Can Liquid Metal be used on 100% Copper IHS/Heatsinks?

R

Rev. Night

[H]ard|Gawd
Joined
Mar 30, 2004
Messages
1,489
Hi all,

I am about to tackle a few liquid metal projects (1 cpu, 2 gpu) using TG Conductonaut and I thought I had all my bases covered. Now I am reading that TG only recommends using it on nickle-based copper IHS/heatsinks. All 3 of the intended items are 100% copper, not nickle based. Does this mean that I can't use liquid metal on 100% copper? Will it need to be replaced in a year? I have zero interest delidding this yearly. A few reddit threads are saying that GA + CU = CU Alloy, can lead to issues.

FYI, I will be using Kapton tape around all dies/heatsinks to help protect the pcb. Delid will be sealed in the 4 corners by crazy glue.

CPU: I7-6700K Delid. Copper IHS from Rockit 88
ith_08583632-5194-455f-9022-114a1332b219_1024x1024.jpg

GPU 1: Sapphire Pulse 5700xt
tc6g2p65f6n31.jpg

GPU 2: Sapphire Nitro 480x with Arctic Cooler Twin Turbo II
coolerturborear.png
 
Yes copper is fine. So is nickel plating. Aluminum is what you don't want to use it on.
 
Right, exactly, this is between die and IHS only. The heatsinks for the GPUs, the contact plates, are copper.

How long does it last? I read some people having to reapply the LM after a year or two.
 
Rev. Night said:
Right, exactly, this is between die and IHS only. The heatsinks for the GPUs, the contact plates, are copper.

How long does it last? I read some people having to reapply the LM after a year or two.
Click to expand...
I believe it is good practice to reapply it after 2 years. Fr what I read it doesn't degrade that much after 2 years and can be used for long with slight increase in temps the old we it gets.
 
Nenu said:
There can be a small amount of pitting on the copper surface after you clean off the old LM.
There is still a reaction but it isnt on the same scale as aluminium.

This should give you a good idea
https://www.gamersnexus.net/guides/...cts-copper-nickel-and-aluminum-corrosion-test
Note, they didnt test how effective the stained copper is when used again with LM.
Click to expand...
Yes it will permanently stain any material. I repasted CPUs before with lm and never notice a degradation of performance from stained IHS. It did take a lot of elbow grease to get off all the old LM but couldn't get the stain out. Was after about a year when I did it.
 
So I will have to delid the cpu again in 2 years? Thats fine, but it means I have to hold onto the delid tool and not resell it
 
i never understood the LM craze yes it helps with temps a bit but not enough to mess with...that 5c droop wont get you anything
 
for GPUs, yeah you are right, im only doing it because I have it on me. This is mainly for CPU delid
 
Rev. Night said:
yeah the stock paste that Intel used to bring die to IHS is really, really bad.
Click to expand...
It is a paste that is made to last. It is not bad just a different application. It won't dry out or "push out" that de8auer likes to call it. It is meant to never be changed. Most consumer pastes and LM should be changed ever couple years. Granted Intel was a bunch of cheap asshole and should of just soldered the IHS.
 
tunatime said:
i never understood the LM craze yes it helps with temps a bit but not enough to mess with...that 5c droop wont get you anything
Click to expand...
Like other said it is the best for deliding. If you going that far you might as well use LM. I never used it outside of deliding. I went from mid 90s on a OC 7820x under water to low 70s with LM and deliding.
 
yeah i know. I'll have alot of it left and alot of thermal tape left too, so i might as well use it.
 
you will get likely get good gains on the GPU, too. I took 10c average off my RTX 2060 with Thermalright Silver King liquid metal. And about 13c off my Kaby Lake i5 7600k. They also cool down WAY faster, after you are finished loading/stressing.

I do not recommend using crazy glue. for 4 reasons:

1. Some super glues will eat through your CPU's PCB.
2. You probably will not get all of the glue off, if you have to re-lid. Which will mess up the height of the IHS.
3. Super glue will probably come undone in a couple of months and need to be re-done. I know, because I had to do it.
4. You should adhere the IHS along all sides (with a small gap). Because the CPU's PCB flexes in the socket and can pull away from the IHS, which will lower the contact between the CPU die and the IHS
 
Interesting points. According to Rockit, whose copper IHS I am using, they even use a form of crazy glue (the same type I bought "Loctite Ultra Gel Control Super Glue") in their instructional video:

 
Rev. Night said:
Interesting points. According to Rockit, whose copper IHS I am using, they even use a form of crazy glue (the same type I bought "Loctite Ultra Gel Control Super Glue") in their instructional video:

Click to expand...

Yeah I've seen that and....I suspect they probably don't still recommend super glue. But probably didnt' want to redo the video.

anyway, this is what I used, after my own superglue failed

https://www.amazon.com/Permatex-821...t=&hvlocphy=9061079&hvtargid=pla-437600037492

liquid gasket maker is what Kyle/[H]ardocp recommends. Get one rated for high heat. I chose that one because its rated for high heat. And also, I thought the "high torque" aspect might help keep the CPU's PCB from flexing away from the IHS and breaking contact. Still working over a year later.

I also used a copper IHS upgrade. If you look at the underside of the IHS, it indents to make space for the CPU die. Rather than applying the gasket maker directly to the borders of the IHS, I applied to to the inner edges of the indent. This way, I would get the cleanest and lowest contact. Don't forget to leave a little gap for air and gases to escape.
 
As an Amazon Associate, HardForum may earn from qualifying purchases.
This. It is what I used when I went to redo the LM later. It popped right off with no damage using the tool. I really don't recommend using any kind of adhesive.
 
Last edited:
Interesting. I sent an email to Rockit asking them their opinion. Will post up should they respond
 
Be careful with the CA adhesive it is flammable I would have to look up the auto ignition T. I am a retired professional Chemist, it will burn if hot enough.
 
Ultra clean and low, after applying it to the inner edges.
WP_20190726_07_03_08_Pro.jpg



chameleoneel said:
Yeah I've seen that and....I suspect they probably don't still recommend super glue. But probably didnt' want to redo the video.

anyway, this is what I used, after my own superglue failed

https://www.amazon.com/Permatex-821...t=&hvlocphy=9061079&hvtargid=pla-437600037492

liquid gasket maker is what Kyle/[H]ardocp recommends. Get one rated for high heat. I chose that one because its rated for high heat. And also, I thought the "high torque" aspect might help keep the CPU's PCB from flexing away from the IHS and breaking contact. Still working over a year later.

I also used a copper IHS upgrade. If you look at the underside of the IHS, it indents to make space for the CPU die. Rather than applying the gasket maker directly to the borders of the IHS, I applied to to the inner edges of the indent. This way, I would get the cleanest and lowest contact. Don't forget to leave a little gap for air and gases to escape.
Click to expand...
 
As an Amazon Associate, HardForum may earn from qualifying purchases.
chameleoneel said:
Ultra clean and low, after applying it to the inner edges.
Click to expand...

Curious, would it be find if I put Kapton tape around the cpu die and on the interior of the IHS? For the cpu, I don't want LM spreading onto the chip, and for the IHS, its always easier to tape out where you want stuff to go before you paint. I bought some just for this purpose. The only thing to keep in mind is not to put too much down so that it causes a gap. Like around where the IHS goes.
 
I wouldn't leave any tape on there, if that's what you mean. I doubt any tape will hold under such high heat.


For the exposed stuff around the CPU Die, I just covered it in more gasket goo.
 
ok so results! I was able to do several COD sessions at 40% PWM and 50% PWM (horizontal line fan curve across all temps) and then a gradual curve from 20-45% PWM (realistic fan curve, because who wants to hear full power at idle).

Stock Paste/Stock Fans:
Core: 75C
Junction: 90C
PWM set to 20-50% but kinda loud

Noctua NT-H1/Stock fans:
Core: 73C (improvement of 2C)
Junction: 83C (improvement of 7C)
PWM set to 20-50% but kinda loud

Noctua NT-H1/Custom Arctic P12 fan bracket:
90% PWM: 66C, 76C
70% PWM: 69C, 78C
50% PWM: 73C, 83C
45% PWM: 76C, 85C
40% PWM: 80C, 90C
Somewhat cooler temps, significantly cooler if you go up in PWM, but its alot quieter.

TG Conductonaught/Custom Arctic P12 fan bracket:
50% PWM: Starts at 67, 76, ends at 69C, 78C. (Improvement of 4C core, 5C junction)
20-45% PWM: 76C, 85C (Improvement of no C to 4C core, 5C junction)
40% PWM: 78C, 88C (Improvement of 2C core, 2C junction)

So where does this leave us? Improvement from Stock paste to LM is huge, up to 6C core and 14C Junction, even more if you really speed up the fans. Noctua NT-H1 will offer middle ground improvement with TG Kyro offering maybe a C or 2 better. The custom 120mm fan bracket does a great job over stock Sapphire Pulse 92mm fans. Overall, this is a good project to do especially if you are about to do a CPU delid, since you will have all the stuff you need already.
 
You must log in or register to reply here.
Back
Top