Overclock.net banner

Does Liquid Metal TIM make sense if not deliding the CPU?

6248 Views 26 Replies 7 Participants Last post by  epic1337
Hi Guys,

A quick question: I bought some Thermal Grizzly liquid metal for using in a custom water cooling loop using AquaComputer waterblock, on a Ryzen 3950x (and in future a 5950X). It seems that liquid metal is a very popular solution to use when deliding the CPU, and use it between the actual silicon and the thermal plate (and that makes sense especially when they are using a thermal paste and are not being soldered together).

But my question is: does it make sense to use liquid metal if I do not intend to delid the CPU? Should I expect some better performance still? or it's not worth the trouble (and risk of shorts...). Would I be better off just using some high performance thermal paste? And are AMD CPUs soldered thermal plate? or they use thermal paste internally?

Thanks
21 - 27 of 27 Posts

· Otherworlder
Joined
·
7,566 Posts
They are my images (it also says it btw)

;)

Here is another one by the way.... Kryonaut

:)

View attachment 2512900

If you have used Conductonaut before (or any liquid metal) you should be able to tell by how the IHS looks

;)
by the way, X pattern is more than good enough, would have less edge spillover that way.

 

· Registered
Gigabyte X570 AORUS MASTER, 3950X, Corsair HX1000i, BeQuiet DarkBase Pro Rev.2, Gigabyte GTX1060 3GB
Joined
·
84 Posts
Discussion Starter · #22 ·
My takeaway from this thread is so far:
- the liquid metal will not affect the WB, since it is nickel plated
- the liquid metal will have an esthetic (but not functional) effect on the IHS of AM4 processors.
- liquid metal tends to have worse coverage that regular thermal paste (due to surface tension), but that will be mitigated by careful application and the VARIO type waterblock
- cross shaped thermal compound leads to best coverage (nice gif, I had no idea and was always applying a dot in the middle)

Summary:
- for my combination of processor and WB, it is perfectly alright to use liquid metal, provided I can live with the side effects on the IHS, and I exercise extreme care when applying.
 

· Registered
Joined
·
2,627 Posts
by the way, X pattern is more than good enough, would have less edge spillover that way.

Ive tested many scenarios, using just X is perfectly fine, however as the TechN waterblock has a bow in the middle I add the little extra just incase

;)

My takeaway from this thread is so far:
  • the liquid metal will not affect the WB, since it is nickel plated
  • the liquid metal will have an esthetic (but not functional) effect on the IHS of AM4 processors.
  • liquid metal tends to have worse coverage that regular thermal paste (due to surface tension), but that will be mitigated by careful application and the VARIO type waterblock
  • cross shaped thermal compound leads to best coverage (nice gif, I had no idea and was always applying a dot in the middle)

Summary:
- for my combination of processor and WB, it is perfectly alright to use liquid metal, provided I can live with the side effects on the IHS, and I exercise extreme care when applying.
Yup, thats pretty much it!

Have fun whatever you decide to go with

:)
 

· Graphics Junkie
Joined
·
4,525 Posts
using liquid metal on a non-delidded CPU would not be worth it unless that CPU was soldered to the heat spreader. If the CPU uses regular paste under the IHS then that's where the bottleneck is and that's what needs to be fixed. Yes there could be some temp benefit from doing liquid metal on top of the heat spreader but it would be negligible.
 

· Registered
Joined
·
248 Posts
Guru3d tested this fairly recently and found ~2.5C improvement over Kryonaut Extreme, or ~4.5C over arctic MX-5 on a 3700x with an AIO. I'd guess the gains with zen3 will be slightly higher since that seems to run hotter than zen2. Probably only worth it if you are chasing those last 3 degrees though, which for all practical purposes will not be noticeable when using the system.

 

· Facepalm
Joined
·
11,672 Posts
by the way, X pattern is more than good enough, would have less edge spillover that way.

X is a good pattern.
On very thick pastes however, you need to use X + drops in quadrants. I've seen X pattern with TFX simply not cover the entire IHS (causing two cores to have higher temps than proper) because parts of the IHS didn't get the paste spread. X + drops in quadrants fixed that.

Another method I've not seen mentioned, with very thick pastes, is large drops (rather than the 5 drop method shown in your gif) which may use potentially less paste than X + Quadrant drops.
 

· Otherworlder
Joined
·
7,566 Posts
X is a good pattern.
On very thick pastes however, you need to use X + drops in quadrants. I've seen X pattern with TFX simply not cover the entire IHS (causing two cores to have higher temps than proper) because parts of the IHS didn't get the paste spread. X + drops in quadrants fixed that.

Another method I've not seen mentioned, with very thick pastes, is large drops (rather than the 5 drop method shown in your gif) which may use potentially less paste than X + Quadrant drops.
yes indeed, for very thick pastes or heatsinks with rather lacking mounting pressure, the X pattern would not spread properly enough.
 
21 - 27 of 27 Posts
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top