YOU GUYS asked me to try this... so I did

so i was sent an article on twitter by quite a few of you actually um from igor's lab which was kind of a collaborative testing sanity check of something that bill zoid said regarding 12th gen temperatures and we'll take a look at that today funny thing about this this actually isn't new we've talked about this in the past with other versions of intel we've talked about with amd um but it seems like all of these things have kind of like created the perfect storm with 12th gen and it kind of leads to well just intel being intel evga is proud to announce their new x570 for the wind motherboard with multiple rgb and argb led headers fin stack heatsink switchable bios and 15 phase digital vrm the x570 for the wind motherboard brings evga's legendary build quality and performance to ryzen processors to learn more and see the full feature set follow the link in the description below uh okay so some of you might now either have read the article and already know where this is going but we want to test this for ourselves and those some of you may have no clues what we're talking about when i even reference the past first things first uh this 12th gen intel is a bigger die because of the fact that it's got e cores and p cores you'll notice it went from a square to a rectangle so it's longer with that a new socket type of course we're talking about lga 1700 which is found right here in our apex board yes we've already know these boards have caught fire we already know why the capacitor is backwards we talked all about it our board's fine but with constant socket changes which for the most time for the most part intel didn't matter because it retained the same size and we're talking about just the small square size of a typical intel cpu but now that they're longer the only thing that changed on the previous size were the amount of pins in that array but the size overall in the substrate stayed the same now that we have a larger substrate and more pins 1700 of them to be exact we no longer have this nice square thing we have to have tension on to keep cool we now have this rectangle compound that with the fact that we have shown a million times in the past going all the way back to like my x99 cpus that they're not flat that the ihs on them become very concaved or domed i've had some that were completely concave where if i start lapping it it's a perfect square starts for me or in this case a rectangle will start forming and then it takes forever to reach the center and what that shows you that it's either concaved or domed if it starts to appear in the center with the sanding marks it's domed if it appears on the edges it's concaved the flatter the surface the better the cooling because of even tension and even distribution of heat across the cooler if you have a dome section that means you have an air gap which has to be filled in by thermal paste thermal paste will do its job but it's not going to be a very efficient transfer of heat because it's traveling through more thermal paste to transfer the more metal we have touching without thermal paste needing to fill those microscopic peaks and valleys the better heat transfer we're going to get couple that with 12th gen now having a really tall socket where it's much taller than it was previous generations of sockets we now have an issue and the issue is not just the fact that intel socket is taller the issue is the fact that we have a million different types of coolers out there which now have to be compatible with a taller socket and a lower socket from previous generations and what a lot of brands have done is they have just basically said oh yeah our previous mounting mechanism works and we'll just ship that and call it 1700 where what they'll account for is the extra spread in fact if you look closely at this motherboard here you'll see now instead of one set of holes you have two sets of holes which kind of overlap so you get this sort of a eight shape or like a snowman looking thing the problem with that is a lot of cooler manufacturers took that and went oh as long as we just make it wider and we have a bracket now that has adjustable slides on it see all these hear the wobbling that's because all of these can slide they go okay as long as we can match the width everything's fine the problem is if we take a look at typical asymtech retention system and i'll use this one right here and i'll kind of put one together here real quick it threads through the motherboard into this which goes on the back of the motherboard you thread this down until it stops right you got your cooler that goes through it you take this nut and you thread it down until it stops that's all fine and dandy you know there's no springs or anything in there it's just go until the threads stop that's all fine and dandy with the old sockets that had no z-axis height change but now that we have 1700 which has a z-axis height change that's taller cranking down a retention system like this which a lot of manufacturers are shipping with their 1700 compatible coolers is putting far too much tension on the cpu and what's that causing it is causing an actual bow to happen with the cpu itself now our cpu or my cpu here if i look at it is not bowed i know you guys can't see this on camera but i'm just sort of standing checking it myself i'm lining everything today i'm just standing checking it myself yeah that's flat there's no bow there you can't see this on this camera either maybe we'll try and get a single shot of it though but if i take a look at my edge here i am seeing all kinds of light which tells me my ihs is concaved i've got light in the center and not on the edges so we got a concave so this is something we've known with intel cpus for a long time i've lapped them for generations we've done videos about it we've showed you how i've dropped those which is 10 celsius by just lapping the cpu but what igor's lab and build zoid has found is that by raising the height of the retention system here means that the cpu is not sticking as high above it bringing it up closer to the cpu is reducing the tension and causing less of a bode effect because it's not over tightening as it comes down the issue i believe fully stems from these old types of ace attack retention sockets and and hardware that you'll find with your standard blocks if we take a look right here the torques torques torques and torques that makes the retention system come off they put a one millimeter washer around those and put it back down so what that does is it brings the whole socket system up a little bit higher which is supposed to alleviate uh some of the pressure so it just brings this piece right here the door closer to the top of the cpu they say that's led to a 5c reduction in temperatures we'll be using a 360 aio right here from fractal design this is actually an old setup will use the old system that i know people are doing are taking their old coolers plopping them on 1700 and then saying temperatures are bad so let's test it with the old hardware and we'll get our baselines and then we'll test it with the spacer mod and then we'll test it with a lap mod so i just tested the cinebench run it's looping right now it's set for 30 minutes remember it will take time for all the coolant to go reach its max temperatures but it's funny because i use the kingpin thermal paste the temperatures are already so so good here i mean look at this low 70s 70s 70s 70s 75 75 67 74 73 right if you look at the chart though you can see they are trending upward right see how this if you draw a line through the average it's going up so we need to let this go until it completely stabilizes and see where it's going to go like right off the bat i would say this is not a problem and the funny thing is i'm using the old hardware which is cranked down as hard as it should go which is very tight this is what i hate about when we're sent articles saying look blah blah blah is saying x y z x y z is hot yada yada i mean enter whatever is trending there when we did our 12th gen video people were messaging me going dude it's so hot it's super hot why didn't you talk about the temperatures because this is what we were experiencing these temps i would not call these bad i mean sure they're water cooled right now in the 70s which some people would say it's warmish i don't even have the fans at 100 they're at 75 and it's a basic asetek aio there's absolutely nothing spectacular about this so this is not temperatures that would make me raise an eyebrow and there's something's wrong here um like when we did the 12th gen test in the corsair one and that was hitting 100 c that's obvious because of the size of the cooler it makes me wonder if maybe there's a mounting issue with that in terms of this particular issue that was mentioned but i'm still going to go through with the with the washer test to see if we see any sort of change in the temperature whatsoever and i did actually enable the asus enhancements not the core clocks so for instance we look at core clock here you'll see we're running at 4.9 if we didn't drop it yeah 488 0. so that's actually 4.9 look at that locked speeds nothing changing our e-cores are down to 36.85 so 3.7 no throttling no fluctuating no issues whatsoever but we'll let it go for like another five or five or ten minutes just to make sure that the temperatures are not going to continue to climb and then we'll show you the mod that uh igor's lab and build zoid were talking about now and the thing about this kind of test is it's gonna be so much variance between motherboard types are they all using the exact same retention system remember how threadripper had some that were troublesome and some weren't i don't know if that's the case with this particular retention socket so it's been running for a while now as you can see and our temperatures are not climbing whatsoever so if we look at 11 and 12 again yeah they didn't they never spiked any higher 78 78 80 was still our max all right so you know what we need to do now we need to stop the test and look how fast those temp shot down 25 22. all right well next thing we got to do now is we are going to do the washer method and then after that i really want to lap the cpu like i wouldn't have even looked at these temps and like i have to lap this cpu some of the other ones like the 9900k and whatnot we definitely had to lap those but i have to tear this apart now so i'm leaving the cpu in the socket because i want to protect the pins that's something that igor's lab had even mentioned but i'm going to go one step further and then i'm going to actually like electrical tape down the cpu real quick the socket's a little different than it was in the past in the past it like fully surrounded the cpu i'm taking my ifixit with the torx head these are very loose i can that's about as tight as they were so just keep in mind how tight they were so that's the bottom and i think i'm going to go ahead and find washers that will fit that okay so i actually happen to find some plastic washers that from an old water block we just need to screw this back down there we go so not too tight just snug so that's that one like if i were to do that now it'd be a very uneven amount so we have to repeat the process on this side over here so they say one millimeter is ideal if you go higher than that the issue is now you don't you don't you can't have the retention go up higher than the cpu or you'll be touching the retention and not the cpu itself so now same thing i'm going to put the same amount of kingpin on there and who knows maybe this is the reason why we were cooling so well was i was using the kpx which i use on everything now because you know kingpin fanboy here all right let's just start the test here and let's see what the initial spike is that's already 5c lower because it spiked last time to 70. i hate when smart people are right okay so our temps as you can see are just not climbing anymore uh 72 72 70 69 it's about the same honestly if we look at 11 and 12 again the two hottest cores 78 78 they maxed out 79.79 we can't call that one degree anything other than potential margin of error it's almost like this cpu doesn't have whatever issue other cpus are potentially having maybe this one's a little less concave than others maybe the tim inside better applied it doesn't say confidential on it we for ever since 10th gen i've been getting ones that are stamped like the retail versions so i don't know i don't know necessarily what it is that's cause because i mean i just did the spacer mod and nothing really seemed to change uh the only thing i can suggest at this point now is i'm gonna try lapping it just to see if anything improves there but that doesn't mean that this particular fix isn't going to work for you guys uh it clearly worked for igor's lab and it worked for buildzoid let's take it back apart again and then let's uh let's lap it so one thing that's really interesting here if we look at this pattern on the cooler you can see how in the center it's a little thicker right there that's actually indicating our our concave right there so because there's concave which means there's more gap between the cold plate and the cpu ihs means there's room for more thermal paste there which is why it's thicker there and you can see it's thinner around that because that's where it's higher that didn't really appear the first time it appeared the second time with it being under less tension so i'm really kind of curious now as to how this is going to perform after lapping it so i want to see what the initial pattern looks like and you have to do this on on a sheet of tempered glass because it's perfectly smooth the hardest part here is getting this the pressure down even on the cpu if you start putting more pressure on one side you can actually elaborate cook it uh kingpin has a neat sanding disc that does all this you just hold it on there but i've had good success doing it this way multiple times and the water it's distilled water and it ain't going to hurt nothing it just keeps the sandpaper from getting clogged up so the initial pattern is exactly what i expected it to be if you look on that you can see the four corners are definitely higher all right so as you can see i flattened the ihs down to where the copper is uniform now and i know it looks super rough that's because of the fact that it was 400 grit and then we went all the way down to 1500 uh it's very very smooth those are just scratching marks um from the sanding i know people are going to be like but that's going to affect the cooling no it's not any more than the machine marks in your cold plates affect cooling it's more important for it to be flat than it is for it to be polished like a mirror you don't need it to be like a mirror so anyway here we go repeating the test same standoffs and stuff we had before only now we have a lapped cpu predictions on this one me i'm thinking maybe a couple degrees at the most i mean there was a pretty big concave in fact when we tested it now with the light like phil was saying that there was like the teeniest tiniest of a sliver on one side and that could be the ruler itself because again this is not the most scientifically precise straight edge but the fact that you could see light through there before like a lot of light versus now i'm feeling pretty confident with this so let's watch in real time as i start this test before they spike to about 65 66 and then they sort of normalize around 72. remember this is lapped cpu yeah really unchanged the question's gonna be what happens now over time which i think will be realistically about the same there's another theory right now that i was thinking about as i was doing this that could end up potentially being an issue and that is if i lapped enough material down and then we had raised that socket up if i've now kind of cannot you know sort of counteracted that to where maybe it's too level now and now maybe the cpu cooler isn't getting a good enough mount but i mean that would mean i would have had to take a millimeter of ihs off and i don't think that's the case because i don't think it's even that thick so this is what we're going to call it uh i'm essentially calling this one no change lapping not i mean a void your warranty obviously but if you are not scared to do it and you have a very weird concave cpu i recommend doing it we do it with all our intel cpus but look at the temperatures 72 is the max on core 1 and 2 70s here 75 on 5. go down to 11 and 12 which is our hottest 78 78 exactly as it was before max of 79 just like our last test so unchanged voltages are locked fan speeds were locked everything was locked so that it could not fluctuate to create any sort of a weird effect in our testing here that particular article obviously it helped build zoid and it helped igor's lab uh didn't really have any change for us maybe 2c 3c at the most and it's hard to say whether or not that's margin of error the temperature in this room has not changed so ambient temp has no effect on this um that's our findings take it as you will 12th gen still hot because of that 227 watts of power draw 2 9 29 28. let's call 230 230 watts of power draw under load anyway all right guys there you go take that as you will i'm not saying the article's wrong i'm just saying it didn't help our particular cpu or you know its effect on this motherboard thanks for watching guys and we'll see you in the next one

Channel: JayzTwoCents

Views: 425,906

Keywords: 12900k cooling mod, 12900k, intel, intel cooling mod, z690 cooling mod

Id: tUEDU6oQmzs

Length: 17min 23sec (1043 seconds)

Published: Sat Jan 15 2022