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

The old CM Mastergel maker didn't last long for me either. Temps started to climb 9 months in.

The new version is much improved in that regard. Harder to spread though.

 

I tried the new Coollaboratory Liquid Extreme and hated it. It’s like they purposefully tried to copy conductonaut and all it’s flaws. It’s annoying to try and spread compared to liquid ultra, so I’m going to stick to my tried and true TIM.

 

Did you get your paste?

 

Yup, testing now. They do indeed look the same.

I've only used the ZF-EX for now and it looks and spreads like old mastergel maker. Not nearly as viscous as i was expecting.

 

You didn't do the good old finger test?
TFX and ZF-EX have exactly the same thickness ?

 

Haven't applied TFX yet but it looks the same out of the tube at least

 

Even if they are not the same, it must be a slightly tweaked TFX (slightly higher W/mK), so it's no surprise they look the same. The real question would be if they also perform the same.

edit:

Arctic MX-5 will be released in March: https://www.amazon.co.uk/dp/B08T64M68V

Hopefully it performs better than MX-4.

 

Yeah, hopefully it doesn't dissapoint like kryonaut extreme

 

I've tested Kyronaut Extreme by the way, in the first test it looked almost identical to the non Extreme version. Somehow expected but on a notebook CPU I was hoping it could perform differently. They are much better than the MX-4 of course.

 

Yeah any high-end paste will be better than MX4 nowadays. GD900 performs identically to MX4 and its 2 bucks for 30g of that stuff

 

Did you buy ZF-EX from that seller? Im tempted to buy but only 3 sold and 2 reviews....

 

I bought it from https://www.aliexpress.com/item/4001065680665.html but I would not do it again...the shipping speed was the slowest shipping in the world but at least it came.

So I gave you https://www.aliexpress.com/item/1005001411781566.html instead.
I mean you get money back guarantee anyway so just buy it. Slow boat is slow boat. You spend thousands of dollars on BGA throttlebooks...don't tell me you're complaining about $25 dollars now.

 

How long did it take? Was it during december?
Im not complaining about the price (5,5€ btw) but rather buy from someone with more items sold if I can

 

Just used TFX and it spreads really well. Looks similar to ZF-EX but feels and spreads differently. Where ZF-EX was sticky like other pastes TFX feels sort of like kinetic sand.

Preliminary temp results show it beating the mastergel maker by 2C, maybe because i was able to make the application really thin since it spreads so well.

 

Just recieved new KPX and the claims were true, it's super soft and liquidy now

 

Wait what claims?
does that mean it's bad now?

 

That KPX is no longer super thick, some guy on discord emailed kingpin and he confirmed that it's thinner now



Not sure how this affects temps and longevity which is what I plan to test.

 

Ok, good to know about that change (or bad to know).
Anyway, I just tested two syringes of TFX with one syringe of ZF-EX.

They are exactly the same paste.
I rubbed it on a countertop then smashed and spread it with my fingers and looked very carefully at it.
They are 100% the same.

Any change in consistency that is in the tube (dry or wet) depends on how dried out the paste is.
The tube of TFX I got from Nan's Gaming Gear is the wettest, same consistency as the other two tubes I bought from Nan's in the past. It cures after a week of use and gets a lot dryer on the chip or die (also drops temps a couple of C).

The tube I got from Aliexpress, China, was, well..it was a lot dryer. Dryer than even the ZF-EX. No way you're spreading that no matter what you do, but the X method or 5 dot method still works fine. Same temps. I assume it was sitting around for awhile in storage or something.

The ZF-EX seemed to be in the middle. Dryer than the Nan's tube but not as dry as the TFX From China.

All felt the exact same and had the exact same appearance once rubbed deeply on your fingers. You could tell it's the exact same paste.

 

I think this is a bad news for my laptop, I'm not expecting it can match Thermalright TFX. This is a change for the mainstream, better application.

Is there any info about Snowman MTG 5/10, it's rated 14.3W/mK: https://de.aliexpress.com/item/4000...&terminal_id=5d7b78d8d26c4455990dc5ea5e207df7

 

Luumi did test the Sonwman MTG5/10:



However his results are not really representative for a laptop.

Kingpin KPx application from ZoLKoRn:



No idea if he is using the new batch but looks quite easy the application.

 

By the way I have a new CoolerMaster MasterGel Maker Nano, will check out next weekend if there are differences compared to my 3 years old MasterGel. I have also some other paste as well which I haven't tried out yet. Hopefully I can test them all out next weekend, I need at least two attempts for every paste to make sure it's correct. And initial temps only, I can't test sustained performance. Overall I have 17 newly purchased paste in my collection, mostly high rated pastes.


Alphacool Subzero Thermal Grease 16 W/mK
Phobya NanoGrease Extreme 16 W/mK
Thermalright TFX 14.3 W/mk
Thermal Grizzly Kryonaut Extreme 14.2 W/mK
Thermalright TF8 13.8 W/mK
Alpenföhn Permafrost 2 13.4 W/mK
IONZ IZP14 13.4 W/mK
EC 360 Ruby 13.4 W/mK
Thermal Grizzly Kryonaut 12.5 W/mK
AABCOOLING Thermal Grease 5 12.5 W/mK
Prolimatech PK-3 11.2 W/mK
CoolerMaster MasterGel Maker Nano 11 W/mK
Antec Formula X 11 W/mK
Gelid GC-Extreme 8.5 W/mK
Arctic MX4 8.5 W/mK
EVGA Frostbite 2 1.6 W/mK
Noctua NT-HT2 ?? W/mK

 

After 7 hours of retesting I can share my results.

Average of the 4 cores first and then each core separate. The last 4 in my test reached 100 degrees on core 1, this core throttled a bit. They would have reached slightly above 100 without throttling on this core. On a further note, this is an initial temperature test, I cannot test sustained performance.

1. Thermalright TFX--------------------------------------86.5 degrees (83/91/85/87)
2. CoolerMaster MasterGel Maker Nano-------------87.75 degrees (83/93/86/89)
3. Thermal Grizzly Kryonaut Extreme----------------87.75 degrees (83/93/86/89)
4. Noctua NT-HT2-----------------------------------------88.25 degrees (83/93/87/90)
5. Thermal Grizzly Kryonaut----------------------------88.5 degrees (85/93/87/90)
6. Thermalright TF8--------------------------------------89.25 degrees (85/95/87/90)
7. Alpenföhn Permafrost 2------------------------------89.5 degrees (86/95/87/90)
8. Prolimatech PK-3--------------------------------------91.0 degrees (87/96/89/92)
9. IONZ IZP14----------------------------------------------91.25 degrees (87/97/89/92)
10. EC 360 Ruby--------------------------------------------91.25 degrees (88/96/89/92)
11. Gelid GC-Extreme-------------------------------------92.0 degrees (88/97/89/94)
12. Alphacool Subzero Thermal Grease---------------92.0 degrees (89/96/91/92)
13. Phobya NanoGrease Extreme-----------------------92.0 degrees (88/97/90/93)
14. EVGA Frostbite 2---------------------------------------93.5 degrees (88/100/91/95)
15. AABCOOLING Thermal Grease 5------------------94.0 degrees (90/100/91/95)
16. Arctic MX-4-----------------------------------------------94.25 degrees (89/100/91/97)
17. Antec Formula X----------------------------------------94.75 degrees (92/100/92/95)

 

And as I said, TFX is the best paste
Amazing how poorly Phobya Nanogrease Extreme(Alphacool Subzero=same thing) did.

 

Yes the Phobya is a surprise to me as well, I purchased it because of a good result in a laptop test on youtube but it doesn't work for me. I wonder if it's related to the smaller 10nm Intel chips.

 

Great job, temps are really toasty considering turbo is disabled though. Also that core to core difference is huge

 

It's a Dell, the cooling setup is really thin for an 28W+ device especially the single heatpipe and the cooling vents between the hinge: https://www.notebookcheck.net/fileadmin/_processed_/3/a/csm_MG_2226_559b3c5128.jpg

But here you have to keep in mind it's Prime 95 Small FFTs with AVX512 which is extremely demanding on Intel CPUs, the temps are 10+ degrees lower with Cinebench R20. Also everything is heated up including the desk and the heatpipe. And by the way in tablet mode the temps would be like 10 degrees lower (more breathing room for the fan and heatpipe). However I rarely use it in tablet mode and preferred the classical desktop mode on a desk for this test. The core delta looks normal given the high temperatures, there is a ~5 degrees delta in the 70-80 temperature range with Cinebench R20 using a top performer paste. From a previous TFX Cinebench R20 run: 71/74/71/72

 

Prime 95 has a AVX 512 already? That explains the toasty temperatures.

 

Yes it supports AVX512, it can be disabled.

 

I don't get the whole stressing of your systems to produce high temps to show how a paste performs. It's more validating to take a normal workload over 2-3 days to find the best option.

I recently went through several "pastes" and even one performed well for 3 days but day 4 it went to crap and failed to transfer heat effectively causing yet another swap to a different brand.



Currently using the Antec Formula X and it's keeping things consistent.



The MAX temps hitting suboptimal readings is typically due to reopening everything immediately after a reboot w/o letting things settle. If I show patience max temps typically sit at ~90C after an extended period of time.

I've tried several different tubes of paste with this laptop over the last couple of years and they all pretty much do the same thing for temps unless it's utter crap.

I'm running a I7-9750H at max power and with Antec the first time I saw the CPU hit 4.25GHZ. I've noticed that the paste performs well in triggering fans and reducing temps quicker than in the past. I find running fans @ MAX reduces the effectiveness of the heat sink in pushing the heat out the exhaust. Bursts of fans seems to work more effectively in reducing increased temps vs blowing hot air across the heat sink constantly.

 

@Clamibot

I just got done repasting and padding my M17 R1 with the products I described on that other thread.

For all interested, I gave these two items a go:

https://www.tglobaltechnology.com/product/tg-pp-10-silicone-thermal-putty/

https://www.coollaboratory.com/product/coollaboratory-liquid-metalpad/

I did not do a deep dive into the thermal performance of my laptop because I ran into a problem that I can't fix immediately, which I'll get into later. I ran TS Bench, the stress testing utility built into Throttlestop, Cinebench R23, and Uningine Heaven Benchmark.

I think the liquid metal pad is the real deal. The problem I had is Alienware's absolute crap heatsink isn't flush against the CPU die. I had a 30 degree temperature differential between the hottest and coolest cores. The temps I did get on the cooler cores was unbelievable. Like 65C under load with Cinebench. The GPU saw better performance as I believe the heatsink is a better fit on that side of the cooling solution. The metal pad itself is like aluminum foil, but far more fragile. It bends and tears very easily and, in my opinion, would work great on a LGA socketed CPU in either a desktop or laptop. I think you need very even clamping pressure to make the most of this product. Unfortunately, because of the extremely lopsided defect in my heatsink, after testing I broke it down and went back to ICD.

Now, interestingly, the metal pad literally fuses with the CPU die. You can visibly see where the CPU got hot as the metal pad consistency changes at those points. It is brittle at those points and sticks to the die and heatsink. I think if this product is well applied, it is superior to true liquid metal TIMs in that it is far safer (no spillage) and way easier to remove.

The second product is a silicone thermal putty that has a 10 W/mk heat transfer rating. This isn't as good as Fujipoly, but it is idiot proof in trying to figure out what thickness thermal pad to apply. It is a very loose putty and will leave a mess when you're working with it. But that looseness means it compresses when you clamp down the heatsink, fills gaps, and otherwise makes sure those memory modules, VRMs, and chokes are all properly covered.

I intend to return to the metal pad after I figure out how to level my heatsink.

 

ARCTIC MX-5 is in stock on Amazon.com

Viscosity differs according to the page.

MX-5 550 Poise
MX-4 870 Poise

 

MX-5 has less thickness than MX-4?
Isn't that bad?

 

Not if you want it to resist pumping out. The higher the viscosity, the better.

It's bad if you want to spread the paste yourself, but I don't spread thermal paste manually. I don't see the point in doing that. I just let the heatsink spread it for me when I squash it onto the CPU.

 

I mean, MX5 being less thick (lower viscosity) is bad, not the other way around.

 

Oh whoops. I could've sworn you said more thick.

I looked at your post again and now I see that you said less thick. My bad.

 

My guess is that they are positioning it to compete with the likes of kryonaut and new kpx which are softer pastes that work better at subzero temperatures since they can be reflowed more easily should they crack at subzero temperatures.

 

It is available on amazon.de, if all goes well I will test it in a few days. Hopefully it is better than MX-4.

 

Hi guys, any suggestion on what non conductive (first time i repaste anything, and i'm kinda scared of spilling some) paste for my omen dc1041nl? Currently on -140mv, capped to 3.5 / 3.7 ghz, it easily does 85°c on both cpu and gpu, and i'd like to lower it down. I see many suggest ic diamond but availability in italy is what it is...

 

Thermalright TFX

 

@ peterthedoor

Antec Formula X seems pretty consistent and has a good density to it making it easy to apply. Really any paste from a known brand will be better than what's on there now. They're all going to be within a couple of degrees of each other.

 

I finished testing the new Arctic MX-5 and also tested the MARS Gaming MT1 which is rated 9 W/mK. Environment and ambient temperature didn't change from this test. The MX-5 is noticeably thinner than MX-4, very easy to spread and by the way it's a blueish paste. It's not a good paste for laptops but considering it has slightly lower temperatures compared to MX-4 in my test with a more watery paste they have improved something. It's very cheap (8g less than 10€) and very easy to spread, this paste is for the mainstream and should be a solid choice for heatspreader CPUs. For a laptop I wouldn't use it.

1. Thermalright TFX--------------------------------------86.5 degrees (83/91/85/87)
2. CoolerMaster MasterGel Maker Nano-------------87.75 degrees (83/93/86/89)
3. Thermal Grizzly Kryonaut Extreme----------------87.75 degrees (83/93/86/89)
4. Noctua NT-HT2-----------------------------------------88.25 degrees (83/93/87/90)
5. Thermal Grizzly Kryonaut----------------------------88.5 degrees (85/93/87/90)
6. Thermalright TF8--------------------------------------89.25 degrees (85/95/87/90)
7. Alpenföhn Permafrost 2------------------------------89.5 degrees (86/95/87/90)
8. Prolimatech PK-3--------------------------------------91.0 degrees (87/96/89/92)
9. IONZ IZP14----------------------------------------------91.25 degrees (87/97/89/92)
10. EC 360 Ruby--------------------------------------------91.25 degrees (88/96/89/92)
11. MARS Gaming MT1----------------------------------91.5 degrees (87/97/90/92)
12. Gelid GC-Extreme-------------------------------------92.0 degrees (88/97/89/94)
13. Alphacool Subzero Thermal Grease---------------92.0 degrees (89/96/91/92)
14. Phobya NanoGrease Extreme-----------------------92.0 degrees (88/97/90/93)
15. Arctic MX-5----------------------------------------------93.5 degrees (88/100/91/95)
16. EVGA Frostbite 2---------------------------------------93.5 degrees (88/100/91/95)
17. AABCOOLING Thermal Grease 5------------------94.0 degrees (90/100/91/95)
18. Arctic MX-4-----------------------------------------------94.25 degrees (89/100/91/97)
19. Antec Formula X----------------------------------------94.75 degrees (92/100/92/95)

 

Sorry for my ingnorance/ utter noobness on the matter... but how does it perform longevity wise? And is it easy to install?
To be honest, it's the first time i repaste anything, and i also hope to not do it too frequently

 

How's your TFX / ZF-EX doing for you?

 

Don't try to spread TFX. You will fail miserably.
Just apply it in a straight line across a BGA CPU, or in a large intersecting X pattern (you can also use 5 large drops, one in middle, 4 diagonally between the middle and edge of the IHS or large GPU die) on IHS or GPU dies. Then tighten the heatsink.

 

Excellent

It spreads really well actually, way easier than old KPX and is one of the better spreading pastes since its not that sticky.

 

Whenever I try to spread it it just sticks to the spatula unless I press SUPER hard then it's just not worth the effort to waste paste trying to do all that work. Maybe they changed the formula?

I do have two tubes of TFX. The newer one is less thick than the one I ordered from China.
Is the ZF-EX sticky for you? Because it is for me (it's like the same as the TFX I got from China).

 

I've tried 3 tubes of TFX so far and all had stuck to the spatula, as you said it required more paste than usual and hard pressure to spread. TFX is the thickest paste from all I've tried.

 

Hmm, I'll take a video the next time i apply TFX and let you guys see if its the same consistency as yours.

 

Can you check on youtube from other TFX/ZF-EX tests if the consistency looks the same as yours? Here are some:

sample 1
sample 2
sample 3

 

The last sample I bought from Nan's Gaming Gear about a month ago was more fluid and came out of the syringe easier and was easier to apply than the two I bought previously last year. I noticed a strange pumpout type effect with it also if I applied it on my GPU and then ran the GPU at very high temps (>75C) before letting it cure, almost identical to what I saw with Kryonaut Extreme. The one I bought from Aliexpress that came two months ago was about as dry/thick as the Thermagic ZF-EX samples I bought and definitely seemed more resistant to that.