Delid Thermal Compound

[1982 Original]

I’m getting ready to delid my 12700kf using the RockitCool kit. Obviously the majority of people use Liquid Metal between the IHS and the die. Can a standard TIM be used?  Has anyone had success using something other than Liquid Metal?

 

 

[Shimejii]

It can be used, usually if you are going this extra mile anyway, you might as well use LM for the best performance. Its pretty dumb to do this just to half ass it unless you are doing XOC

 

A lot of time TIM like that is used for Liquid nitrogen OC and such since LM and stuff performs muuuch worse at colder temps.

[OhioYJ]

The last delid I did was on a 7700k, I tried a couple different types of liquid metal, an all copper IHS, and just using thermal paste rather than liquid metal. I tried Kryonaut, and Arctic Silver 5 between the die and IHS. I thought I had the results of everything some where, but so far I've just found the two types of liquid metals. From what I remember I ended up with liquid metal, because the thermal paste was no better than stock.

 

If you are going to delid you need to use liquid metal. If you aren't going to do that, I would leave it alone.

[1982 Original]

I am only interested in LM alternative because of the non-conductive properties and I’m not a fan of how it stains the surfaces.  It’s not that I’m trying to do a “half-assed” job. I will likely use LM, but was wondering how necessary it was to get decent results or if someone had a non-conductive compound on a delid that worked well. 

[OhioYJ]

2 minutes ago, 1982 Original said:

I’m not a fan of how it stains the surfaces.

I'm not either. This was my 7700k freshly delidded, no marks (clean die):

 

 

One week later I took it apart to switch from Liquid Ultra to Conductonaut, this is as clean as my die would come, it actually appears scarred / corroded after only a week:

 

 

The copper IHS I believe was part of that, so I swapped back to the stock nickel plated IHS. Used the Conductonaut, and it's been sealed up for over 4 years now. That PC my son uses it now over at my in laws after school and such. Still no temp problems. Still never needed to be serviced. I'm not touching it until it has a temp issue.

[RONOTHAN##]

30 minutes ago, 1982 Original said:

Can a standard TIM be used?

It can, but generally it will be worse than just having the IHS soldered like the 12700K is at stock. 

 

Delidding those chips doesn't make sense unless you plan on doing liquid metal and doing heavy overclocking at ambient. If you're not planning on doing either of those, there's no point. 

[1982 Original]

2 minutes ago, RONOTHAN## said:

It can, but generally it will be worse than just having the IHS soldered like the 12700K is at stock. 

 

Delidding those chips doesn't make sense unless you plan on doing liquid metal and doing heavy overclocking at ambient. If you're not planning on doing either of those, there's no point. 

For me, it’s something fun to do. I got my 12900ks to 5.4 already, also using the contact frame. I’ve never done a delid or a lap, so I want to experiment with that world. I’m also not rubbing nickels together. If I wreck something, I’ll just pickup a 13th gen anyway. 

[Beerzerker]

11 minutes ago, RONOTHAN## said:

It can, but generally it will be worse than just having the IHS soldered like the 12700K is at stock. 

 

Delidding those chips doesn't make sense unless you plan on doing liquid metal and doing heavy overclocking at ambient. If you're not planning on doing either of those, there's no point. 

I must disagree - LM is not a requirement in any way for better temps from a delid and the results of it vs a good, standard TIM aren't that different enough to justify the potential risks of using it - At least to me they aren't. 
It does give better temps but how much improvement you're looking for is a big factor in what you'd do and go with anyway.

Like the OP I too don't like what it does and the fact it can be destructive to certain metals/alloys such as aluminum since it contains Gallium. Doesn't help that if your cooling setup happens to be made from a material it reacts to then you'll have problems down the line at some point. 
I'm not aware of any standard TIM's that will do that.

Here's an article that was done not too long ago about it.
Liquid Metal Thermal Paste Explained & The Dangers Explored - Digital Advisor

As for delidding with regular TIM I've done it with MX2/MX4 and the results were good, in fact my 7700K has been done that way and it runs just fine with MX4 under the lid.

[RONOTHAN##]

1 minute ago, Beerzerker said:

As for delidding with regular TIM I've done it with MX2/MX4 and the results were good, in fact my 7700K has been done that way and it runs just fine with MX4 under the lid.

That's the thing, you did it with a 7700K. The 7700K still used Intel's terrible thermal material, so basically any aftermarket paste was an upgrade. The 12700K, however, is a soldered IHS, so all of those pastes are just a straight downgrade compared to solder, and LM is only barely an improvement as well. 

[Beerzerker]

33 minutes ago, RONOTHAN## said:

That's the thing, you did it with a 7700K. The 7700K still used Intel's terrible thermal material, so basically any aftermarket paste was an upgrade. The 12700K, however, is a soldered IHS, so all of those pastes are just a straight downgrade compared to solder, and LM is only barely an improvement as well. 

That's another point you just made about the results of LM vs solder in terms of temp improvement - But at the same time there are exceptions.

I must say if a chip's lid is soldered on, there is a risk of destroying the chip in the process of delidding but at the same time it can be done.
A guy over at TPU did it with significant gains in their results and it's really a good read, this info coming from someone that's already done it is worth the read, esp to someone that doesn't know much, if anything about it.
Delid 12700K Results - Surprising | TechPowerUp Forums

Yes, my 7700K was a chip TIM'ed with Intel's brand of pigeon-poop TIM but that's not my point - My point is if you're going to suggest LM, that's fine but at least be sure to mention both, the good AND bad points about it so an informed choice is made by the one(s) who's asking about it.
That's all I'm doing here.

[1982 Original]

There is one guy on Twitch who I have seen delid 12th Gen multiple times without even heating it up first. The solder seems to let go just fine. I do plan to apply some heat prior myself. It also seems like the general consensus that a delid on 12th Gen is good for 8 - 10C. I’ll certainly share my findings later this week. 
 

5.4 thermal throttles with synthetic tests for me, gaming hovers 75 to low 80’s. I’m curious is the extra 8-10C will prevent that thermal throttle on Cinebench or get me to 5.5 while gaming. 

[Beerzerker]

Yes, be sure to share the results of your delid so others here can reference it along with any relevant details of how it went and so on.

[1982 Original]

Update: I can definitively say that standard TIM under the IHS is worse that the stock solder.  I used Noctua NT-H2 and I probably lost 10C on cooling over the soldered IHS.  I tested 5.2Ghz at 1.35v right before doing the delid.  It would top out at 89C after 10 minutes on Cinebench R23.  Those exact same settings after the delid thermal throttle on the first pass.  So, unless my application is terrible, which it could be, it's worse than before.  

 

I'm still waiting for the conductonaut to be delivered later this week.  I'm optimistic that will improve.  Linus saved about 10C when he delided his 12900K in the RockitCool video.  There's another guy on Twitch that I've seen delid several 12th gen processors live and they netted about 10C using the RockitCool delid kit. 

 

The delid itself was super easy.  No issues there.

 

Update: Using liquid metal did lower temperatures 10C over the stock solder.  I tested again today after applying LM and letting it cure overnight.  5.2Ghz all core, 1.35v, topped out at 79C on a 10 minute loop of R23.  Exactly 10C reduction.  It also pulled about 10-15 less watts as well, even though the voltage and core clock was the same.  A guy I follow noticed the same thing.  He thought that the lower temperature could be the reason why it pulled slightly less wattage.

Edited November 12, 2022 by 1982 Original
Updating results.