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

i was only able to find it on a vietnamese shopping site: https://shopee.vn/Keo-tản-nhiệt-batman-ZF-12-12wmk-4grams-i.63612027.1860975318

 

Me to but I dont think they will sell to my country whereas Aliexpress and eBay will. Hope there is more somewhere else that's available worldwide

 

I know. These results are insane but without any photos, I don't believe much or may be there's something wrong with testing method

I couldn't find it on any site such as Taobao or Tmall. You can take a look at the link below (Shopee VietNam and no oversea shipping), He says Silicone Grease & Radiator make this paste together: https://shopee.vn/Keo-tản-nhiệt-batman-ZF-12-12wmk-4grams-i.63612027.1860975318

Will ask the shop manager where he gets this paste from

 

Cool Thanks tungd678. Please keep us updated

 

I asked shop's owner and he answered that the manufacturer sent him this paste for "PR purpose" so no international selling now

 

Thinner thermal paste will never work properly on bad designed notebook heatsink. This is common knowledge but not all understand this. A short explanation from Der8auer (shareholder at Grizzly).

People buying etc notebooks with TRIPOD Heatsink design/uneven pressure (or bad designed) on Cpu die should stop buying/using thin creamy thermal paste. The very best would be that the Notebook manufacturers stop saving money on the cooling. But of course... Use a thick thermal stamp is a lot cheaper for them aka the preferred way. Same also using long pipes and Unified heatsink design which is cheapest way make a notebook cooler.

 

Never knew Der8auer had shares in TG.

 

yeah only found out myself while watching that vid. but makes total sense

 

Just look at advices and you can see what’s behind reviews (as well behind the scene). 1+1 is always 3 None thermal paste is good for all cases. This is common knowledge @Mr. Fox @Ultra Male @Falkentyne + many others which know about this topic.

 

Hello fellas, very instructive thread!

I'm researching the best fire and forget one to repaste relatives and friend's laptops, so that means the paste with the least amount of pump out effect on laptop heatsinks so the paste can last upwards of 3 years minimum.

So far IC Diamond 7/24 seems to be the one with best longevity according to this thread, but scratches the CPU/GPU dies and has a thermal conductivity of only 4.5 W/mK. What about the Phobya NanoGrease Extreme, does it scratch/corrode the CPU dies ? I see it has much better thermal conductivity and similar viscosity to IC Diamond but what about the pump out effect, could it last upwards of 3 years and become a contender for most enduring paste?

I think it would be very useful to edit the first post, adding a longevity category for most enduring thermal paste for laptops based on pump out/ push out effect. That way enthusiasts can identify the best paste for their system based on thermal conductivity, as they can reapply it every 3 to 6 months themselves, and also the most consistent and long lasting paste for secondary systems or relative's PCs.

 

ignore the thermal conductivity values they put on the packages, they dont mean jack squat

also, the scratching effect of ICD is only "bad" when it comes to visuals. but it wont actually damage any of your hardware, so you can just stick to ICD, no worries

 

But only *we* know that! But outsiders will treat it as paste caused the chip to be scratched!
I am using CM MakerGel Nano and it has same composition as ICD but much smoother.

 

So minimal pump out effect just like with IC Diamond? Well might as well pick it up if it will last 3 years or so.

BTW are t he new CM Mastergel Maker with the novel flat-nose syringe applicator thingy using a different composition or it's the same old one as OG Mastergel Maker?

Anyone used the new flatnosed syringe if so, what's your feedback while applying it on the smaller, narrower rectangular delidded mobile CPU-GPU dies?!

 

Conductivity values are meaningless, only thing you should look at is how viscous the thermal paste is. The more viscous it is - the more it will last in notebooks because notebooks have heatsinks with low mounting pressure and uneven mounting pressure (most manufacturers do not use a 4-point mounting for CPU heatsink plate) and need a thick paste to compensate for those.
https://www.tomshardware.com/reviews/thermal-paste-comparison,5108-12.html

This is not true. Use a common sense for a bit. Ideally you want a perfectly flat surface for both heatsink and the CPU surface because metal-to-metal heat transfer is most efficient. That's a fact, not an opinion ;-) Let's say you made a scratch - that area will definitely have less contact area now compared to perfectly flat area and even if you will fill the scratch with thermal paste - the paste will have less thermal conductivity than a direct metal-to-metal contact ;-) Of course, a few tiny scratches will not have measurable effect, but let's say you put a lot of them and some of them will be big - there will definitely be a difference. This is why many users of desktop CPUs used to do "lapping" or polishing of both heatsink and CPU IHS (and get up to 10C difference according to their own tests) and why MSI is now doing the same to heatsink on their GT76:
https://us.msi.com/Laptop/GT76-Titan-9SX

 

we are talking about providing long lasting thermal paste to regular joe users / fire and forget but sure, for enthusiasts every degree saved is worth it.

Sent from my Xiaomi Mi Max 2 (Oxygen) using Tapatalk

 

Still using the old one w/o any issues.

 

No, you are right and wrong here. Lapping is to get rid of concave or convex surfaces. This reduces the distance across the surfaces from each other. But, unless the surfaces are so flat that they have an attraction to reach other, where say the IHS will stay with the heat sink without some other substance between them, you will need some sort of material to facilitate the heat transfer. Scuffing can help with this, allowing pastes, especially liquid metal, to "wet" or adhere to the surface better. It allows for filling in voids better and giving surface contact area that the paste helps move the heat to.

So there is a bit of both there. Personally, I would lap to mirror, then use a diamond paste to put an extremely fine scuff on the surface. You get the surfaces flat to maximize surface contact while also giving the imperfections to allow the paste to adhere to both surfaces to get great heat transfer.

 

Hello When using CM maker gel nano or GC extreme do we spread or is it the same as ICD where u just apply a line in the middle and let it do its thing?

 

Spread method is better.

 

Thanks for sharing this video, really interesting

 

Is a mirror finish heatspreader beneficial of you're using liquid metal? Or is it just better if you're using paste?

 

I'd be interested to see if the high polish would seal against migration reducing draining the LM over time into the metal. That's where someone could innovate - make a surface that doesn't melt or migrate with LM.

 

I doubt it, since the migration is a chemical process.

And there is a way of protecting against LM corrosion - nickel plating

 

Then you answered your own question - polishing the surface to a mirror finish - removing the plating corrosion barrier - would be a bad idea for using LM over leaving it sealed and protected.

That is unless you polished it - and then plated it, again.

 

I was talking about unplated copper heatsinks. Like on the GT76.

Obviously you don't sand plated surfaces XD

 

It's also obvious you'd want to lap a surface over leaving it raw - unless it came pre-lapped.

If your question is - would making the surface mate more closely by lapping it squeeze out the LM causing it to migrate out from between the mating of the CPU / heatplate, then that would be just as true before lapping - which is why you use barriers to stop leakage of LM.

A better mating across the surface improves heat transfer, so yeah, using LM doesn't make it any different than regular paste, except LM leaks with nothing to "hang on to".

That's why I don't recommend using LM at all, it's an unnecessary risk for leakage and damage on the motherboard - and anything else it drizzles onto.

Try lapping + a good enthusiast paste instead of LM, and if see for your specific usage going without LM is good enough.

 

I sorta see where the confusion stems from.

Let me rephrase

How much would a lapped surface reduce the temp difference between using LM and paste?

 

It depends on the surface(s) before and after lapping. That's why I suggested trying it out and seeing your results. I've seen LM give 0c difference and 10c+ - a lot depends on pressure on the surfaces as well. It also varies according to thickness of the paste.

I would suggest trying LM first (if you must) before lapping - a couple of pastes you like as well - then lap and repeat - come back and let us know.

@der8auer did this a few times already and shows similar results for lapping improvements reducing paste improvements - it's a process of catching up from both ends, better thermal transfer from different means.

 

Mirror finish is terrible. Nothing to do with absorption. Mirror finishes stop LM from even adhering to the surface thus making spreading it almost impossible. Then when it is spread finally, when the surface gets squeezed thin, LM starts adhering to itself rather than the surface, leaving completely exposed streaks or spots, because there is absolutely no grip from the surface whatsoever. This is horrible for temps.

Buffing a surface with 1500 grit sandpaper makes LM love the surface, makes spreading easy and helps it adhere to the surface far better.

 

Yeah I've heard of that, but doesn't it work just fine on the cpu and gpu dies which are already mirror finish?

 

No, @Falkentyne is correct, you need to "rough up" the surface to keep LM in place if the surface won't hold LM in place.

Adding the barrier can help, but nothing can keep LM vs gravity from working together to eventually leak - if it's going to it's going to. LM finds away.

That's why conductive pastes aren't a good idea. No matter how good you are, eventually a random flick of the stick and the silvery globules of doom will take out a motherboard component - maybe it's needed, maybe not.

Roll the dice on a $4000 laptop for a few degree's of C? Nah, the benefits in FPS don't warrant it.

 

Just finished a little experiment on heatsink lapping.

I used my old Lenovo y410p with an i7 4700mq as the test subject.

Here are the results after 20 mins of prime 95 at max fans

Kryonaut
Pre lap
Core 1 max - 77
Core 2 max - 79
Core 3 max - 79
Core 4 max - 77

Post lap
Core 1 max - 74
Core 2 max - 75
Core 3 max - 75
Core 4 max - 74


Conductonaut
Pre lap
Core 1 max - 67
Core 2 max - 69
Core 3 max - 67
Core 4 max - 68

Post lap
Core 1 max - 65
Core 2 max - 66
Core 3 max - 66
Core 4 max - 66

So about 3-4 degrees cooler with paste and 1-3 degrees cooler with LM after lapping. So sadly lapping is no replacement for LM. I had hoped to see a better result with paste.

I used 1500 grit to 7000 grit wet and dry to achieve a mirror finish on the heatsink. And even with a mirror finish the Liquid metal still held on just fine, no issues with that.

 

well, lets put it like this: the fact that u were able to work with LM on your unlapped heatsink means u already had a pretty good fit and won the heatsink lottery to start with thus, post lapping results were expected to be marginal. nice data though!

 

I've never had any laptops with poor heatsink fit yet. All seem to have worked well.

All were pga/bga laptops tho. No clevos.

 

The alienware 15 r3 and 17 r4 are the worse ever. Best solution for a average person like me was ICD thermal paste. Just want to share info. Dont buy those laptops. hehehe..

 

How bad was it on those?

Newer alienware laptops (m15/m15 r2) seem to have pretty ok heatsink fits.

 

here is more info.
http://forum.notebookreview.com/thr...ven-core-temps-due-to-uneven-heatsink.797477/

 

I just repasted my Clevo P960 with Kryonaut. I have been using it with CM Mastergel Maker since august but I noticed thermal throttling. So I am in the market for a new cooling solution or how to tweak my existing cooling system. The TIM on the CPU was dry-ish in the center mostly and on the GPU was even worse. Funny thing tho that i had no issues with GPU temps.
I am really considering in using a graphite pad instead of TIM. I am a little afraid of using Liquid Metal since I am afraid it could spill (I travel almost daily with my laptop and most of the time I do not use it on a flat surface).

 

I used the coollaboratory metal pad kit on my gram, seems to work pretty well. The kit comes with metal pad (it seems to be a liquid metal type substance in a sheet form) and you sandwich a piece of shim between two pieces of metalpad that you cut to the exact size of the die surface area and then place the heatsink back on top. It seems pretty safe as the substance I am guessing liquifies and "cures" (I don't think it's really curing so I air-quote that) under heat of the CPU. The shim keeps it sticking to the area instead of outflowing.

This laptop has pretty poor heat management and it seems to work just SLIGHTLY better than the kryonaut I tried first. The heatpipe/sink still gets saturated very easily so it doesn't help all that much in this unit but it does seem to keep normal non-high-load operating temperatures slightly under what kryonaut was.

 

Why would there be an issue? Liquid Metal will stick even to glass surface. It does not need the surface to be rough, anyone who says this is wrong. This is why it is always better to have as smooth surface as possible since even the tiniest scratch will form a gap which will lower thermal conductivity even if it is filled with thermal paste.

 

Sure. Don't compare liquid metal with ordinar thermal paste.

 

Why are you copypasting these garbage "advices"? The manufacturer only cares if you will save a few seconds when spreading the liquid metal alloy but it does not give a **** about compromised thermal conductivity which is caused by every tiny imperfection you will add by scratching the surface. If you care that you will spend few seconds less when applying liquid metal alloy (which, by the way, WILL "wet" or in other words stick to the metal surface of heatsink and CPU no matter how much you polish them, go read up on properties of Gallium and its alloys) - sure, put as many scratches as you want. If you care about actual temperature decrease - manually scratching the surface is not only unnecessary but irrational to do since you are decreasing direct metal-to-metal contact between solid metal parts and any thermal paste including liquid metal alloy will have worse thermal conductivity and since liquid metal alloy WILL spread and stick to the surface of heatsink and CPU without doing this (once again, go read up on properties of liquid metal alloys based on Gallium). If you are afraid that liquid metal alloy will "leak out" when transporting laptop - just do not apply too much of it (practice your spreading until you'll learn how to do this) and clean the surface of the CPU and heatsink properly and it will NEVER leak if you do both of these properly. Once again, there is absolutely no NEED to damage in any way (including the way described in "suggestion" you copypasted) the surface of heatsink or CPU by manual scratching when applying liquid metal or any other kind of thermal paste, for any purpose other than MAYBE potentially saving a few seconds while spreading the liquid metal alloy (which will spread anyway).

P.S: I know that once again, people WILL have reading comprehension issues and WILL ignore each and every fact that I stated but hey, at least I have tried ;-)

 

Not so sure how familiar you are with Liquid metal, longevity and end results but I and a few have used it a very long time. And with temps and performance results none can beat. Not even close.

Another thread with about same topic. Read also bro @Falkentyne’s posts http://forum.notebookreview.com/thr...r2-owners-lounge.829502/page-32#post-10978040

 

I've used LM for a few years now and I've never had to rough up any surface, they've always held on well.

I'm not sure how long a mirror finish to mirror finish application will last but the temperatures do seem to be ever so slightly better. More direct metal to metal contact is always good.

Though as my results showed, even a mirror finish to mirror finish contact still has enough micro gaps for there to be a sizeable temperature difference going from regular paste to liquid metal.

 

I've experimented with multiple methods on my 9900k. Over 10 delids and relids and also experimenting on heatsinks. Roughening the surface gave me best results. Mirror finish gave me bad longevity.
Take a mirror finish surface and then try to spread the LM *very thin* and watch what happens. Most people don't do this.
It will start streaking and collect to itself leaving exposed copper/nickel. Even if you put enough that it doesn't do that, if you then put it under high pressure and heat, if there is any tiny imbalance (like a convex CPU die), you can check in 1-2 days and unlid your CPU and see an exposed part without any LM. That's because it pooled to the side and towards itself because there was no traction or micro-ridges and your temps will be higher from this. This effect is greatly reduced by using a surface roughed with 1500 grit sandpaper, and the 'imperfections' you created simply get filled up by the LM absorption. This helps maintain temps. If you do the above spread thin test on 1500 grit buffed surfaces, you will notice the 'streaking clear' issue stops much sooner. And the LM adheres better also.

A fully mirror surface works if both surfaces are very flat with perfect contact pressure. But you're only going to get that on a sanded flat CPU and custom bought IHS. The problem is if there is any convexity or concavity at all, even a tiny bit, I found 1500 grit buff sanding works the best.

I've spent many, many hours testing such things on both CPUs and Video cards (Vega 64, etc).

 

Lapped and LMed my GS75



Stress testing with prime 95 small fft + heaven 4.0

Edit: used 1500 grit all the way up to 7000 grit wet/dry. Spent 30 mins each with 7000 grit on the cpu/gpu heatspreader.

 

I didn't make a detailed record of the per core temperatures on my first fresh LM application but the max was 86°C cpu and 72°C gpu

Post lap 30 min stress

https://i.imgur.com/o5NTstA.png

Not sure why this forum is so finicky when it comes to uploading pictures.

A much larger gain than I expected. Might be due to MSI's rougher than usual heatsinks.

 

Hi guys, I repasted my GE65 Raider 9SF (i7 9750H, RTX 2070) with Artic MX-4 thermal paste because, as a desktop PC guy, it's what I usually use (my current lifestyle forced me into buying a laptop instead of upgrading my PC). I did it today and Apex Legends doesn't make my CPU heat up to 95ºC anymore, it stays at low 80s / high 70s, and at a stable 4 GHz with ocassional drops to 3.9 GHz.

The thing is, after the fact I've seen some posts online about how MX-4 is terrible for laptops because of low mouting pressure, and will degrade super fast after a few weeks. Is this true?

If my MX-4 application will really degrade fast, what other thermal paste would you recommend for my laptop so that I can replace it before it happens? I'm not ballsy enough to use liquid metal. Purchasing a 2000€ laptop was absolutely insane on my part and the biggest risk I'm willing to take is the act of opening it up and repasting it in itself, but there's no way in hell I'm going to risk applying liquid metal on it. If this laptop dies I die along with it, it just can't happen under any circumstances, it's not an option. So yeah, traditional thermal paste only please. Thanks!

EDIT: Also, another question. I decided to ignore the...stuff that was on the VRMs, since it didn't look like regular thermal paste. So, I just left it in there and pretended it didn't exist. But eventually I'll have to replace that stuff... What is it? Do I just replace it with regular thermal paste? Thanks

 

It does happen, I've worked with guys / gals trying to get control over their quickly deteriorating temps - repaste and a few months later it's already to go again.

I've had good luck with keeping the paste away from the edge of the CPU / GPU - the part where it meets the air seems to cause wicking away of the moisture, drying out quicker.

I am weird, I apply my paste by hand - no simple X or dots, I use a cut card to squeegee the paste as thin as I can across the whole surface meeting the heat plate, then use an xacto knife to cut away a perimeter about an 1-16th" - 1/8th" of cleaned surface - so no paste will squish out or "oreo" around the edges.

I've also suggested and people have tried sealing around the edge with K5 Pro or another high temp sealant - to keep air from getting in under the edge between the CPU and heatplate. This has worked reportedly for some, but I haven't needed to try it.

I've used AS5 / MX4 and NT-H1, but today I'd use Kryonaut:
https://www.amazon.com/Thermal-Grizzly-Kryonaut-Grease-Paste/dp/B011F7W3LU

There are lots of pastes that will work, but all tend to pump out or dry out over time, more or less fast depending on the thermals and exposure to air. If you need to pull apart the heatplate from the CPU / GPU once seated fully, you need to clean the old paste and re-apply.

Use common sense though, I've recovered by re-smoothing out the paste with a card to a thin covering and made that clean lip around the join edges and it's been fine.

If you can do just the CPU or GPU independently, I'd do one at a time, they need different techniques - you need to check the fitment of each separately - usually the GPU is larger and has more "slop" in the fitment, so it may take a thicker application of paste to get a good mating.

One last thing, perhaps it should be the first thing, I don't recommend re-pasting for most people. It's just not needed. As long as you aren't thermal throttling constantly - to the point where your CPU is losing performance - it's not really needed.

A lot of people get sucked into reducing temps when it's really not *required* to do so. They justify it as lengthening the lifetime of the CPU / laptop, but I've seen people's old laptops that have been running at high temps for years perform just fine and other parts - like the battery - die long before the CPU / GPU.

Also, the thermal paste that comes from the factory is designed to last a long time with stable temps - not provide great temps, but stable temps. That's because they can't have their laptops needing repasting every 3-6-12 months like enthusiast pastes require so they optimize the paste for longevity, not peak efficiency.

So, is it better to enjoy the laptop as is, undervolt perhaps, tune the fan curve for better cooling, FPS limit your games to the refresh rate (this works wonders), lift up the rear of the laptop higher than the front to start convection cooling - can provide great results simply by lifting the rear quarters off the table to allow better intake of cool air and venting the hot exhaust up and away from the table surface.

All those tricks and more will result in better temps without needing to re-paste.

Please let us know what you do and how it works for you. And, if you have any more questions come back - there are lots of people here that can help you.

Edit1: Oh yeah, that other paste, it's likely a thick paste meant to replace thermal pads in applications that don't need that much thickness to mate the chip and the thermal plate, K5 pro works as a replacement for that:
https://www.amazon.com/viscous-thermal-paste-replacement-Aspire/dp/B00UTX7K2E

K5 Pro will "harden" over time to a plasticky consistency - it seals against moisture / air - which is why I suggested it for around the edges of the CPU / GPU to act as a barrier to air which dries out the thin enthusiast paste. I wouldn't use K5 Pro as the CPU / GPU paste, although people have successfully done so. It's easy to remove when it has set, you may have to apply lifting pressure to unseat it from the heat plate later, but don't be alarmed it peals right off or rolls up into little balls - it's easy to clean up.

Edit 2: Wow, missed this one, "MX-4 2019" - a new formulation "better than liquid metal"... IDK anything about it, but may be worth a try? Here's some info lifted from the Amazon product page:
https://www.amazon.com/ARCTIC-MX-4-2019-Performance-Durability/dp/B07L9BDY3T
on newegg:
https://www.newegg.com/arctic-cooling-ac-mx4/p/2MB-000S-00016

• The new version of the #1 bestseller: the MX-4 Edition 2019 convinces everyone with its usual quality and performance, that have always distinguished it
• Better than liquid metal: composed of carbon micro-particles which lead to an extremely high thermal conductivity. Its that heat generated from the CPU or GPU is dissipated efficiently
• Thermal compound: The MX-4 Edition 2019 formula S exceptional heat dissipation from the components and supports the stability needed to push your system to its limit
• Safe application: The MX-4 Edition 2019 is metal-free and non-electrical conductive which eliminates any risks of causing short circuit, adding more protection to the CPU and VGA cards
• High durability: in contrast to metal and silicon thermal compound, The MX-4 Edition 2019 doesn't compromise over time. Once applied, you do not need to apply it again as it will last At least for 8 years

Will wonders never cease, I'm gonna have to try it myself

 

Unfortunately, I already repasted lol. That's why I'm asking if I did well in applying MX-4 or if i should start thinking about yeeting it and replacing it with something better. But if it's good then I'll just leave it there.

Regarding your advice on how to keep the temps down, I'm already using a cooling pad with adjustable height, and the limiting the FPS to the refresh rate thing... my refresh rate is 240 Hz... so yeah... lol

Thanks for all the advice though!