Thermal Woes

[Hopper64]

as I said before,Thermal Grizzly liquid metal do not cure or dries up.(I successfully removed monoblock to test for contact between surfaces and also I had no problem removing Supermacy EVO from i7 IHS 7700k that had the same liquid metal between them after 6 months)

Thanks. I had the wrong impression with regard to liquid metal. *

[xarot]

Hi there,I did have the same struggles with my 7900x,until I put very generous amount of liquid metal on cpu die(ofc,be careful not to spill over) and also this is very important ,I have placed liquid metal between IHS and monoblock(I got advice for this from someone here on this forum),and my idle temp droped from 35 to 21C and intel stress test on 4,4Ghz/1,20V temps are around 35-40C, before they were around 85C.You will see massive improvment with custom loop.I have 1x420mm,and 2x360mm radiators,GPU block and CPU/VRM monoblok with one D5 pump.I have used thermal Grizzly liquid metal which never cures or dries up.

I think your temp issue was because of if you have delidded your chip, the IHS will be just a bit lower than before, and then the monoblock mayl not make proper contact unless you change the thermal pads to slightly less thick ones. You may have worked around this issue by using a lot more LM where needed, making a better contact.

[Marko911]

I think your temp issue was because of if you have delidded your chip, the IHS will be just a bit lower than before, and then the monoblock mayl not make proper contact unless you change the thermal pads to slightly less thick ones. You may have worked around this issue by using a lot more LM where needed, making a better contact.

Generaly speaking, a "lot more term" is highly subjective because my LM level were barely visible compared to the amount the YT video deliders have used.My IHS height is the same because I have sealed IHS with sillicone every time I did delidding. So IHS contact with monoblock and supermacy EVO was always good.But temps were very uneven between cores.Now,they are pretty much the same.(liquid metal isn't thick as thermal paste so it can't resist IHS pressure from the top,so if you put too much LQ it will squirt on CPU PCB,if too little-uneven contact)...In conclusion,after my third delidd I did manage to put the optimum amount of LQ on the die.

[xarot]

Generaly speaking, a "lot more term" is highly subjective because my LM level were barely visible compared to the amount the YT video deliders have used.My IHS height is the same because I have sealed IHS with sillicone every time I did delidding. So IHS contact with monoblock and supermacy EVO was always good.But temps were very uneven between cores.Now,they are pretty much the same.(liquid metal isn't thick as thermal paste so it can't resist IHS pressure from the top,so if you put too much LQ it will squirt on CPU PCB,if too little-uneven contact)...In conclusion,after my third delidd I did manage to put the optimum amount of LQ on the die.

Thanks for the reply. I see, I personally hate LM and I wish I didn't have to use it. I've delidded 3 Skylake-X chips and one Kaby Lake-X chip. I've seen even 20c difference in cores with Conductonaut, it also comes down to the contact between IHS and the waterblock and the paste spread, waterblocks and IHSs are no flat by default.

For the 7900X CPU, I think cooling it using direct die seems to be the only way to cool it properly with a lot of rad space...I've tried it already. My 7980XE seems a lot easier to cool (18 vs 10 cores anyway), going to try direct die on that in a few days.

[Marko911]

Thanks for the reply. I see, I personally hate LM and I wish I didn't have to use it. I've delidded 3 Skylake-X chips and one Kaby Lake-X chip. I've seen even 20c difference in cores with Conductonaut, it also comes down to the contact between IHS and the waterblock and the paste spread, waterblocks and IHSs are no flat by default.

For the 7900X CPU, I think cooling it using direct die seems to be the only way to cool it properly with a lot of rad space...I've tried it already. My 7980XE seems a lot easier to cool (18 vs 10 cores anyway), going to try direct die on that in a few days.

I couldn't agree more about i9 and direct die cooling But I wanted monoblock so had to be a bit more creative
For IHS height compatibility and monoblocks I would definitely recommend using sillicone to seal it off.
Every i9 except the 7900x have bigger die surface so its much easier to cool, so I think you're in for a treat when you start naked cooling your 7980XE
Wish you all the luck with that!

[Abaidor]

I couldn't agree more about i9 and direct die cooling But I wanted monoblock so had to be a bit more creative
For IHS height compatibility and monoblocks I would definitely recommend using sillicone to seal it off.
Every i9 except the 7900x have bigger die surface so its much easier to cool, so I think you're in for a treat when you start naked cooling your 7980XE
Wish you all the luck with that!

BTW in case you are not aware of it Watercool.de have a VRM block for the Rampage VI Extreme so you can get this + a CPU block if you want. I might do it in the future but for the moment I went with the EK Monoblock and I still putting together my build that got delayed due to excess workload and several mods that I wanted on my case.

[LiveOrDie]

mm that's another possibility. the fans are in pull mode and there doesn't seem to be much air coming out of them. even if I put my hand right under the Rad I can't feel any airflow, but maybe this is normal with this AIO. there isn't too much info about it out there apart from a few glowing reviews.

how can i test where the issue is? i dont want to throw money at a loop if there is a greater issue at play here (like airflow issues in the case). also I'd probably look to reuse the existing fans on a custom loop but if the fans are crap then is there any point doing this

Its nothing to do with rad space its all to do with there just not enough surface space to transfer the head from the IHS to the monoblock or any cooler, My CPU gets up to around 85C at times i run mine at 4.7Ghz 1.221v, I delidded my CPU on day one it dropped my temps by 15c, its pretty clear OP has got a crappy delidded chip there.

[Bmacc]

Its nothing to do with rad space its all to do with there just not enough surface space to transfer the head from the IHS to the monoblock or any cooler, My CPU gets up to around 85C at times i run mine at 4.7Ghz 1.221v, I delidded my CPU on day one it dropped my temps by 15c, its pretty clear OP has got a crappy delidded chip there.

thanks for the reply but i still think it's something with my particular cooling setup rather than anything with the CPU. I would be very happy with 85C temps but i get way higher ones at lower voltages/clocks.

single Cinebench run @ 1.19v



XTU hits 95C+ within a couple of minutes.