[Video Tutorial] How to repaste laptop with liquid metal to reduce CPU temperatures

Hi NBR,

I just uploaded a video tutorial on how to safely repaste your laptop(s) with liquid metal compounds for beginners. This is extremely easy to follow and pause along. I have broken it down so that any person who would like to achieve thermal bliss can attain. If you do follow the tutorial, or do repaste with liquid metal, leave your results in the comments section or here in this thread.

Thanks all and I hope your laptops do not overheat

 

- If you plan on moving your laptop around, would you recommend any liquid metal (LM) between CPU and heatsink?

If you have a poorly connecting laptop heatsink, and the laptop is on the go being moved, shoved into bags, thrown into an overhead airplane compartment, etc., I think users may want to really think through this decision. Even with the tape around it, leakage can occur. Also, if you use a thin, viscous TIM like Conductonaut, the risk is even greater for the possibility of leakage. Having this stuff get loose and bounce around where it could possibly come in contact with other components inside your case. Yikes!! Let's just say that damage from this is not covered any warranty.

To me, for the mobile user, I would think it would be best to only use LM inside a delidded CPU, where the CPU has been sealed against possible leakage?

 

I have been asked this very question in the past, and while I do agree with the viscosity issue and potential 'seepage' out on to the PCB, I have NEVER experienced this before in my 3 prior laptops and desktop that I repasted with LM. Maybe I am a lucky person? I think if the appropriate precautionary measures are taken to avoid any initial excess spewing on to the PCB, you will be ok. If someone else has had this experience, i.e. liquid metal running down on to the motherboard, I would love to hear about it. Electrical tape is one method, which i find easiest, but nail polish is another. with tape, you know exactly what you are covering!

Secondly, and somewhat aside, traditionally LM was used for delidding and then became used for desktop thermal paste more recently in the overclocking community. Most desktop cases have the motherboard standing vertically, and if gravity were to work it's magic, we would see it leaking on to the motherboard... and we know the rest. However I have yet to hear of such an instance... again, I would not discount someone telling me it happened, but I haven't heard of anyone experiencing this.

If people do not feel comfortable applying LM, by all means avoid it! I can only speak from experience and I have nothing but great things to say.

 

I agree with the gravity statement, but a laptop may be subject to a little bit more abuse (thrown into a bag, carried in all types or orientation, screened/jostled at airports, dropped, etc.) Just thinking out loud.

Also, at least in some laptops, the actual heat sink may have manufacturing flaws (where the sink is not a flat plane), the mounting screws not placing enough pressure on the CPU/GPU to keep the LM in place, or other inconsistencies that should make LM users a bit more cautious.

I wonder if @Mr. Fox, @Papusan, @Donald@HIDevolution, or @woodzstack have any insight or horror stories regarding LMs usage in other parts of a laptop.

 

No horror stories, but I've seen enough that I will only use CLU going forward. I was using Liquid Pro, Phobya Liquid Metal and Thermal Grizzly Conductonaut for a while. With all three of those I found the paste was very watery with no viscosity. I did not think anything about that until later, when I found all three of them had a small trail of liquid metal going across the heat sink (my Clevos are upside down, with the CPU pointing down so gravity pulls toward the heat sink, not the motherboard). That has never once happened to me with CLU, but it is much thicker than the other three. I will not use the other three any more based on my observation. I will always use CLU whenever possible. Had that happened on a laptop with the CPU on top of the motherboard like my Alienware systems were, that trail of liquid metal paste might have gone down into the CPU socket and caused some major devastation and destruction. Seeing it happen more than once, CLU is the only safe option in my own mind.

 

Thanks for the input! I haven't had issues with conductonaut in the past, but I am a sample of 1.

 

People are gambling if they use other Liquid metal than Coolaboratory Liquid Ultra. For maybe 1 - highest 2C degrees better cooling isn't the risk worth it. You can't compare a desktop with Liquid metal who sits on the desk vs. laptops who will be used a totally different way. And most of the heatsink in desktops have normally better fit than most part of any heatsink in laptops.

 

Oh, you're totally welcome, bro. I know a couple of other guys saw the same thing. I even made a video showing the three and at that time paying extra for CLU seemed like a waste of money. Now I think the extra cost is worth every dime. Thankfully, nothing bad happened, but it certainly got my attention.

As a general rule, any liquid metal paste is unbeatable in terms of its effectiveness. They all work amazingly well and are excellent products, but the thick texture of CLU makes it the safest option for a laptop as far as I am concerned. The only caveat with them (including CLU) is the heat sink must fit snugly and make good contact in order to be effective. If the heat sink fit is sloppy (a common encounter for many Clevo owners) it does not work well at all. In fact, in that case IC Diamond is hands-down best. You can use it to plug the gap and get decent results that last a long time where other pastes will quickly fail.

 

Certainly, my only deciding factor for thermal grizzly conductonaut was its availability at the time of ordering (and past experience). My next paste job will definitely be with CLU based on what you ( @Papusan) and @Mr. Fox have said!

 

Nice seeing people listening on others viewpoints. +rep

 

why would you LM anywhere but the GPU die or CPU die ?

 

yeah this contact is not always a guarantee, it literally varies from one heatsink to the next enough to make a difference.

 

He pointed out that Liquid Metal *can* end up in other places than where it should be

 

OH haha I misread LOL

 

Yeah I'll add to this. I've only used Liquid Pro and Liquid Ultra.

I've always used Ultra but I really liked how the Pro spread was a little easier to apply and finished in a nice coat. However, upon inspection of my first application I found a little ball or metal beading around quite far from the die which was the first time I've experienced such a thing. So I'm back on Ultra now.

Have y'all noticed that... anyone who used metal in a laptop NEVER goes back. "He who dare, wins" I guess.


@nravanelli

Nice vid too + Rep

 

He who dares to enter reaps the rewards of low thermals.

Thanks @TBoneSan ! Just trying to start my YouTube channel on a high note If anyone has any comments/suggestions for future tutorials - hit me up.

Feel free to share my video on twitter/facebook, wherever. Exposure is always nice!

 

Another alternative is to use a thinner version of Fujipoly (30X-m or 50X-m), (0.3 or 0.5mm thickness) instead of Scotch 33 electrical tape. It will compress between the edges of the chip lid and the heat sink, creating a dam that will also minimize any escape of liquid metal. It will require quite a bit of trimming with a razor knife and you may want to create a template for what fits right around the chip lid out of craft paper so you can minimize any seams or gaps when you cut in to the Fujipoly.

Additionally it will act as another way to "wick" heat away from the CPU PCB. This will be a slight benefit.

This dam of Fujipoly would keep any liquid metal from escaping, also it is going to keep additional oxygen away from the liquid metal.

I would suggest measuring the height of the CPU lid and using a "TIM" that is just slightly thicker. Fujipoly is compressible by at least 50% of its thickness (taken from the manufacturers web site for Fujipoly).

The idea behind liquid metal is to not use it as a gap-filler. The amount you put down on the chip lid and on the heat sink is known as "wetting" and should not be a ball of metal gobbed up on top of the sink. It relies upon the "intimate contact" between the chip-lid and the heat sink. If you have gaps that require a 'filler" then that is a problem you need to resolve way before you start putting any sort of thermal goo or liquid metal down to bridge the gap.

-------------------
Something to consider with liquid metals that are based on an eutectic mixture known as galinstan. This is a mixture of gallium, indium and tin where the resulting alloy (combination) has a lower melting point than any of its components (around -20 C for galinstan). Manufacturers can "tweak" the melting point and viscosity by very slight changes in the alloy content of the three primary metals

One of the negative things about gallium is that it loves to alloy with other base metals as well. This will be most pronounced with something like aluminum where it will literally turn an aluminum heat sink in to brittle little pieces that snap apart like crackers in your fingers in a matter of days. Even with other metals like copper (and nickel to an even lesser extent) is that it does "alloy" with the metal (hence the wetting effect). Normally this would create a tarnish or maybe some minor spalling if there was a poor plating job (the chip lid is actually nickel plated copper but it is a good nickel coating).

 

I'm a little miffed to have read this thread AFTER I ordered me two syringes of Conductonaut. I was planning on smearing it everywhere in my P775, but it now seems I'd be smart to limit myself to the die-IHS.
But what could I do best to seal that CPU back up again without it lifting up the IHS or making it impossible for me to repeat the repaste?
I'm thinking a Fujipoly dam around the die under the IHS and then spotglueing the IHS back on. Might add to the cooling effect and the compressibility of the Fujipoly will keep the IHS snug with the die.

Thoughts?

Edit: can I say great minds think alike without sounding cocky? Seems everything I think was in this most excellent previous post.
Next step will be to decide what to do on top of the IHS... CLU again I suppose is the safest? Or are there any tricks to help there as well?

 

Safest bet is CLU vs. Conductonaut if you haven't a very good fit between IHS and heatsink(less chance for leakage). But all types of liquid metal would make bad temp results if you haven't good fit from heatsink on top of IHS.

 

----------------------
Actually what would work better is a non-corrosive silicone adhesive. Those are the types that do NOT smell like vinegar (Acetic Acid).

When you apply liquid metal to the top of the chip die and to the inside area of the IHS where the die will come in contact you would make a super tiny little track of silicone adhesive along the bottom flat edge of the IHS that will be in contact with the die-carrier. Leave one tiny little gap where there is no adhesive so any internal pressure from the silicone curing process can escape. When the IHS is then applied to the carrier it will smooth flatter and there will not be any cleanup or mess.

If you use a black silicone adhesive and really nice technique it would be almost impossible to tell the difference between the Intel work and your own in assembling the package.

I am sure that some people will put the adhesive across the entire mating area of the IHS and 99.99% of the time it will not be a problem. But if you look at the IHS when you first crack it open you will see that they had a tiny gap as well.

You are not trying to make a "waterproof" seal. If you get water that far inside of your computer then you have lots of other issues going on. The idea is for the IHS to be sitting as close to the die-carrier as possible, not to be moving around and to have most of the space where air can move in and out of the package closed off.

 

Sounds like a great idea. Now to look for that kind of adhesive, preferably with a tiny little nozzle.

 

So liquid metal is far superior than IC Diamond and Collaboratory?

 

Coolaboratory ~ Coolaboratory Liquid ultra = Liquid metal But Coolaboratory make also standard conventional thermal paste

 

My question is: is it safe to use liquid metal on a laptop that will sometimes be transported/shipped in Winter?
So the laptop might go down to 0-5 C. I have read that LM is for 10C+ operation only. But what about storage?
Could it become solid and splitter off the CPU potentially causing short circuits on the laptop mainboard later?

 

If the melting point is down around -20C you should be ok. Different LMs may have different MPs so that would be the starting point to look up

Condensation is the problem while operating at subzero temps but you don't get there unless you're extreme cooling, water chiller, peltier, LN2 etc


Re the viscosity (or lack of) in Conductonaut I think more care is needed to make sure the die/heatsink interface is as small as possible, and the right amount is used (unlike paste, you can repeatedly check your fit). CLU it seems has a bit more tolerance for both.

I took a little while to build up how much conductonaut was needed for proper consistent contact on my CPU and GPUs and I started small in area too, I did not paint the heatsink outside the bounds of the die/IHS. I have since delidded with it too and the temps are simply amazing. Struggling at stock to run a FPU test without thermal throttle, to a 15% OC that runs cooler. Idle delta down from ~30 to ~10 and that includes a reduction in fan speed - I now have a desktop chip whose fan is almost silent and it's no special 6700K clocker

When I didn't have enough and there was a bit of a gap between the slave GPU and heatsink (fixed with bending), I could clearly see where the stuff would 'pool' on the side of the die that was lower due to the cooler stand, but it has never run beyond the edge of the prepped surface area. If I dropped my laptop on its edge I'd be checking for LM run due to the force possibly breaking the surface tension causing an escape, but then again if I did that *touches wood* I'd have much bigger issues than LM running.

 

Guys can I use this kind of nail polish?

 

Any nail polish with a sense of humor and that is poisonous to children is acceptable

 

Just wanted to make sure. Srry kinda newbie :/ thanks for response btw.

 

Actually I've used LM plenty, especially for customers, but I personally still use Grizzly paste or gelid extreme.

 

much better then ICD7. Infact I think ICD7 is entry level. Anything worse then it shouldn;t be used . period. However I also hate what ICD7 does to a surface in the long run too.
On GPU die's and CPU's die's directly it scratches them, not much an issue on metal, like you can lap it to get rid of that.

 

Where have you been? I left so many commendations for you!!!

 

Hi Mr. Fox, I was wondering if you could give me the name or link to your video showing the differences between these liquid metals? I could not find it on your Youtube channel MrFoxRox.

 

Hi Bro. I saw your post on my YouTube channel making reference to this thread.

The only video I can identify with a comparison of three different thermal pastes was this one.



This was not a comparison of 3 different liquid metal pastes. I think my post in January of 2017 in this thread might have been poorly worded. If that is the case, I apologize for not being clearer in what I was saying. I cannot remember doing a video comparing multiple brands of liquid metal. Perhaps I did and cannot remember, but I looked through my library and did not identify a video like that.

My opinion of liquid metal has changed over time. What you quoted above is no longer my opinion. CLU used to be the best. It was thicker and less likely to run or drip. They changed the composition of it and that is no longer the case. Phobya Liquid Metal, Liquid Ultra and Liquid Pro are thinner and runnier than Conductonaut. I now recommend Conductonaut above the other liquid metal options.

 

Great video and conversation.
But I have a question regarding longevity.
I just purchased a gaming laptop and I plan on repasting with CLU. I'm planning on keeping the laptop long-term, and even if properly applied and not needing to repaste often, I'm worried about how long the heatsinks will last.
Do you guys have any input?

 

Heat sinks do not typically wear out as they have no moving parts. As long as the gas in the heat pipes does not seep out they will last many years. The bigger question for every laptop owner is will their machine have a heat sink that fits well and designed properly. There seems to be a trend toward a "heat sink lottery" that is sometimes harder to win than the silicon lottery.

 

This indeed.
Alienwares have warped heatsinks and improper thermal pad thickness that cause high core temp differentials (bad contact pressure on the side of the heatsink close to the VRM area), and MSI has outrageously convex heatsinks. They use a gigantic thermal stamp by "Lairdtech". The heatsinks are mass produced and the exact same heatsink main block is used on the GT72VR, GT73VR, GT75VR and the 6 core GT75, with the only differences (GT73VR and GT75VR heatsink/radiators are identical), being the heatpipes or heatpipe positions. (the GT75 version has an extra heatpipe going from the main CPU radiator to a sub radiator that is sharing space with the GPU VRM radiator). But you can bet your buns they are as convex as ever.

 

Not the only reason, bruh. You forgot the TRIPOD heatsink design. It will never be perfect even pressure with only 3 legs. Have you seen any desktop MB/Heatsink with 3 or less mounting fasteners? A bad heatsink design for use with Liquid metal.


And the brand New TRIPOD heatsink for AW's new Coffee lake models ain't much better!!

 

Thats what I've been recommending for the last 2 years or so.

 

looks like the N150SD or N170 series...
is it ?


(h I see, it's an Alienware, but very similar)

yeah you can NOT use LM on these, the tripod will not make perfect contact, infact you might want to simply use ICD7 as the thermal compound, seriously, because it's thicker. Thinner pastes like TX-4 from Tuniq, Gelid Extreme, Grizzly's Kryonout - will not work well.

 

Any tips on applying LM on Aorus X9 DT?