Which Thermal Paste to buy and apply (Traditional and Liquid Metal)

ANYTHING is better than Kryonaut, even a cheap toothpaste would last longer than Kryonaut.

They have released a new Kryonaut Extreme now though which we have ordered to test and see how it does.

 

Thanks for testing. You should put the syringe in a plastic bag and put it in an cup of warm water before the application (make the paste softer to spread). And you should test it on the processor. The GPU is normally easy to cool even with middle quality thermal paste. The problem is the Cpu.

 

You made two mistakes here.
first, as I said several times, do NOT spread TFX. For direct die CPU's, apply a straight line going to the edges (e.g. BGA). For GPUs or IHS CPU's, apply two intersecting lines in an X pattern, going from edge to edge,
And if the surface is even larger or the X you made is a bit too thin, you can apply four small drops to each quadrant left over along the X. The pressure of the mounting will then spread it for you perfectly.

Second, after this application, use it a week like this then test. Results will be impressive.

 

Thanks. For seringue in a plastic bag and hot water ( i did test that with phobya nanogrease extreme) and could not spread it. Wasn't hot enough. Maybe i did not let it sit long enough. That's why i used the heat gun but next time i will be more patient and do it with plactice bag.

Also i just retested it and temps seems to be improving like @Falkentyne said it takes a week to improve.
For the CPU i use LM that's why i only tested on the GPU.

 

I knew you would find someting wrong with my procedure. lol. I am happy you did.

I tought for laptop we needed to spread. Next time i will follow your recommendation because this paste is almost unspreadable whitout heat. I wonder why tey supply a spreader with it.
I will retest this week but like you say it seems to be improving already. Thanks for your help [/QUOTE]

 

I tested TFX vs Kryonaut on both my ES 10900k + Liquid Freezer II 360 in Maximus 12 Extreme (ES board), in the case in a Define 7 XL, and Retail Z490 Aorus Master (supplied by Gigabyte) + my store bought newegg crappy retail 10900k, with both a NH-D15 in open bench, and an Eisbaer Extreme 280 AIO (open bench).

After a week of curing, TFX was about 2C better than Kryonaut on both setups

 

LOL
I knew you would say something like that. But seriously, kryonaut was there for only one week so it was at it's best performance.

 

That's impressive.
Would you say it's even better then Phobya nanogrease extreme?

 

Phobya had worse temps than Kryonaut for me and it also degraded the same way.

 

Kryonaut is very hard to beat for a non LM paste when it is freshly applied. Some have reported that that it degrades over time (i believe them) but i still use it from time to time just for testing purpose or to compare. The fact that Thermalright TFX could perform just a little better (after fresh repaste) a little better then fresh kryonaut is impressive. Like @Papusan said it would be more conclusive if i tested on the CPU(he is right again) instead of the GPu but the fact that it performs at par and better on a GPU is impressive
I too have experienced that Phobya nanogrease did not always perform as well as kryonaut. I found that phobya nanogrease performs at his best when there is very thin layer using the spread method (not the X). Maybe i am wrong on this one too? not sure.
I will edit my results in a week just to see if it actually improve over a week (i hope so). Even if it would not, i would still be happy with the results. 1 or 2C less on the GPU could translate no thermal throttling while gaming for my system.
Thanks for your help again @Falkentyne. You are my reference when i need advise for repasting in the best possible way.

 

Just to make it clear, I never saw premature degradation with Kryonaut on desktop high pressure heatsinks. If there was it may have been like 2C over a few months. But I'm sure some people saw it.
The premature degradation with Kryonaut happens with laptops, where you get skyrocketing temps after just a few days (with bad core temp deltas in addition to that). I saw that happen with both Kryonaut and Phobya. And initially when switching to liquid metal, it happened with LM also--namely it turning as hard as a rock in a few weeks, and that took a lot of work to address.

With LM on a desktop on top of the IHS and a high pressure mounting heatsink, with some extra prep work (foam dams, etc), this never happened at all.

 

Hello,
I've purchased MSI GE66 I7-10875H (8core) recently and decided to repaste it with MX4 which is recommended all over the internet, at first temperatures were few degrees cooler than stock thermal paste and was able to maintain higher turbo boost before throttling... however after two weeks temperatures are far worse, I've repasted it again alas, after another two weeks temperatures are again high and can't maintain high turbo boost.

After encountering same issue two times in a row I decided to search the internet and looks like it's the famous pump-out...
Read all this thread, but still can't find any conclusive information, what's regarded the "best" long lasting thermal paste with good thermal conductivity. (Don't want to risk it with LM while laptop is under warranty)
I've ordered TG Kryonaut and Thermalright TFX for testing, hope one of them lasts longer than two weeks with MX4.

 

For me personally, I've never had a problem with Kryonaut pumping out in laptops. It works great, but my application on the CPU keeps burning up every 3 weeks. It degreades extremely quickly when exposed to temperatures above 176°F (80°C). That is why I switched to Phobya Nanogrease on the CPU side. I haven't repasted the GPU in 3 years and temps are still the same as day 1.

At first I thought I just got a crappy batch of Nanogrease paste, but I decided to perform an experiment to save myself the trouble of adding yet another paste to my collection. I have 2 CPU heatsinks for my Ranger. One I know is flat, and the other had pump out issues with the paste. With the latter heatsink, the paste would pump out after a week. I decided to swap the heatsinks for testing, and the paste has not pumped out yet on the first heatsink (after almost a month), leading me to believe the second one is warped.

In conclusion, no matter what paste you use, it will eventually pump out if your heatsink is warped, even on desktops. A solution to this would be to sand down your heatsink until it's flat, although you'd have to look really closely to see what regions are higher than others.

For me, Nanogrease performs better than Kryonaut by a little, but not much. There's a new Kryonaut extreme paste that looks like it has properties similar to Thermalright TFX. TFX is an extremely viscous paste, so it should be very resistant to pumping out.

 

Not sure about Thermalright pastes but MX4 is the worst paste for laptop and thermal grizzly is bad for longevity as well.

The best paste I've used so far is New mastergel maker. The one that comes in a flat tube. Don't get old mastergel maker, that one sucks.

 

Well well well, looks like Asus IS using a nickel plated coldplate for their liquid metalled laptops.



That spill though

 

Foam dams

 

I'm currently testing several thermal pastes on my Dell 7506. The cooling setup on this device isn't great on this, the hottest core can reach over 90 degrees in Prime95 even without Turbo. The good thing is there is no thermal/power throttling on this device with disabled CPU Turbo, it holds 28W in Prime95 but runs hot. It's good for testing.

Thermalright TFX is the best so far followed by Thermal Grizzly Kryonaut and Coolermaster MasterGel Maker Nano. On the hottest core TFX is about 2-3 degrees cooler. Also tested Phobya NanoGrease Extreme and Gelid GC-Extreme but for some reason they don't work on this laptop, I tried several times but they are really bad even the MX-4 is better. I will check some more paste in the next days and weeks.

 

On Dell 5570 I did notice Cooler Master makergel nano had worse temps than stock toothpaste dell applied. I re-did it and applied light coat on both the die and heatsink and that did it.
Cooler master maker gel nano has a new version that's like cotton candy and hard to spread. With this consistency, coating die and heatsink simply gave outrageous temps and had to use cotton earbuds to flatten the paste literally until a thin coat was formed. Took too much time though to figure it out!

 

Actually the Cooler Master Makergel Nano I have is from 2017 and easy to spread, the other pastes are brand new (except from Thermalright CF III). No idea if Coolermaster changed something to the worse but might explain something.

 

Yes the new mastergel maker is more viscous compared to the older one but its a change for the better in my testing, longevity has improved and temps are also slightly better than the old one.

I think you guys are getting poor temps due to insufficient mounting pressure, I always provide extra pressure for 10 minutes before reassembling the laptop to make sure the paste is squished thin. When mounted right mastergel maker is comparable to kryonaut thermally but with better longevity.

 

I can't complain about Coolermaster, I mean from 9 different paste it's the second best together with Thermal Grizzly Kryonaut in my testing. TFX probably benefits from the thickness in my case, in a traditional test with a desktop CPU or CPU with heatspreader the others might do better.

 

what is that your using in place of the thermal pads and did you notice any improvement using it?

 

That looks like K5 Pro to me.

 

New CM nano is like Cotton candy.
Repasted it on AW BGA temps are same as old paste 28-32c on idle AC/battery, 65C-70C on max load (CPU) and 71C (GPU max load). Did test it with ambient temps of 27c.

 

Phobya NanoGrease Extreme 16W/mK
Alphacool Subzero 16W/mK

I have both and I can say they are identical, it's the same.

 

Why wouldn't they be the same?
They are the same company. They just changed the name for marketing.

 

I didn't know this! I tested both and it was immediately clear to me they are the same.

 

I brought the $699 Gateway from Walmart which is pretty much a rebranded tong fang build. With a i5 10300h and RTX 2060 (115W but capped at 90W). I brought MX-4 and repasted but haven’t seen any decrease in temperatures, I’m suspecting it also has to do with the pressure from the heatsink not being adequate enough. There is only 1 spring loaded screw on top of the cpu where as the others (7 other screws) just hold the heatsink in place.

My question is, does anyone know if the thickness of Phobya NanoGrease Extreme be able to overcome the lack of pressure and create a better seal between the CPU die and the heatsink?

I am constantly hitting thermal limits while the CPU boosts even with undervolting to the maximum stable voltage.

The GPU stays cool at approximately 76-78*C.

Thanks

 

Well for some Phobya NanoGrease Extreme seem to work, for me it didn't (subpar laptop cooling with low pressure). Also I wouldn't say it's a thick paste. I would choose either Thermalright TFX or Coolermaster MasterGel Maker Nano which are both thicker. TFX is the thickest and hard to spread. The Coolermaster is a much easier one and when I use a thicker layer of paste than I usually do the temps are quite close to TFX. Both are like 5 degrees better than the Phobya NanoGrease on the hottest core in Prime95.

 

Looks like 3 screws on the heatsink over the CPU, 1 spring-loaded and 2 applying tension on the retaining arms.

 

Thank you. I'll check out both and decide which ever is more cost effective.

Ah I did not realize that those two are applying tension. After I repasted and threw everything back together I felt as if the heatsink just sat on top. However, if I press and apply some downward pressure on the heatsink right on top of the screw hole where the springload one sits, I can hear and feel the heatsink pushing down. In this case, not sure if I'm flexing the motherboard or if the heatsink is actually being pressed upon.
Thank you for pointning this out.

 

I'm a Thermalright TFX fanboy. This is the best paste ever. I think it's better than Phobya Nanogrease Extreme (alphacool subzero). And TM's w/mk for it is pretty accurate. Nanogrease extreme is nowhere near 16 w/mk. I found it to be worse than Kryonaut in a direct comparison.

TFX seems to have very long durability:

https://forums.guru3d.com/threads/t...ivity-of-14-2-w-mk.435511/page-2#post-5861772

 

No problem. And this is why I advise applying some downward pressure on the heatsink when screwing it in.

 

Hi everyone, I just finished a tube of GC-Extreme and am looking for something with more longevity. I see TFX and New MasterGel Maker are modern recommendations. Is either of these any better than the other price differences aside?

Thanks! (I mostly re-paste MacBooks, don't know whether that affects the recommendation any insofar as heatsink or clamp quality.)

 

The last time I opened up the laptop to repaste, I made sure I applied downward pressure on the heatsink when screwing it and also downward pressure on the screws, so much that I almost stripped the cheap screws used in here. I will purchase a better thermal paste from Micro Center today if I can find some of the recommended ones in this thread.

I just ran some AIDA64's system stability test and the thing is, without GPU usage, the CPU sits at 75-80*C with turbo boost at 3.5GHz.
With the GPU at 100% usage, it will raise up to high 80s, occasionally hitting 94*C before throttling down. Without the GPU being used, the CPU uses about 37-40W of power.

With the GPU in use - the CPU pulls in about 28W. The core temperatures are also in range with the research I've done. Core 1 and 3 runs cooler at about 4-6*C than the rest of the cores.

Now here's my theory, the heatsink is saturated with heat from both the GPU and CPU and is not cooling fast enough even with 'Turbo Fan' on mode.
With the GPU usage at 100*C, the CPU hits and stays at the upper 80s C, while GPU is at no usage, the CPU tends to run a lot cooler at 75-80*C.

Do you think this theory is correct or even somewhat plausible?

 

Have you undervolted the Cpu? If not do that. As well uncheck the BD-PROCHOT box in ThrottleStop to see if throttling point for the Cpu can be raised.

 

Prolimatech PK-3 and Noctua NT-HT2 are not that bad on my laptop, however they can't match the TFX and Coolermaster. So for me...

1. Thermalright TFX/Coolermaster MasterGel Maker (2017)
2. Thermal Grizzly Kryonaut/Thermalright TF8/Noctua NT-HT2/Prolimatech PK-3
3. EC 360 Ruby/IONZ IZP14/Alphacool Subzero/Phobya NanoGrease Extreme
4. Gelid GC-Extreme/Arctic MX4/Thermalright CF III

I see only two potential thermal paste left who could match or beat the TFX/Coolermaster on my laptop, it's the Kingpin KPx and Thermagic ZF-EX. Unfortunately both are not directly available in my country.

 

Yes I have. That's one of the first things I've did plus another -50mV in the bios.

Attached a screenshot of my throttlestop screen. I've been stable on -125mV on the core and -100mV on the cache. I haven' touched the System Agent or IccMax settings since I don't think those are really necessary and haven't seen any change while adjusting those.

 

ZF-EX is apparently repackaged TFX. I have some of it and have it on my 3090 FE and another CPU.

 

And same performance as TFX? I thought it's a slightly improved TFX because it's rated 14.6W/mk instead of 14.3W/mk. In this test performance differs over 90 degrees:

 

That's impossible. I saw that video. That was intentional.

I've tested them both and they feel exactly the same. I tested them on a 10900k and they were 1C apart, 89C TFX (cured, after 4 months) vs 90C ZF-EX (probably 89C if I had let it cure more than a few days) running AVX small FFT prime95

Same temps on 3090 Founder's Edition.
Both hopeless to spread. Both feel like putty after you let the paste cure a week and then remove the heatsink. Both seem to drop a C or two after curing.

And this reply today:

https://forums.guru3d.com/threads/t...ivity-of-14-2-w-mk.435511/page-2#post-5873484

They are the exact same paste. Don't believe marketing, please.
Of course you can buy it yourself and test it (I have more TFX coming tomorrow, and I have one unopened tube of ZF-EX left).

https://www.aliexpress.com/item/4001065767770.html

Nan's Gaming Gear is out of 6.2g tubes of TFX but there's here for more high robbery prices.
https://www.amazon.com/dp/B08CGPD9RN/

 

No if they are the same I don't have to try. Kingpin KPx is a very thick one, so it might be a good choice for laptops as well. I might also have to try a new Coolermaster MasterGel Maker to see if there are differences compared to my 2017 version.

 

Really? Cuz It's less than half the price of TFX

 

Well you can't really compare china prices with USA prices.

On Aliexpress, 3x2 for 6 grams of ZF-EX is $24 ($8 per package).
6.2 grams of TFX is $23. 2 grams is about $12

https://www.aliexpress.com/item/1005001411781566.html
https://www.aliexpress.com/item/32955394312.html

It's probably just the same paste factory supplying them. I've used them both. They are literally identical in every way.

 

I just ordered a 2g tube of both to test. ZF-EX for 5 bucks and TFX for 10 bucks

 

Hi, can you link the sellers you used? The quoted links are more expensive. Thank you.

 

Local links so I'm not sure if they ship internationally

https://shopee.com.my/product/95940644/2329199585?smtt=0.116877937-1609659668.9

https://shopee.com.my/product/296623518/3662242919?smtt=0.116877937-1609659731.9

 

If anyone is interested, here I share my experience from CM Mastergel Maker (old version - if there's any difference between that and the "new" flat version - doubt it) and the Phobya Nanongrease Extreme:

I own two Laptops, an HP Envy x360 with a Ryzen 2500u and a Lenovo Y740 with the 9750h and 2080 mq.

Initially I repasted my Envy -even though it never had thermal issues as its a 15w CPU that boosts for a couple of minutes to 25w- with the MasterGel and kept it on there for over a year with no drop in performance. Could have kept it on there for years I assume. That said I only could get it near 80c using stress tests in performance mode.

So I thought "great paste" and applied it to my Y740 when I got it back in march last year (spread method).

Only after a week or two the paste degraded and when loading the CPU it spiked immediately over 90c. I thought I did something wrong, even though I repasted many laptops before. When I removed the heatsink the paste looked like it boiled away. It didnt push out but it just became basically transparent as if only the silikon in it was left.

I applied it again and the same thing happed after only a few short stress tests.

Then I tried the NanoGrease which lasted longer, about 15 hours of AC Odyssey. Everybody knows how taxing it is to the CPU. In performance mode it was almost always sitting at its max of 50 watts TDP. Then the same happened again.

While until now I tried to spread both pastes as thin as possible, I then applied the Nanogrease a bit thicker.
Now after more than 40 hours of Cyberpunk 2077 temps didnt change at all since applying it freshly. Even after the CPU spiked to 99c playing the initial release of the game which I blame the heavy use of AVX which they removed after the first or second update.

Conclusion:
1. I don't think the cold viscosity matters too much but more how the paste behaves under high temperatures. Still, higher viscosity seem to have an advantage there.
2. Longevity of the paste depends of the CPU temperatures . I think below 80c the (low temperature)viscosity doesnt matter too much if its not too runny. There has never been a temperature difference on the GPU side which runs much cooler (much bigger die).
2. Thin application doesnt always mean better since the excess pushes out as soon as it gets heated up (maybe then creates a barrier around the die and prevents the vanishing of the TIM between die and heatsink?)

Maybe one part comes into play here, but could be a coincidence:
After I repasted the laptop the 2nd time with the Phobya, I have let it heat up for a few seconds to check if the application was correct and then didnt use it for a few weeks. Maybe that somehow helped to settle and "dry" the paste in the correct places, who knows.

So, big post for a little paste geekyness. At least you may had a good read

Cheers

 

How are the initial temps with Phobya Nanongrease Extreme on your Y740? Initial temps are 5-6 degrees down to TFX/Mastergel on my laptop, this is no option for me no matter if it won't dry out or not. Also I do believe the TFX won't dry out either because of its high viscosity. A thin application can be worse actually, I have to use a bigger layer than I usually do with the Coolermaster which helps.

 

I havent noticed any (obvious) initial difference between the CM and the Phobya. The latter just lasted while the CM didnt. I dont know if the CM lasts longer on a thicker application.

I've never had issues with paste drying out other than with stock pastes.
Common sense also tells me that the higher the viscosity of a paste is, the more its prone to dry out. Stock pastes for example are very high viscosity which makes them dry out quicker then wetter/more liquid ones.