking at 79 °C cpu package temp while water is at around 27 °C => Delta Water / CPU = 51°C. Delta coolest / hottest core = 67 °C vs 79 °C = 12 °C.
Okay so what if we put the cpu inside a "fake" loading mechanism while laping, to simulate the pressure of when it is actually inside a socket. In my head the cpu would now be more or less flat when it is inside a socket and be curved when it is outside of one.
Idk if this would help, if any, but could be fun trying it out.
What about threadripper with all it's pins, does it bend too ? hmmmm... I think I have seen a very sturdy frame with multiple pressure points... maybe not eh ? hmmm...
I hope this is not a stupid question: How does Carbonaut carbon thermal pad work on the LGA1700 socket with and without washers?
Hi
Where can you buy plastic or metal washers in these sizes?
I didn't even get that far. After no posting with the first cpu I tried, I took the cpu out of the MSI Z690 Edge wifi DDR4 mobo and saw that the top left corner of the substrate was bent up towards the IHS. I'm guessing since the substrate was bent it wasn't making contact with the pins, so I ordered a second 12700k and a different mobo and I got a noctua air cooler instead of the NZXT X72 that I was using. Still waiting on the air cooler to be delivered so I can eliminate what parts are causing this. Very frustrating to go through all of this. At this point it's almost worth not upgrading. Any tips to make sure I can get a post and maybe avoid bending the freakin substrate would be appreciated.
Meanwhile, *Pin Grid Array* and horizontal pressure for electrical contact.
yeah PGA is immune to these kinds of issues.
This makes more sense than's Jay's video.
but Jay's video is what motivated me to make this.
@@ActuallyHardcoreOverclocking I see, yea I saw that headline earlier and watched Jay's video first out curiosity. Made no sense to me to be placing washers on a CPU retention bracket if the issue is some coolers applying too much mounting pressure, or at least the impression I got from his video. Your video cleared that up with something that makes sense.
I thought the other videos did an awesome job... not sure what your video brings to the table??
Reasons why LGA1718 / AM5 looks like a square octopus
Wow those little pins have some powerfull springs.
wow I didnt know that about the socket pins thanks.
My 13900k has great temps with an x73 kraken I was shocked to see it absorbing 253w and not even throttling 😅
12:00 and thus lapping your CPU (if you like to void your warranty) is hereby debunked for normal/gaming use cases. because if you still lap the CPU, you might as well void your CPU Cooler warranty as well and lap it since that wasn't straight to begin with..
The ONLY effect lapping has or CAN have is better temps when you use LN2 paired with hardcore overclocking which.. again, in normal/gaming use cases is NOT the case.
READ ON:
the washer MOD everyone is hyped about, MAY or may NOT work for you.. but you don't void warranties by adding/removing the washers..
However, the actual trick to this MOD is to do this AT THE START -when you actually bought a fresh and new system/parts and still need to build it..
if you do the washer mod AFTER the fact, the chances are that the CPU is already bend and wont bend back. so.. your luck may vary.
If you DID do this mod from the start and your temps are.. off the charts high say 80+ when stressing, remove the washers.. add thinner washers.. and test again.
test until you either are satisfied with the temps and/or have no washers installed (so no mod) which means the waterblock/cooling was already matched with the IHS.
I'd love you to make a video about RAM cooling. Disassembling the heat spreaders, swapping TIM, putting fan on them and so on.
RAM doesn't really produce much heat so just pointing a fan at it will have more of an effect than any complicated procedure will.
Older generations produce some heat and I've got the disassembly figured out by now. A lot of patience is needed. The issue is to find the right TIM. If to thin it often won't make any contact with the heat spreader.
And now the socket design of AM5 with the multiple cutouts starts making more sense...
the cut outs are there because AMD put all the filtering on the top side of the CPU
Yes exactly! But, even without it, the next ryzen ihs and substrate is mostly square like lga 1500 processor, so it doesn't matter much.
Good explanation.
@@ActuallyHardcoreOverclocking is that to prevent the heat from the capacitors from soaking into the ihs?
@@christopherjackson2157 no they literally ran out of space.
what exact size washers did you use? as in OD, ID, im going to test this on my 12700kf with a LGA1200 AIO cooler, thats VERY flat.
Just to have an idea how thick are the washers you used?
nice vid, and there is 7700k , delided, that only works on water cooling block, beacause if you tighten it too much, it refuse to boot with both mem sticks :) i applied silicon under IHS like factory did, but, didnt help. well, its water coooled, screws are loose, buuut, it works, if you touch it (repaste), you gonna curse a lot. 4,8 with normal voltage
I use an nh-d15 which is supposed to have a curved cold plate and I still see spread like you show. Hit about 90 in cinebench and 60s to low 70s while gaming. I'm sure I'd have gains from doing this but not sure it's worth it for my use case.
Make sure the fan speed is able to hit max. (1400-1500RPM) If you use the lownoise adapter it will be 200-300 RPM lower and that could be the reason for the 90 degrees.
90 is still ok though, i think 94-96 is the throttle temp.
Are you concerned about the cpu and mobo bending/warping?
This is very good info thanks bz, I didn't knew the LGA1700 is going to have this much pressure than the previous sockets. Well I guess I have to find washers(hope the plastic kinds don't crack) for my already lapped raystorm waterblock and if I'm going to have somebody asking me to help them build their new pc with this LGA.
Also quick question for my clutch electric screwdriver: how much torque(N•m or ft•lbs) is it supposed to have to safely rescrew the loading mechanism back?
IDK about the torque I just hand tightened it.
@@ActuallyHardcoreOverclocking oh thanks bz! I guess I have set the clutch at around 0.3 - 1Nm(0.2 - 0.73ft.lbs) hopefully the lowest torque on the screws could hold it.
Also my driver is the makita DF012DSE its an awesome pen driver around computers.
Make sure you're accounting for the lower z-height of LGA1700 vs earlier sockets, presumably with washers adding to the spring pressure of the Raystorm mount.
Nylon washers won't crack, but may squish over time.
Never use an electric screwdriver without having a torque spec. It saves like 10s over hand tools. Don't use a wrench, either.
@@concinnus I know about the lga1700 z-height situation it's around 0.717-0.783mm lower than lga1200, that's why this washer mod is an option just in case if I upgrade my brother watercooling PC with the lapped raystorm cuz those new lga1700 spring pins are quite tough & I don't mind the issues of this mod to made it fit. I don't have 12th gen, I'm still on 10th gen & ryzen zen3 with a itx.
When i was replacing the tubes for my bro's PC, I noticed the raystorm mount springs with the nylon washers was squeezing the solder mount of my bro's z490 godlike marking rings on them, thankfully nothing bad happened and I loosen them a bit after that with nearly the same temps.
The Makita driver I mention in my comment before does have a auto stop clutch and torque specs in the manual which is amazing cuz this tool is very gentle on sensitive electronics like motherboards and tough enough to drill fan screws and other stuff, big time & life saver around computers, but it need a set of driver bits like Hexs & other Phillips bits especially the taller bits for tall cpu coolers.
@@emini6 To be clear, the washer mod Buildzoid is talking about is to raise the z-height of the retention bracket, whereas the washers I was talking about were to lower the z-height of the waterblock.
What about removing entirely the retention mechanism or adding padding between the retention mechanism and the cpu so pressure is more even?
I am kinda wondering what’s the deal on the itx boards? For instance I have a z690i aorus ultra itx ddr4, I’m using a heatkiller iv waterblock, my temps are not outrageous, but I am sure it can’t be better. I’m using the update 1700 backplate, springs and washers from heatkiller as well. Idk if my thinking is correct, but the smaller itx size would mean the board has less flex, also my waterblock it’s got a really good mounting system, it’s almost impossible to get a bad mount on heatkiller iv block. Can anyone chime in on this? Forgot to mention the z690i aorus has a robust amount of solid heatsinks and a full aluminum backplate so, I’m not sure if this mod would work for me, I think my mount might be as good as it will get.
So for AMD cpu with flat (either factory or sanded) heat-spreader the optimal water-block should actually be also flat then because AMD doesn't use LGA socket. Correct me if i am wrong please.
yes for flat AMD CPU you want a flat block.