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

I've ordered KRYO 33 now, I believe this could be a good laptop paste because it's obviously designed for laptops and GPUs. The consistency should suit a laptop, I guess that's why they are basically saying a drop in the middle doesn't work. And they have a history of other viscous paste.

 

? I can't find this anywhere. Link?

this? this is a kit...

https://www.amazon.com/Ultimate-Repaste-Compatible-Cryptocurrency-Alienware/dp/B091M7RTPR/

 

I've ordered it from Amazon.de

Because KRYO 33 is very new it seems to be not available on every Amazon shop, I found it on Amazon.co.uk as well.

 

https://www.reddit.com/r/GamingLaptops/comments/jzhskx/if_youre_wondering_what_the_best_paste_for/

Interesting, Hydronaut is better than Kyronaut for direct die CPUs (much better longevity) according to Der8auer. Hydronaut is also more viscous than Kyronaut and Kyronaut Extreme according to specification. The only thing is I can't/won't test longevity because I'm changing the paste too often at the moment. I will give it a try though.

Degradation in laptops is quite common I believe. This Noctua NT-H1 result is shocking. Or in this test Kingpin KPx after 6 months degraded 5 degrees. Some say that Arctic MX-5 has a long lifetime, however if it's 10+ degrees worse from the start on in a laptop with uneven heatsink who cares.

 

Waiting patiently for your Kryo33 results.
And I'm still waiting for my 0.2mm Indium foil to arrive. Holy cow why does everything turn into a slow boat when I need something?
https://www.amazon.com/dp/B092YNNG5Q

 

Stopped reading there. Guy is sponsored by Thermal Grizzly so don't take anything he says at face value with regards to thermal paste.

 

He isn't only sponsored. The point is he recommended Hydronaut over the other Thermal Grizzly pastes for direct die cooling. Furthermore Hydronaut is more viscous than the others according to their official specification which is generally a good thing for a laptop, in particular when the heatsink isn't fully flat. I think it's widely known that the regular Kyronaut struggles over 80 degrees which isn't suited for a laptop in most cases. You also didn't read the longevity report from Reonu, he tried 8 pastes and Hydronaut gave him the best longevity result. As I said I cannot or won't test longevity for now, I want to see how it performs in my initial temperature test compared to the regular Kyronaut or Extreme version.

 

They announced a new thermal paste:
https://streacom.com/products/tx13-high-performance-thermal-paste-without-the-waste/

Anyone knows the viscosity of TFX? Cant find on internet.

 

Tested KRYO33 today with several others. In my initial Prime95 temperature test on my Dell 7506 it's behind the best but still quite good and in the longrun it might be more competitive after some weeks or months.

Thermal Grizzly Hydronaut is too much behind the best even if it has a very good longrun performance. This paste is noticeably less viscous, so it might work better in other devices.

Xynetic CTG8 is better than Xynetic CTG9, it's surprising but it's correct. I already tested them two weeks ago with the same result. Xynetic CTG8 and Maxtor CTG8 seems to be the same and from this result it's obvious to me that the Maxtor CTG9 wouldn't improve the temps over CTG8 on my device.

I'm using a newer batch of SYY-157 here which is slightly more runny compared to my first SYY-157 tube, earlier test was about 0.5 degrees worse averaged over all cores but as always with such small margins it could be a margin of error.


KRYO33
76/82/78/80

Thermal Grizzly Hydronaut
78/86/81/83

SYY-157
73/81/75/78

FuzeIce Plus
71/80/74/76

Maxtor CTG8
78/82/79/80

Xynetic CTG9
79/86/81/83

Xynetic CTG8
76/83/79/81

 

Deepcool G40
74/81/75/77

Prolimatech PK-3 Nano
74/81/76/78

SYY-157
73/81/75/78


I have to try Deepcool EX750 at some point if it's available on Amazon or ebay. Pretty sure my older SYY-157 tube would beat both by a small margin. Thermalright TFX and FuzeIce Plus seem to be the best for me and then closely followed by a few others which are similar in several of my tests (SYY-157 old, IC Diamond24, Cryorig CP5, Coolermaster MasterGel Maker, Noctua NT-HT2, SYY-157 new, Deepcool G40). All of them are really good, although I wouldn't recommend IC Diamond24 and even Thermalright TFX which is a very hard and expensive paste, it's not for everyone. SYY-157 is an outstanding paste (even if it's 1 degree down on FuzeIce Plus/Thermalright TFX) because it's so cheap and very easy to use.

 

How many computers will the 2g supply of SYY paste? I have 4 I want to paste.

 

Difficult to say because it depends on the chip. My old 4g SYY-157 was good enough for 7 applications and I have roughly 20% paste left in this tube, my i7-1165G7(+on package PCH) is a small chip.

 

You should buy enough paste for the job and not try to get by on the bare minimum size. 2g isn't very much. That might be enough for two LGA CPU's or five BGA chips or three GPU dies.

 

Hello everyone, I need help on choosing thermal paste based on the problem of this topic:
http://forum.notebookreview.com/threads/help-please-extremely-fast-thermalpaste-pumpout.833872/

(since this I Lapped and evened out this ****ty heatsink as far as I could, but still, MX-4 Pumps out after 2-3 weeks.)

I read a few threads and pages of this thread but still can't decide. please help and thank you.

the ones that I can buy:
Cooler master mastergel Regular and Pro V1 and V2
Deep cool G15
Noctua NT-H1
Cooler master Ice Fusion V2
Cooler master PTK-L01 ! (first time seeing this)
Corsair XTM50 and TM30 (But I rather not choosing them cuz of ridiculous price, around 60% higher than CM pro)

tested:
TR ChillFactor 3
MX-4
halnziye hy880 (this one was very viscous but its thermal performance wasn't good)

 

MX-4 is hot garbage.
Chillfactor 3 is WAY old. It has long been obsoleted by TF8 and TFX.
I'm sorry that you're upset that certain pastes are not available in your country but if you want to solve this problem, you're going to have to stop complaining about it and find a way to get those pastes. If we can put people in space, you can get thermal paste in your country.

Try
Thermalright TFX
Thermalright TF8
SYY-157
Innovation Cooling IC Diamond
IC graphite thermal pad (this will at least stop pump out)
FuzeIce Plus
Deepcool G40
Thermagic ZF-EX

 

Well, I'm not upset or anything it's just that if I want to buy from another country it will take around 30-45 days to reach here + a really high tax and I need my laptop for my university project (ML+Fluent) which I should be doing right now!

so How about I start with G15? I read that it's actually G40 in lower volume! and it's cheap.

and another thing, someone is coming to my location from the US and I asked them if they could buy me a thermal paste/pad.
I wanted to buy carbonaut (if they agreed) but I don't see this in the recommended ones, is IC better than carbonaut? and do you think I should buy TFX? (it is viscous but it's still a paste!)

my heatsink pressure is pretty good and after lapping and tilting, I dare to say it's almost leveled, almost (I hope Im right).

Oh and yes I had that Chillfactor for around 5 years now


Thank you

 

IC and carbonaut are basically the same thing performance wise but no one has done side by side testing of them together.
Both require a very flat heatsink and lots of mounting pressure.
Both of their w/mk values are total baloney hogwash. They measure them on the X/Y plane (horizontally), while the only thing that matters is the Z axis plane.
Any "lower" z axis published values are usually at pressure ranges far beyond what can be done commercially.
IC pads can usually be re-used at least once on the same mount. Carbonaut can not.

 

Thermalright CF III was really poor in my test, it was worse than Arctic MX-4 which is bad. Deepcool G15 should be the same as Deepcool G40. This is a very good paste for me with very good initial temps, so there is a chance it's good for you as well. Keep in mind I can't test longevity and by looking into your thread I think your heatsink might be worse than what I have. Deepcool G40 is a relatively thick paste but it's not as thick as some other good performer like FuzeIce Plus (One enjoy), IC Diamond, Thermalright TFX, Cyrorig CP5, SYY-157.

 

Thank you
Do you know if this line is true? (Thank you Rooter1234)
"G15 is G40 in lower volume"

And do you think I should get TRX or carbonaut? (or FuzeIce Plus, Cyrorig CP5, SYY-157)

based on heatsink shape and these temps (MX4 after 20 days) Prime95 Blend after 10min (I get same temps with 60-62W on small fft test)


vs first day small fft:


Also, any idea for the heatsink will be appreciated. (That topic was for a long time ago, now it won't pump out that badly, one time it was good for 3 months with mx4)

 

https://www.gamerstorm.com/product/ACCESSORY/2019-04/1291_10859.shtml
https://www.gamerstorm.com/product/ACCESSORY/2019-04/1291_10858.shtml

G15= 1.5g
G40= 4.0 g

 

Need a picture of the actual heatsink itself and the screw mechanism (and it has to be extremely clear)

On some spring-loaded top mount heatsinks, you can increase mounting pressure by applying a small flat washer (like 0.2mm thick) on top of the heatsink before putting the screw on. This may require removing the "c-clip" that attaches the screw to the heatsink. In this type of heatsink, the spring tension is what creates downwards pressure, by "pushing" the heatsink down away from the top of the screw, as the spring is compressed, so applying a small flat washer will compress the spring even more, giving more pressure. This works best if the heatsink is as flat as possible (that may require some VERY MINOR sanding, and the less you sand, the better, because anything sanded beyond the bare minimum will -reduce- pressure. You should be able to find washers that small possibly at an auto parts shop. Some building hardware stores may not sell washers that small. I forgot the spec, I think they're called M2 or M3 diameter washers or something. Ace Hardware in USA actually carries washers that small. Home Depot does NOT in many locations (no idea about Lowes). Such washers may be on Amazon (or even ebay).

You have a 4C delta on fresh application so it looks like the heatsink is decent. Might just need some good thick pastes then. Or you can try foam dams (very light density polyurethane foam--air conditioner filter foam works well, and if it's too thick, just trim it down)+ kapton polyimide tape (3M tape only) to cover the SMD's around the chip core + liquid metal.
Note: Liquid metal works best if you "buff" both the CPU and heatsink with 1500 grit sandpaper to roughen it up. LM absolutely HATES polished surfaces because it destroys its wetting capability.

 

Thank you for detailed explanation

I have those washers (not the flat ones, used them for Ln2 pots) and tested them, pressure was so high that thermal paste actually pumped out faster!

My heatsink looks like this:






Spring on the upper side screw is a lot stiffer than the bottom ones.

And do you think I could use pure gallium instead of LM? Its solid in room temperature (which prevents pump out when moving the laptop) and intel used it too, so what do you think?

Oh and I roughened up my heatsink before but never thought of die will it help with thermal paste too?

And I found halnziye hy-a9 here is it any good?
http://www.halnziye.net/product/196.html

https://www.amazon.com/Halnziye-Thermal-Silicone-Performance-Overclocking/dp/B07HMD9YLM

 

No you can *NOT* use pure gallium. You risk crushing the core when it expands when it freezes or losing contact and getting skyrocketing temps, and you can't apply a bond otherwise. The same issue why LM can't be used at subzero temps. Can't answer your other questions, I'm sorry. And your temps shouldn't rise 'faster' with washers if the balance and fit is proper, unless the paste is just plain terrible. I'm not an expert on this sort of difficult stuff you're asking.

 

Thanks a lot, I am grateful to you, your answers were a great help.
well, it was mx4 so a terrible paste was used.

if anyone has tested this before please help and if not I will buy this and G15 saturday and I will send the results here.
halnziye hy-a9
http://www.halnziye.net/product/196.html

https://www.amazon.com/Halnziye-Thermal-Silicone-Performance-Overclocking/dp/B07HMD9YLM


Ok so I bought G40 and a country made thermal paste to test
I found Liquid Ultra should I buy that too? Is it good? (I mean in comparison to TG)

 

Halnziye HY-A9 wasn't good for me: http://forum.notebookreview.com/thr...and-liquid-metal.806840/page-53#post-11096136

 

hi
pls exist any better LM "than TG Conductonaut" ????
,.it used to work great, but now it still creates dry spots "a few months, but also the day after the exchange"

this is some new composition?? .,,the old TG worked well for 1-2 years
,.now this "dry zones" create some like ""black dust"" and it is not possible to apply LM without full cleaning

or exist any thermal paste for custom OC ?? 7820HK 1.25V-1.3V // 75-95W
old MSI madness returns

 

Hello again and thank you.

Ok, now I am here with the first result (my country made thermal paste) and man I couldn't believe it,
Based on Review websites here it should be around 2C of TG kryonaut so here are the results:

10 min small FFT of Prime98 at 70W (Igpu is on so some power goes there which results in showing lower TDP)


same test with MX-4 (around 7 min) day 1 of application


50W CPU Prime98 small FFT + FureMark GPU at 85W (new paste)


Gpu Stats (sorry forgot to take screenshot when gpu was under load)


and here is the picture of today's MX4 after 2 to 3 weeks: (It was wattery!)


and here are the pictures of the new thermal paste (just for seeing pressure dist/ opening right after screwing in)





now I just should wait and see the results after some time. (I didn't open the G40 yet), wish me luck.

 

after a full investigation ,, the problem is a randomly bent mobo "on VR - random 0.5 -2.5mm"
GT75 7re "cpu + VR --one heatsink"

the question is
I used a mix of paste and pads - this now works perfect
,.,but how long can the paste last on VR ??
mosfet - pads .,,.chokes -TG kryonaut ,,.cpu TG LM

i try find new "ULTRA soft pads" , but it will be difficult to assemble correctly
now -80w 70C "core temp difference 2C --best results

edit
black dust on cpu - removed "burns?" , .plastic scraper and cleaning " technical petrol"

 

your results look good
CPUs below 90C are very nice results "90-94C is max" / GPU RTX is around 85C MAX "87+C temp point before slowdown" -- OK

I do not recommend LM ! works randomly and requires protection / control
,,.Also, don't often use these "benchmarks stress ",.,this will quickly shorten the life of thermal paste and pads

normal your usage ,.,.and I think it will work well

 

Akasa released a new paste called T5 Pro-Grade+ with a viscosity rating of 6,000,000 (mPa.s/ 22̊C). I know that the viscosity rating is hard to compare over other manufacturers but that's a high number. Just for comparison: Deepcool EX750 is rated 5,600,000 and Z10 even lower.


It's designed with hybrid silicone and nano-diamond particles and unsurprisingly outperforms all the previous Akasa pastes according to them. I've searched for it but seems like it's not yet available.

 

Just finished a test with 3 of my favourites and also KRYO33. Thermalright TFX was 1 degree better this time but this is not a big deal. KRYO33 actually is not as thick as SYY-157+FuzeIce Plus and this might be the reason why it's ~2 degrees down, just because my device really prefers highly viscous pastes.


KRYO33
76/83/79/81

SYY-157 new
74/82/76/78

SYY-157 old
74/81/76/78

FuzeIce Plus
74/81/76/78

Thermalright TFX
73/80/74/76

 

Ok - other question.
I am going to replace thermal paste on VRMs and 'other stuff' visible here:


and for CPU goes Thermal Grizzly Conductonaut (LM), GPU Thermalright TFX but for others I have no idea what to put...
I don't want 'play with' CoolerMaster Maker or something thick because it will be nightmare to spread + it could lift/worser mount pressure.
Could you advice any thermal paste that will be 'ok' / easy to apply?

Maybe Noctua NT-H2 ?
Edit: NT-H2 is no go due high pump-out risk... Maybe Gelid-GC Extreme?

PS: VRM's also reach temps like 80*C soo treat them like CPU die (no-go for MX4 and other watery/conductive things)

 

There is a paste called K5 Pro from computer-systems.gr on Amazon which is specifically designed for pad replacement and they have some videos on youtube. I don't know if this is good and if the usual paste could replace a pad. Gelid-GC Extreme wasn't thick (despite the difficult application), it's was relatively runny when I immediately removed the heatsink after a Prime95 run and I guess that's why it doesn't work for me. Because of that I have doubts it's a good choice for pad replacement with possibly bigger gaps but this is more a guess. Not every thick paste is a nightmare to spread. From my paste collection I would try Cyrorig CP5, it's thicker than SYY-157+FuzeIce Plus and it was quite easy to apply for me. But I have no idea if this works for VRMs and other stuff and longevity is another unknown variable.

 

K5 Pro is a good option instead for pads. More forgiven with different high between the components.

 

Thank you both for advices.
I have ordered K5 Pro on amazon (20g tube with express delivery on wednesday for 12.9$ / 10.9€)

 

I'm planning on using IC Diamond 7 for a repaste on my laptop, and ordered some K5 Pro to keep handy if the thermal pads are not in reusable shape. Is ICD7 still a good idea for laptops? This will hopefully be the last time this thing is ever opened again before it's junked in another 2-3 years. How long will a good application of ICD7 last?

It came with Gelid GC Extreme from HIDEvolution back in 2018.

Edit: As an aside, if I were to ever try liquid metal on a laptop, is Coollaboratory Liquid Ultra preferable because it's allegedly thicker and easier to apply than Thermal Grizzly Conductonaut?

 

Having used both CLU and Conductoaut, the difference is negligible between the two - CLU * slightly* easier to spread, Conductonaut (supposedly) * slightly* higher performing. I prefer CLU for liquid metal applications because I want to believe that being imperceptibly more viscous (vs Conductonaut) means it is also (imperceptibly) safer. I prefer Thermal Grizzly's traditional pastes for high performance, non-liquid metal use. Old standby pastes like ICD7 or Artic Silver are still good as a alternative to the more expensive stuff for situations where you're not worried about a degree or two difference, not overclocking, etc...

 

Thanks. I might try LM on a desktop in the future to experiment and get used to it. I usually do more overclocking there anyway.

So I've got SYY-157, Thermaltake TFX, and IC Diamond 7 on the way. Which of these lasts the longest?

FWIW, the Gelid GC Extreme that HIDEvolution put on my GT75 has held up pretty well for the past 2.5 years. I haven't stress tested it in Prime95 or anything crazy like that, but the basic XTU/CPU-Z stress tests are fine while maintaining the overclock. Though the silicon itself hasn't kept the OC as well (it reboots more often when overclocked to +400MHz on all cores, but still rare enough that I can crank it up for a gaming session and then turn it down after to prevent random crashes while I'm away from the machine).

 

How good are the stock thermal pads on the MSI GT75's video card (GTX 1080)? I have Fujipoly Extreme pads on CPU but stock pads on GPU. They look to be in good shape. Should I leave them as-is or replace with K5 Pro? The GPU core doesn't seem to get that hot to begin with, so I'm not sure if it's a good idea or not.

 

Leave them as-is. Why fix a no-problem? Stock thermal pads is good enough. They are often soft and this help on unevenness as different height between the components.

 

No need to change the stock pads unless you're doing Liquid Metal repaste, or shunt mods or Bios HW programmer TDP mods and the stock pads end up ripping up.
I believe all of the pads are 1.0mm, that was on GT75VR and GT73VR. I am unsure if this is the same for GT75 Titan.

If you end up destroying the pads, use Gelid Extreme for replacement.

 

Thanks!

Also, is TFX really better than Kryonaut in terms of pure thermals (on CPU)? Should I be using it for my desktops as well?

 

TFX should be equal to or better than Kryonaut (did not test Kryonaut Extreme) on air. On my 10900k and AIO, Kryonaut was 2C better than TFX at first but after 1 week, TFX was 1C better than Kryonaut (tested with Arctic LF II 360 AIO with AIDA64 Stress FPU). Sub-zero, Kryonaut Extreme wins (according to Elmor from Elmor Labs). TFX seems to be poor on subzero.

 

Awesome, can't wait to try TFX. Should I spread it or use the X method on mobile CPU and GPU?

Also, FWIW to any people finding this thread through search and Google, when I took a closer look at the GC Extreme on the notebook GPU from 2018, it still looked good. The stuff on the die had good consistency, not totally dried out. The stuff that squishes out past the die had been dried out and annoying to clean off. One caveat though, the GPU (this is a GTX 1070) rarely, if EVER, hit high temperatures. It was almost always 70s and lower, often lower than 70 C even while gaming.

As for the stock MSI paste, I took the heatsink off the mobile GTX 1080, then stuck it back on, and it's still just in mid-high 70s during load (3DMark) and low 70s or less with fans on 100% (on a GT75 Titan).

Basically these mobile GPUs from this generation run crazy cool. On the CPU it's a totally different story. You can tell when the temps start crawling up after about a year or two depending on use. Stock paste will throttle in 3DMark, even on an 8750H. I'm excited to try TFX as soon as it gets here.

Slightly unrelated, I had IC Diamond 7 (albeit a slightly old tube) on my OC-ed 9900K (I believe Falkentyne helped me OC that on another forum lol) from 2019 and I had to repaste that last Christmas. It would throttle under stress tests. I used Kryonaut and temps were, and remain, amazing. Stays under 90 C even pushing 200-210 watts sustained loads on an NH-D15S (which is slightly scarred from the botched IC Diamond removal). If TFX can cool that well, I'll be one happy camper.

I'm also thinking of trying IC Diamond's Graphite thermal pad in the future for all non-OC applications. Has anyone tried using that on VRMs and stuff?

 

I ordered two tubes of TFX from Amazon and they've both gotten lost. I swear, this weird thing they do with multiple shipping companies for one shipment (handing off to USPS for the last leg of the trip) is dumb and frequently results in lost or delayed packages.

I have some SYY-157 though. Will it last as long as TFX on a notebook? I need it to last at least 3 years like the Gelid GC Extreme did.

 

I put SYY-157 on the GPU. It's pretty sticky in its own right. Sticky enough to make spreading it annoying as hell. Not as thick/sticky as IC Diamond (and I presume TFX is something like that), but still I'd heat up the tube or paste in the future before applying it if using the spread method.

It (GTX 1080) hits low-mid 70s at around 200 watts and creeps up after a while of load. The heat soak is real. Hitting the 100% Fans button cools it down very quickly. I find myself hitting it just to get the idle temps down after a 3DMark or gaming run.

I've always used the pea or X method with IC Diamond with good results FWIW.

If the TFX ever gets here, I'll put that on the CPU which will be the real test. This thing gets very, very hot.

 

My IC Diamond 24 is very thick but doesn't stick as much to the spatula as TFX. I have no problem with SYY-157 spreading, this is an easy paste for me. The small chip size of TGL-U helps.

 

I got the TFX. It was a little weird to warm up in warm water, the tube's end is open! So the paste is technically open to the air. How long does it last like this? Anyway, warming it up in 60 C water made it about as pliable as SYY-157, which was easier than IC Diamond. I guess it worked well enough since temperatures are a bit better than before.

I think in the future, I'll warm up the TFX with hotter water (70 C) then immediately put it down and use pea or X shape and put heatsink on to squish it. It's just so much easier than trying to spread it. Can't mess up the pea or X method. Every time I spread paste I feel like it's not smooth enough.

 

Bare die, just spread it. It's easy to spread if you "overdo" it slightly. In other words, basically apply squiggly lines across the die (Bare die CPU BGA, or GPU core) in a pattern, then just spread it with the spatula. It's going to be a bit thick but that's fine. It will cover the entire die and the excess will be squeezed out and it should last a long time.
Bare die spreading is always better than X pattern or line pattern. But you can start with the squiggly lines and then give motions like this.

This picture is the best way to do it.



into:



I think these were Seanwee's pictures.

IHS desktop LGA Intel/AMD CPU? X pattern with dots in quadrants, *OR* 9 large dots (3 top, 3 middle, 3 bottom). Both work well. Maybe the 9 large dots will use a little less paste than X + dots in quadrants but I didn't test that.

 

Thermalright TFX
74/80/75/78


Akasa 5026
75/86/79/81


AAB Cooling TG4
85/96/87/90 - stopped after 2 minutes (water paste)


Genesis Silicon 851
76/83/78/81


Genesis Silicon 800
87/95/88/90 - stopped after 2 minutes


Phanteks PH-NDC
75/82/78/80


Dimastech HTX-EE
90/98/93/94 - stopped after 1 minute



Biggest surprise to me is Phanteks PH-NDC, I didn't expect it to be so good. It's not a a thick paste, it's actually relatively thin. Their homepage says it's a Nano diamond-like particle paste, maybe that's why. The new Akasa T5 Pro-Grade+ is also nano-diamond particles paste and (according to the specs) a very thick paste (6,000,000 mPa.s/ 22̊C). Even thicker than JunPus JP-D9000 (5,500,000 ) or Deepcool EX750 (5,600,000 ). Hopefully it's available soon.