What's a good therman grease? Are thermal pads any good?

D

Deadjasper

2[H]4U
Joined
Oct 28, 2001
Messages
2,685
Got a box running at 45c at idle. Me thinks it might be the cheap Chink no name thermal grease I used. The last name brand grease I bought was Thermal Grizzly and it worked great but the next time I ordered it I must have gotten a bad batch because it was impossible to spread out, all it did was roll up into a ball.
 
cyclone3d said:
Depends on the CPU and the cooler you are using. MX5 gives better temps.

MX5 is just as good if not better than Grizzly Kryonaut. I have tried both on the same setup.
Click to expand...
I can honestly say I saw a gain of a few degrees. Dare I say between 3-5 on 6900K 8core16thread.
 
cyclone3d said:
Depends on the CPU and the cooler you are using. MX5 gives better temps.

MX5 is just as good if not better than Grizzly Kryonaut. I have tried both on the same setup.
Click to expand...
My point, there are no ageing problems.
The paste wont dry or push out, a carbon fibre pad will work at its max capacity forever.
Important if you cant do the work yourself again or if doing it again causes damage.
And prevents loss of uptime or function due to this particular heat issue.
 
I know this may be a far-out kind of comment but since I haven't read on thermal paste for years my comments may be useless. However, I remember reading that some people in the past did not like using Arctic Silver because the heatsinks would become stuck or even destroy CPUs when replacing a heatsink. (I know some people play a game before removing a heatsink.) Anyway, I've used Artic Ceramique for years.
 
wandplus said:
I know this may be a far-out kind of comment but since I haven't read on thermal paste for years my comments may be useless. However, I remember reading that some people in the past did not like using Arctic Silver because the heatsinks would become stuck or even destroy CPUs when replacing a heatsink. (I know some people play a game before removing a heatsink.) Anyway, I've used Artic Ceramique for years.
Click to expand...
that can happen with basically any paste, not just as5, and if you follow amd's instructions of "warm the cpu and twist" you wont have an issue. actually had this "issue" yesterday with my 2600x, mx5, which gets quick thick and sticky after a few months, and my block. i let it sit too long after shutdown and it was quit solid but a good twist and she popped right off.
 
Nenu said:
My point, there are no ageing problems.
The paste wont dry or push out, a carbon fibre pad will work at its max capacity forever.
Important if you cant do the work yourself again or if doing it again causes damage.
And prevents loss of uptime or function due to this particular heat issue.
Click to expand...
I know. I also have a couple of the carbon fiber pads. If the cooler and IHS are not flat enough, they give pretty poor temps since the carbon fiber pads don't squish to fill in gaps. Of you have a lapped or flat from the factory cooler and a lapped CPU they work much better.
 
  • Like
Reactions: Nenu
like this
cyclone3d said:
I know. I also have a couple of the carbon fiber pads. If the cooler and IHS are not flat enough, they give pretty poor temps since the carbon fiber pads don't squish to fill in gaps. Of you have a lapped or flat from the factory cooler and a lapped CPU they work much better.
Click to expand...
I spose I wasnt trying to convince you, you have more experience than me.
I was stressing the point for those reading in case there is value for them.
I personally have no reason to use carbon fibre pads, my current favourite is Thermagic ZF Extreme paste, a bit stiff [though wont eject itself as easily over time], well worth the effort applying.
 
I've used an IC graphite thermal pad, the 40x40mm version for about 4 years now on 3 different heatsinks. Never had a problem with it, and it's even torn now. Still use it anyway. I have a good amount of things going on right now across my triple monitor setup, Spotify, RDP session, Firefox, Steam, Signal messenger.. and peeking at hwinfo shows 35C on my i9-11900K with a Noctua NH-D15 chromax.black. This chip hits and holds 5.3GHz on 7 of 8 cores simultaneously on the desktop just putzing around like this. You tell me if that's not good enough or not.
Load temps are pretty much amazing. Maybe 45C in games? Don't have an exact number offhand but believe me this setup is fine.

Otherwise, I'm pretty particular about my computer, and always preferred MX4. It's my favorite paste, I've seen it go 8+ years without needing a repaste. A lot of the pastes that are 'better' than Arctic Cooling's are mostly only better for the conditions they were formulated for, sub-ambient cooling. You'll be doing more repastes, and some don't mind, I do.
I wouldn't be unhappy if MX5 were on my CPU by any means, but I'd definitely reach for it on anything similar to a motherboard or GPU repaste. Use what you want.
 
To each their own. Thermal Grizzly Kryonaut PASTE.
Only problem is it's spendy.
 
Gattlin said:
To each their own. Thermal Grizzly Kryonaut PASTE.
Only problem is it's spendy.
Click to expand...
Also that it is not meant for long term stability due to the paste being pushed out so it should be re-applied at least once or twice per year. E.g. temps will increase by several degrees after a few months. The difference between a good and a great paste is minor so IMO something like kryonaut only makes sense if you have a high end cooling solution. If one is on air or aio then a cheaper paste of good quality like mx-4, nt-h1 etc. makes more sense. Kryonaut, nt-h2 etc. are for those who are chasing the last few degrees and have a cooling solution where it is very difficult to gain better temps.

For reference I am using kryonaut atm. but the difference to nt-h2 was minimal (less than 1 degree and close to margin of error) and maybe 2-3 degrees better than nt-h1 on fresh application. It has lost 2-3 degrees in temps in 4 months though so probably going back to nt-h2 or hydronaut when I do maintenance.
 
MX-5 is great. I used it on my CPU and GPU water block. It dropped my gpu temps by 3c over MX-4. I didn't pay attention to how it faired for my CPU unfortunately. There is no reason to use thermal pad unless you building a system that will not be maintained. Is it really such a burden to repaste a CPU every couple year?
 
Last edited:
If you run your CPU hot a lot, I found thermal Grizzly pumps out and dries up and your temperatures start to rise. It is very good other than this.
Noctua NT-H2 paste is very good. It is what I am using now. I have an aio and my idle temeperatures are < 1c above ambient room temperature.
 
what's a thermal paste that resists pump out. My mx-4, nh1, and hydronaut, were pumping out on this.
 
GeneO said:
If you run your CPU hot a lot, I found thermal Grizzly pumps out and dries up and your temperatures start to rise. It is very good other than this.
Noctua NT-H2 paste is very good. It is what I am using now. I have an aio and my idle temeperatures are < 1c above ambient room temperature.
Click to expand...
Only had the problem with kryonaut. Hydronaut was fine for many months with no change in temps.
 
GeneO said:
If you run your CPU hot a lot, I found thermal Grizzly pumps out and dries up and your temperatures start to rise. It is very good other than this.
Noctua NT-H2 paste is very good. It is what I am using now. I have an aio and my idle temeperatures are < 1c above ambient room temperature.
Click to expand...
The paste crowd has been split 50/50 for years on this. Some use pastes formulated for sub-ambient cooling like a lot of Thermal Grizzly pastes and maybe gain a celsius or two, but they do dry up. At risk of angering everyone, I believe it's just improper use based on a misreading of benchmarks and bad reviewers. Most people, who won't be posting here about it, have moved to just using Noctua's provided paste (which is great) or something like Arctic MX5. Even Arctic's lower end stuff is pretty good. I favor just using pastes that are formulated for ambient temp cooling if I'm doing ambient temp cooling. Nothing wrong with using the "better" pastes though, go for it. Just not enough gain to matter to me.

I'm getting average paste performance out of a IC Graphite thermal pad, maybe somewhere around Arctic Ceramique, and that's working well for me. My system holds 5.1Ghz on 8 cores during gaming and never blips, and runs around 53C at full load last I looked just last night. I'm not sure what my limiting factor is on my cooling, I'm guessing nothing. This particular chip runs at high voltage. An extra degree or two might extend my chip's life, but I doubt it. Limiting factor being just 14nm physics considering I run a NH-D15, but I don't really need it to be any better. Maxing out Intel Adaptive Boost as it's advertised is all I was shooting for with my build.
 
I read a decent appraisal of a 0.1mm 60C melt solder pad and the difficulties applying it correctly.
Failure to take enough care can result in harm to the CPU or mobo without precautions during the burn in period.
The burn in period needs precise temperature control, no 2nd chances or the pad is wasted.

The result is pretty good if you cover the bases:
About 2C better than the best pastes watercooled (probably larger delta with air), doesnt cause any metal damage, doesnt pump out after the initial burn in phase (any initial pump out you can protect from), will last forever until taken apart.
Worth a look if you take enough care
https://www.igorslab.de/en/liquid-m...-how-the-magic-burns-in-safely-goes-tutorial/

I'm interested to see what this can do on a de-lid, unfortunately not covered.

ps
put your English translated German head on, some things dont come out great.
 
Last edited:
I've been going between MX-4 and Grizzly -- hardly any difference between the two. In my experience, I prefer a paste that's going to last as opposed to one that will net me 1-2c cooler, so primarily use paste designed for ambient. The grizzly I've used in a build a while back with a peltier where the improvements were more noticeable.

Wasn't there an article in [H] a while back where they trialled a bunch of pastes and then something ridiculous e.g. mayonnaise, caulk? If I recall, the caulk did better than some of the lower-scoring pastes.
 
Nenu said:
My point, there are no ageing problems.
The paste wont dry or push out, a carbon fibre pad will work at its max capacity forever.
Important if you cant do the work yourself again or if doing it again causes damage.
And prevents loss of uptime or function due to this particular heat issue.
Click to expand...
My favorite thing to see when tech reviewers use the carbon pads is that they aren’t wearing gloves and putting their body oils all over the pad before slapping it down on the cpu. Gotta love YouTube hahaha
 
evhvis said:
Also that it is not meant for long term stability due to the paste being pushed out so it should be re-applied at least once or twice per year. E.g. temps will increase by several degrees after a few months. The difference between a good and a great paste is minor so IMO something like kryonaut only makes sense if you have a high end cooling solution. If one is on air or aio then a cheaper paste of good quality like mx-4, nt-h1 etc. makes more sense. Kryonaut, nt-h2 etc. are for those who are chasing the last few degrees and have a cooling solution where it is very difficult to gain better temps.

For reference I am using kryonaut atm. but the difference to nt-h2 was minimal (less than 1 degree and close to margin of error) and maybe 2-3 degrees better than nt-h1 on fresh application. It has lost 2-3 degrees in temps in 4 months though so probably going back to nt-h2 or hydronaut when I do maintenance.
Click to expand...
That’s interesting, I still have my original spread of kryonaut on my 2700x built in 2018 and my recorded temps have not gone up since. That makes it 3 years old at this point…I wonder if maybe you don’t have enough mounting force on your cpu. Maybe the tighter the gap the more friction there is to prevent the paste from pumping out under heat?
 
housecat said:
The paste crowd has been split 50/50 for years on this. Some use pastes formulated for sub-ambient cooling like a lot of Thermal Grizzly pastes and maybe gain a celsius or two, but they do dry up. At risk of angering everyone, I believe it's just improper use based on a misreading of benchmarks and bad reviewers. Most people, who won't be posting here about it, have moved to just using Noctua's provided paste (which is great) or something like Arctic MX5. Even Arctic's lower end stuff is pretty good. I favor just using pastes that are formulated for ambient temp cooling if I'm doing ambient temp cooling. Nothing wrong with using the "better" pastes though, go for it. Just not enough gain to matter to me.

I'm getting average paste performance out of a IC Graphite thermal pad, maybe somewhere around Arctic Ceramique, and that's working well for me. My system holds 5.1Ghz on 8 cores during gaming and never blips, and runs around 53C at full load last I looked just last night. I'm not sure what my limiting factor is on my cooling, I'm guessing nothing. This particular chip runs at high voltage. An extra degree or two might extend my chip's life, but I doubt it. Limiting factor being just 14nm physics considering I run a NH-D15, but I don't really need it to be any better. Maxing out Intel Adaptive Boost as it's advertised is all I was shooting for with my build.
Click to expand...
You are claiming 53C full load on air, 11900k with Adaptive boost switched on?

What is the "full load".

Is that with the unlimited turbo or with the standard turbo time limit of about 1 minute (and then it clocks down a lot).

How about Cinibench R23 multi-core or even encoding a video in handbrake?
 
You must log in or register to reply here.
Back
Top