10:28 Do you happen to know the measurements or model number of the screws and springs used? I'm running the exact same setup and have run into the same issue.
So with the direct die plate... you can go crazy with the tightening all you want, you actually really want it to be very snug. The Rockit Direct Die plate will not let you crack the die or cause an issue from over tightening, you will strip out the threads before you tighten too much to hurt something ( you can't if plate is installed correctly ). Also I don't understand why don't you use liquid metal instead of thermal paste? I have a 10900K with the same kit and I'm using an EK Classic water block run through EK's 360mm Thick all copper radiator and the res/pump combo that comes with the EK Classic kit. I used liquid metal on my direct die and then ran the Auto OC Bot in the bios of my EVGA Z490 Dark. It set me up on 5.2Ghz @ 1.28v * not bad * but I was then able to click it up to 5.3Ghz @1.375v. I had my 9900k with the same direct die system and cooling running 5.3Ghz @1.320 all day every day with the EK Classic 360mm kit and liquid metal combine with the direct die kit from Rockit
Sounds impressive... Myt SP63 uses a lot of power more beyond 5.0 .. Yours must be a SP80 quailty at least.. My BIOS recommend at LLC6 for 5.0Ghz its over 1.4v haha! I am running 1.31v LLC6 cpu 50x cache 47x 100% stable..
Ended having use two polymer washers instead of the one to make it work with this same AIO and the RK direct die kit. Ran OCCT SSE/Medium and immediately my temps show to 90 degrees. It never even reached 70 degrees delided without direct die. I checked to make sure it was making contact with the aio cold plate and it was but apparently I needed to get a little lower. Ambient temps were fine - low twenties iirc. Ran OCCT again but this time with AVX2/Small/Extreme - fully expecting thermal throttling - but it maxed at 70 degrees. Are you sure about the die being protected while tightening? Direct die frame tightening or actual aio onto the die tightening? Thanks.
@@lassekristensen385 He's SP100, I too can do 5.3GHz @ 1.373Volts LLC5. There are no 10900k's that I know of stable at 5.3GHz AVX. Basically temps become impossible to tame. My SP63 performs close to yours.
@@GodKitty677 I 100% think you are right. I had a SP63 and I think I ran it at 5.1 or 5.0 in the end. this was with watercool and and delid/relid?. Although I was not THAT experienced in OC as now I felt still that I could not be getting any better with what I had without pumping voltage. I have a 14900KS now that sadly has a limit below marketing. I run it at 58/44. This is the limit for it to run flawless and shader compile, ue5 compression etc. Its a great CPU though and I feel that even if only 100Mhz on P and E lower. I spend so much money on direct die, liquid metal, watercooling, "car like" radiator with 3 pumps and I cannot reach what I basically paid for! But it runs amazing like this- I would just not go through all this again! :) Next time it will be a lower watt cpu and regular watercool (by that I mean I keep my car rad and pumps haha!) but no delid, break warrenty, liquid metal, direct die etc. its just not worth it- But this CPU should last at least a few gens more, loads of life and performance left of corse.... But hopefully we can see through AMD that the 350-400watt show is over!
Hey been think of upgrading from Z390 i9 9900K to z490 i9 10900K, of late, anyone know if the rockit kool 9th gen direct die kit if it should work with 10th gen or not.?
I'm thinking that early adopters of the 10900K got much better chips and over time as more have been released theres been a slight decline in their quality. I got a 10900K about a month ago and my chip runs 5.2 w/ 48 uncore at 1.35v no avx offset and tops out at 85C full load using a Corsair 360 AIO normal mounting. However, I've seen the performance of other ppls chips who got them near day 1 launch and they were getting like 5.3 to 5.4 w/ 47-48 uncore at 1.32-1.35v. Then we saw that intel released the 10850K which putting 2+2 together says that Intel only had a limited amount of 10900K's. Its very strange that yours is running that hot at nearly the same specs so I wonder if the 10900K is near the end and intel will push out the 10850Ks soon until they run out. Intel has had fabrication issues for a while now but its gotten a lot worse which is why they kept pushing back 10nm and 7nm. I think its safe to say that 14nm 10th gen is the end of the road for that arch.
This 10900K that we tested in this video was an early one, it was preordered back in may. It does seem to be extra hot though, even by 10900K standards. But your theory makes sense, Intel is definitely having issues pushing out 10900k's!
@@sabishiihito Kitguru did a vid today saying that Intel's next chip for z490 will actually be going back to 8 cores but will have pcie4. So that may be true 14nm will be on 11th gen. I am slightly concerned that the Nvidia 3080Ti / 3090 will bottleneck on pcie3. The current testing and theories says it wont but we wont know until its released and tested.
@@hardwarelens8040 well tbh pre-ordering and actually having it on hand doesnt mean much. If you had received the chip back in May and were just now getting around to OC'ing it that would contradict my theory but you just now received it. I also noticed that the 10850K's are now on Newegg for $500 which again tells me the 10900K's are all about done so what evers left out in the market are the last. Funny thing is no one is touching the 10850K's. They are the same chip just a lower bin that does 100mhz less. Ive seen the test comparisons and I would gladly use one as my use case isn't benefiting that much from 5.2ghz. I went 10 cores because I run a lot of background tasks while gaming, the 2c/4t help take up the extra load of those apps.
@@heyguyslolGAMING just to clarify, we received this particular 10900K on launch day back in may, its just that there was no direct die kit available until recently so we never made a video about it until now. It did seem like we kind of lost the silicon lottery with this one though based on temps and OC results others were getting.
I know it has been a year since you commented on this video, but I'd really appreciate that you explain why regular thermal paste cannot be used on a direct die. Throughout the years of building custom watercooled PCs, I have used arctic MX4 multiple times on GPU's without issues and great temps. Why would a CPU die be different? I am not planning to overclock. Thank you :)
10:28 Do you happen to know the measurements or model number of the screws and springs used? I'm running the exact same setup and have run into the same issue.
So with the direct die plate... you can go crazy with the tightening all you want, you actually really want it to be very snug. The Rockit Direct Die plate will not let you crack the die or cause an issue from over tightening, you will strip out the threads before you tighten too much to hurt something ( you can't if plate is installed correctly ). Also I don't understand why don't you use liquid metal instead of thermal paste? I have a 10900K with the same kit and I'm using an EK Classic water block run through EK's 360mm Thick all copper radiator and the res/pump combo that comes with the EK Classic kit. I used liquid metal on my direct die and then ran the Auto OC Bot in the bios of my EVGA Z490 Dark. It set me up on 5.2Ghz @ 1.28v * not bad * but I was then able to click it up to 5.3Ghz @1.375v.
I had my 9900k with the same direct die system and cooling running 5.3Ghz @1.320 all day every day with the EK Classic 360mm kit and liquid metal combine with the direct die kit from Rockit
Sounds impressive... Myt SP63 uses a lot of power more beyond 5.0 .. Yours must be a SP80 quailty at least.. My BIOS recommend at LLC6 for 5.0Ghz its over 1.4v haha! I am running 1.31v LLC6 cpu 50x cache 47x 100% stable..
Ended having use two polymer washers instead of the one to make it work with this same AIO and the RK direct die kit.
Ran OCCT SSE/Medium and immediately my temps show to 90 degrees. It never even reached 70 degrees delided without direct die. I checked to make sure it was making contact with the aio cold plate and it was but apparently I needed to get a little lower. Ambient temps were fine - low twenties iirc. Ran OCCT again but this time with AVX2/Small/Extreme - fully expecting thermal throttling - but it maxed at 70 degrees.
Are you sure about the die being protected while tightening? Direct die frame tightening or actual aio onto the die tightening?
Thanks.
@@lassekristensen385 My cpu is just as bad it, took creative overclocking to get the frequency higher.
@@lassekristensen385 He's SP100, I too can do 5.3GHz @ 1.373Volts LLC5. There are no 10900k's that I know of stable at 5.3GHz AVX. Basically temps become impossible to tame. My SP63 performs close to yours.
@@GodKitty677 I 100% think you are right. I had a SP63 and I think I ran it at 5.1 or 5.0 in the end. this was with watercool and and delid/relid?. Although I was not THAT experienced in OC as now I felt still that I could not be getting any better with what I had without pumping voltage. I have a 14900KS now that sadly has a limit below marketing. I run it at 58/44. This is the limit for it to run flawless and shader compile, ue5 compression etc. Its a great CPU though and I feel that even if only 100Mhz on P and E lower. I spend so much money on direct die, liquid metal, watercooling, "car like" radiator with 3 pumps and I cannot reach what I basically paid for! But it runs amazing like this- I would just not go through all this again! :) Next time it will be a lower watt cpu and regular watercool (by that I mean I keep my car rad and pumps haha!) but no delid, break warrenty, liquid metal, direct die etc. its just not worth it- But this CPU should last at least a few gens more, loads of life and performance left of corse.... But hopefully we can see through AMD that the 350-400watt show is over!
What's the temp diference between just delided (liquid metal+ihs+thermal paste) and delided & Direct die ?
der8auer made a video testing exactly that.
Hey been think of upgrading from Z390 i9 9900K to z490 i9 10900K, of late, anyone know if the rockit kool 9th gen direct die kit if it should work with 10th gen or not.?
I'm thinking that early adopters of the 10900K got much better chips and over time as more have been released theres been a slight decline in their quality. I got a 10900K about a month ago and my chip runs 5.2 w/ 48 uncore at 1.35v no avx offset and tops out at 85C full load using a Corsair 360 AIO normal mounting. However, I've seen the performance of other ppls chips who got them near day 1 launch and they were getting like 5.3 to 5.4 w/ 47-48 uncore at 1.32-1.35v. Then we saw that intel released the 10850K which putting 2+2 together says that Intel only had a limited amount of 10900K's. Its very strange that yours is running that hot at nearly the same specs so I wonder if the 10900K is near the end and intel will push out the 10850Ks soon until they run out. Intel has had fabrication issues for a while now but its gotten a lot worse which is why they kept pushing back 10nm and 7nm. I think its safe to say that 14nm 10th gen is the end of the road for that arch.
I'm pretty sure 11th Gen Rocket Lake will be 14nm as well, just adding PCIe 4.0
This 10900K that we tested in this video was an early one, it was preordered back in may. It does seem to be extra hot though, even by 10900K standards. But your theory makes sense, Intel is definitely having issues pushing out 10900k's!
@@sabishiihito Kitguru did a vid today saying that Intel's next chip for z490 will actually be going back to 8 cores but will have pcie4. So that may be true 14nm will be on 11th gen. I am slightly concerned that the Nvidia 3080Ti / 3090 will bottleneck on pcie3. The current testing and theories says it wont but we wont know until its released and tested.
@@hardwarelens8040 well tbh pre-ordering and actually having it on hand doesnt mean much. If you had received the chip back in May and were just now getting around to OC'ing it that would contradict my theory but you just now received it. I also noticed that the 10850K's are now on Newegg for $500 which again tells me the 10900K's are all about done so what evers left out in the market are the last. Funny thing is no one is touching the 10850K's. They are the same chip just a lower bin that does 100mhz less. Ive seen the test comparisons and I would gladly use one as my use case isn't benefiting that much from 5.2ghz. I went 10 cores because I run a lot of background tasks while gaming, the 2c/4t help take up the extra load of those apps.
@@heyguyslolGAMING just to clarify, we received this particular 10900K on launch day back in may, its just that there was no direct die kit available until recently so we never made a video about it until now. It did seem like we kind of lost the silicon lottery with this one though based on temps and OC results others were getting.
Direct Die = LIQUID METAL! The thermal paste can simply not transfer heat fast enough.. Its fail to use paste on DD.
Bitch I dont want this, I want water literally running over the core and touching it
try noctua nhd15, you will be surprised :)
You should NEVER use regular thermal paste on a direct die. ALWAYS use liquid metal. In this case, you should have been using the Conductonaut.
I know it has been a year since you commented on this video, but I'd really appreciate that you explain why regular thermal paste cannot be used on a direct die. Throughout the years of building custom watercooled PCs, I have used arctic MX4 multiple times on GPU's without issues and great temps. Why would a CPU die be different? I am not planning to overclock. Thank you :)
280 aio is to low for a 10900k. min 360 aio for the 10900k
um ya your 10900k is way out of tune.