whats the best liquid metal for delid?

[Bananasplit_00]

just need to know what to throw under the IHS when i delid my I7 4790K basically. no clue whats good and whats not, found LiquidUltra for ~$10 with shipping but idk how good that it compared to other liquid metal. will either use more of what i put under the IHS on the IHS or Hydronaut, havent decided tbh. thanks for the help  

[PCGuy_5960]

@done12many2 @MageTank @Lays Should be able to help you. AFAIK, Coollaboratory Liquid Ultra and Thermal Grizzly Condactonaut are very good.

[Groundedbeef]

i used Liquid Ultra on my 7700k  delid, seems to be working fine, less is more, its weird the first time you spread it on the die.

[For Science!]

19 minutes ago, Bananasplit_00 said:

-snip-

I used conductonaut, good stuff.

[MageTank]

59 minutes ago, Bananasplit_00 said:

just need to know what to throw under the IHS when i delid my I7 4790K basically. no clue whats good and whats not, found LiquidUltra for ~$10 with shipping but idk how good that it compared to other liquid metal. will either use more of what i put under the IHS on the IHS or Hydronaut, havent decided tbh. thanks for the help  

Performance wise, they are all nearly identical as far as LM goes. CLU is easier to deal with, Conductonaut is far more fluid, so it takes a little extra care to use, but CLU tends to leave a much more difficult stain once it's used. I'd say CLU is fine for bare dies under the IHS of a CPU, but conductonaut might be better for GPU's, since it doesn't leave an abrasive stain on your heatsink.

[Bananasplit_00]

1 minute ago, MageTank said:

Performance wise, they are all nearly identical as far as LM goes. CLU is easier to deal with, Conductonaut is far more fluid, so it takes a little extra care to use, but CLU tends to leave a much more difficult stain once it's used. I'd say CLU is fine for bare dies under the IHS of a CPU, but conductonaut might be better for GPU's, since it doesn't leave an abrasive stain on your heatsink.

well liquid ultra between IHS and copper CPU block then? bad or good idea?

[Daniel644]

1 hour ago, Bananasplit_00 said:

well liquid ultra between IHS and copper CPU block then? bad or good idea?

from what I've seen in videos online, 99% of the performance improvement when delidding comes from the liquid metal between the Die and IHS, from IHS to cooler, it's not worth the cost or concern of any aluminum being in the cooler, just use a good regular thermal paste there.

[MageTank]

2 hours ago, Bananasplit_00 said:

well liquid ultra between IHS and copper CPU block then? bad or good idea?

Not worth it. LM is expensive, and the difference between LM and traditional pastes between the heatsink and CPU IHS is negligible. Maybe 1-2C tops. Not worth the heavy cost, and potential disaster of you use too much. Gallium bonds to gold, would hate for it to run off and end up in the pins on your board. 

[Th3pwn3r]

1 hour ago, MageTank said:

Not worth it. LM is expensive, and the difference between LM and traditional pastes between the heatsink and CPU IHS is negligible. Maybe 1-2C tops. Not worth the heavy cost, and potential disaster of you use too much. Gallium bonds to gold, would hate for it to run off and end up in the pins on your board. 

$20.00 is expensive? A drop in 1-2 degrees is definitely worth it if you're already in the process. Everything is a potential disaster, depends on how careless you are though.

 

[MageTank]

Just now, Th3pwn3r said:

$20.00 is expensive? A drop in 1-2 degrees is definitely worth it if you're already in the process. Everything is a potential disaster, depends on how careless you are though.

 

$20 a tube is expensive when you are using nearly all of it to spread over top of your IHS. Gallium bonds best with itself, so you will need a thin layer on both the IHS AND your heatsink. Given how large IHS's are, that's certainly a very expensive paste application for 1-2C at best. Also, worth is subjective. While 1-2C difference is worth it to you, it's certainly not worth it for me. Especially once you factor in maintenance. I swap coolers quite often when swapping out CPU's and coolers for various tests, and spending $15-$20 a tube per swap, that adds up. A $10 tube of Gelid lasts me several applications, and performs nearly identical to liquid metal in regards to IHS->Heatsink. 

 

Also, being careless has very little to do with it. When you place it on your IHS, in between the IHS and heatsink, you run the risk of your coolers mounting pressure forcing the liquid metal out of the sides of the CPU. Not only that, but even cleaning LM can be risky if you accidentally smear it into a contact or pin. Sure, you can cover contacts with kapton tape or liquid electrical tape, but good luck doing either of that for your motherboards pins. 

 

So the rundown of your options are: use a cheap silicone paste for roughly the same performance (+/- 1-2C), or invest in $15-$20 worth of liquid metal, some kapton tape or liquid electrical tape, ruin the resale value of your copper heatsinks by staining the surface with a silver/gray color, and run the risk of damaging hardware for that extra 1-2C difference in thermals. Yeah... no thanks. 

[Th3pwn3r]

If you GLOB too much paste of any sort on the coolers pressure is going to cause problems, that's a given. That would be carelessness and not erring on the side of caution when applying the material but you can argue that if you'd like.

 

You mentioned some of the risks that many of us are willing to take. Worst comes to worst, those of us that do it and have a disaster have to buy another processor, it's not the end of the world. If it were a $1,000+ processor then I would be more hesitant. There are risks and there are rewards, it's nothing new in the world. Sometimes it's high risk/high reward, sometimes it's high risk/little reward. You have valid points that people should consider, personally, I'm going to squeeze every bit of performance out. If you're testing using different materials, cheaper materials, better materials or what have you it doesn't really matter as long as SOME of the variables remain constants. For your purposes I guess it wouldn't make sense if you're changing things more than your underwear

[Bananasplit_00]

6 hours ago, Th3pwn3r said:

If you GLOB too much paste of any sort on the coolers pressure is going to cause problems, that's a given. That would be carelessness and not erring on the side of caution when applying the material but you can argue that if you'd like.

 

You mentioned some of the risks that many of us are willing to take. Worst comes to worst, those of us that do it and have a disaster have to buy another processor, it's not the end of the world. If it were a $1,000+ processor then I would be more hesitant. There are risks and there are rewards, it's nothing new in the world. Sometimes it's high risk/high reward, sometimes it's high risk/little reward. You have valid points that people should consider, personally, I'm going to squeeze every bit of performance out. If you're testing using different materials, cheaper materials, better materials or what have you it doesn't really matter as long as SOME of the variables remain constants. For your purposes I guess it wouldn't make sense if you're changing things more than your underwear

i think you are greatly overestimating the economical power of a highschool student here. i cant just go and buy another late batch 5GHz+ capable I7 4790K and Z97 motherboard just becasue i rekt them with LM

8 hours ago, MageTank said:

Not worth it. LM is expensive, and the difference between LM and traditional pastes between the heatsink and CPU IHS is negligible. Maybe 1-2C tops. Not worth the heavy cost, and potential disaster of you use too much. Gallium bonds to gold, would hate for it to run off and end up in the pins on your board. 

thanks for the help  il just use NT-H1 or Hydronaut from IHS to waterblock

[MageTank]

16 hours ago, Th3pwn3r said:

If you GLOB too much paste of any sort on the coolers pressure is going to cause problems, that's a given. That would be carelessness and not erring on the side of caution when applying the material but you can argue that if you'd like.

 

You mentioned some of the risks that many of us are willing to take. Worst comes to worst, those of us that do it and have a disaster have to buy another processor, it's not the end of the world. If it were a $1,000+ processor then I would be more hesitant. There are risks and there are rewards, it's nothing new in the world. Sometimes it's high risk/high reward, sometimes it's high risk/little reward. You have valid points that people should consider, personally, I'm going to squeeze every bit of performance out. If you're testing using different materials, cheaper materials, better materials or what have you it doesn't really matter as long as SOME of the variables remain constants. For your purposes I guess it wouldn't make sense if you're changing things more than your underwear

It doesn't make sense from any practical stance though. IHS's are not perfectly flat, neither are heatsinks. It's different with bare die, which is essentially flat glass. Go look at any Asetek designed water block, and you will notice it's convex. Look at any of Intel's IHS's, and you will notice the center is slightly concave. Using a thin layer of LM on both sides, only works after you lap both to a completely flat finish, and even then, the risk becomes greater since you would need to use even less LM to prevent leakage.

 

It has nothing to do with being careless, and everything to do with your coolers required mounting pressure being too much to prevent leakage. You either end up using too little of a paste to make zero difference (or cause worse temperatures), or you use too much, and run the risk of the gallium bonding with any gold contacts or pins in the future. This is especially problematic with Conductonaut, which tends to ball up more so than CLU, which retains a liquid state far more often, but the risk is still there with all LM pastes.

 

Until you've actually done all of the work required to make LM work between an IHS and heatsink, I don't think you can fully understand the impracticality of it. I've seen people use it for naked die mounting after some serious lapping of their AIO's cold plate, and see decent results, but outside of that, it's just not practical at all. @Lays and @done12many2 can probably chime in here, as they too have probably attempted it at some point, and arrived at the exact same conclusion. These are people that consider themselves enthusiasts, and if they consider it pointless, it's safe to assume that it is. 

 

I'll also remind everyone that CLU eventually stains and creates a sand-like texture on the things it touches. It looks like this: 

 

I tried completely sanding a heatsink that had CLU on it, and was still unable to remove the sand-like texture. It's permanent at that point. The gallium essentially permeates and bonds with the copper itself. If your heatsink is plated with Nickel over top of copper, you might be able to completely sand the nickel coating off, but you are still going to end up with the stain on your IHS. To some, this won't matter at all, but others might see this as a big deal if they intend to resell any of these components in the future, and have to try to explain why this metal looking crust is stuck to the CPU and won't come off.

 

10 hours ago, Bananasplit_00 said:

i think you are greatly overestimating the economical power of a highschool student here. i cant just go and buy another late batch 5GHz+ capable I7 4790K and Z97 motherboard just becasue i rekt them with LM

thanks for the help  il just use NT-H1 or Hydronaut from IHS to waterblock

Honestly, a lot of pastes perform nearly identical to each other when it comes to filling the gap between the IHS and heatsink. I personally use Gelid, but even Arctic Silver 5 still works fine. http://www.tomshardware.com/reviews/thermal-paste-comparison,5108-8.html

 

As you can see here, the difference between Kryonaut and AS5 is only 2C. The difference between the best and worst pastes are only about 4C. 4C is still a decent difference, but not something that will completely make or break a build. 

[Bananasplit_00]

13 minutes ago, MageTank said:

It doesn't make sense from any practical stance though. IHS's are not perfectly flat, neither are heatsinks. It's different with bare die, which is essentially flat glass. Go look at any Asetek designed water block, and you will notice it's convex. Look at any of Intel's IHS's, and you will notice the center is slightly concave. Using a thin layer of LM on both sides, only works after you lap both to a completely flat finish, and even then, the risk becomes greater since you would need to use even less LM to prevent leakage.

 

It has nothing to do with being careless, and everything to do with your coolers required mounting pressure being too much to prevent leakage. You either end up using too little of a paste to make zero difference (or cause worse temperatures), or you use too much, and run the risk of the gallium bonding with any gold contacts or pins in the future. This is especially problematic with Conductonaut, which tends to ball up more so than CLU, which retains a liquid state far more often, but the risk is still there with all LM pastes.

 

Until you've actually done all of the work required to make LM work between an IHS and heatsink, I don't think you can fully understand the impracticality of it. I've seen people use it for naked die mounting after some serious lapping of their AIO's cold plate, and see decent results, but outside of that, it's just not practical at all. @Lays and @done12many2 can probably chime in here, as they too have probably attempted it at some point, and arrived at the exact same conclusion. These are people that consider themselves enthusiasts, and if they consider it pointless, it's safe to assume that it is. 

 

I'll also remind everyone that CLU eventually stains and creates a sand-like texture on the things it touches. It looks like this: 

 

 

I tried completely sanding a heatsink that had CLU on it, and was still unable to remove the sand-like texture. It's permanent at that point. The gallium essentially permeates and bonds with the copper itself. If your heatsink is plated with Nickel over top of copper, you might be able to completely sand the nickel coating off, but you are still going to end up with the stain on your IHS. To some, this won't matter at all, but others might see this as a big deal if they intend to resell any of these components in the future, and have to try to explain why this metal looking crust is stuck to the CPU and won't come off.

 

Honestly, a lot of pastes perform nearly identical to each other when it comes to filling the gap between the IHS and heatsink. I personally use Gelid, but even Arctic Silver 5 still works fine. http://www.tomshardware.com/reviews/thermal-paste-comparison,5108-8.html

 

As you can see here, the difference between Kryonaut and AS5 is only 2C. The difference between the best and worst pastes are only about 4C. 4C is still a decent difference, but not something that will completely make or break a build. 

huh NT-H1 beasts Hydronaut? nice because i have loads more NT-H1 then Hydronaut  thanks for the help, il go with Liquid Ultra from die to IHS and then NT-H1 from IHS to waterblock then  

[orbitalbuzzsaw]

Yeah...no. Just don't.

[Bananasplit_00]

15 minutes ago, orbitalbuzzsaw said:

Yeah...no. Just don't.

any reason not to use LM under the IHS?

[orbitalbuzzsaw]

4 hours ago, Bananasplit_00 said:

any reason not to use LM under the IHS?

Any reason to? Just use regular TIM.

[Lays]

6 hours ago, Bananasplit_00 said:

any reason not to use LM under the IHS?

Under the IHS is where LM does it's job, the speed/effectiveness at which it can transfer the heat to the IHS is much faster than any other :normal: pastes.

[iamdarkyoshi]

CLU did absolute wonders on my 6700K between the IHS and die. I needed like half the heatsink for the same cooling performance after deliding.

 

Also, if you unscrew the bottom screw on the CPU holddown and then close it, you can put the screw in, holding down the IHS without having to worry about sliding the IHS around

[Bananasplit_00]

6 hours ago, orbitalbuzzsaw said:

Any reason to? Just use regular TIM.

that performs worse then LM though. so i dont see why i should use some silicon paste like Linus did in his delid video and not notice an improvement instead of LM which should drop it 10+ C

3 hours ago, iamdarkyoshi said:

CLU did absolute wonders on my 6700K between the IHS and die. I needed like half the heatsink for the same cooling performance after deliding.

 

Also, if you unscrew the bottom screw on the CPU holddown and then close it, you can put the screw in, holding down the IHS without having to worry about sliding the IHS around

yah i have a pretty good idea of how to put the delided CPU back into the socket but thanks for the tip