Liquid metal TIM?!

Framework Laptop 16
1

“The Framework Laptop 16 leverages a thermal system developed in collaboration with Cooler Master, with a liquid metal thermal interface, three heatpipes, and two fans that keep CPU performance stable and quiet.”

This was a nice surprise for me, and I think further cements the desire for performance that Framework has for this laptop.

I’ve never used liquid metal (LM), but I’m aware other companies have. Most notably in the PS5, and some of the laptops from Asus, off the top of my head.

But as with all liquids, it’s prone to leaking, and can short stuff out. I’ve also heard it can “dry out” if not coated evenly, and can also ruin certain metals if it makes contact with them.

I wonder whether FW has their own special blend of LM, how they apply it, what steps they’ve taken to prevent leaks, whether this will extend to the discrete GPUs, and the level of maintenance a user could be expected to perform, if any. Will we possibly need to purchase aftermarket LM?

Thanks

2 Likes
2

Liquid metal thermal interface materials should, in theory, last indefinitely if the two surfaces are prepared properly. The gallium will have a slow reaction with copper but is essentially nonreactive to nickel. It’s very likely that the cooler will be nickel coated copper - this is pretty standard when using LM.

1 Like
3

Liquid metall is not water, or doesn’t dry, it doesn’t flow away and it certainly doesn’t age. The only reason to not use it, it’s that it’s annoying to handle, so yes you need to pay attention to not spread it around when removing the cooler from the Mainboard

1 Like
4

Liquid metal oxidises, needs replacing every two-three years. Phase change material is superior for laptops. Put some Honeywell PTM-7950 on and you will get much stronger longevity with similar temps.

2 Likes
5

I will choose PTM7950.

2 Likes
6

I’ve been running liquid metal for 7 years now. Never did any maintenance, temps as day one.

3 Likes
7

In the same boat as @Christian_Elsner, been running LM in both of my desktops one of which is going on 5 years old and it’s still running as well as it did when I first applied it.

1 Like
8

Yeah, I had some wrong ideas on LM TIMs. I imagined they were more like mercury, in how they flowed. Thanks guys!

Excerpt on LM from a PDF on TIMs

"Liquid metals have high thermal conductivity and excellent
wetting to surfaces of objects. In addition, no soldering or
surface metallization is required, and a very thin bond line
can be obtained.

However, the liquid metal must be contained so that it doesn’t
spread, and it can be difficult to handle and apply.
Fortunately, once applied, its surface tension will typically
keep it where it is desired."
Source

I would have thought LM in desktops would be okay since they don’t exactly move around often. Compare that to a laptop that will almost certainly experience more vibrations and changes in orientation over its lifespan.

Unless the surface tension is enough to keep the LM from moving out from between the chip and heat sink, I’d expect you’d need a more robust solution in a laptop, but that it shouldn’t be too hard to engineer a good one if you know what you’re doing.

You can find posts on reddit of unhappy Asus customers, but I can’t tell if Asus themselves are cutting corners, or applying incorrect amounts of LM, QC issues, or just underestimating how wild it can be. [1], [2], [3], [4][4a]. [4] seems like a fan controller issue but [4a] is a bit concerning if legit.

But they’ve certainly made a fair few laptops with the stuff, so perhaps they know what they’re doing by now.

I want to believe FW is approaching the problems from another angle. It certainly doesn’t look like anything I’ve seen before.

FW L16 Cooling Method
FW L16 Cooling Method651×606 88.5 KB

1 Like
9

Yes, it’s certainly a pain to apply, but I believe framework might actually better equipped to handle that than asus. Less volume also means the line workers are better educated.

10

Also don’t put your running laptop in weird position on the edges/side to avoid leaks… :wink:

11

I understand where you are coming from, but I really hope it isn’t this sensitive. I want to be able to pick up my laptop and carry it around, or close the lid and slide it into a backpack/bag and not have to worry that moving it around will dislodge the liquid metal and ruin cooling (at best) or possibly short something out on the motherboard.

I hope they have something reliable and relatively rugged figured out.

1 Like
12

Although there is truth to this, it also depends on the application.

Liquid metal when mounted vertically can shift to different spots on a chip if it becomes hot while in said position. This is however dependent on how much material is applied as well as how tight the mounting pressures are on the cooler. The PS5 was alleged to have issues with the liquid metal causing shutdowns while being used in the vertical position, but its important to note that it wasn’t because the TIM was interacting with and shorting any components. The TIM had concentrated on one side of the heat spreader and no longer provided enough coverage over the top of the chip. They did do a good job of containing the material so no dead boards or anything like that.

2 Likes
13

I mean my desktop cpu is vertical 100% of the time. How can a position be worse than that?

4 Likes
14

I think biggest issue (which probably leads also in leakage) is, that people use to much of liquid metal. It’s highly viscous, so using to much leads in leakage. My hypothesis is, the reason for this is the force between the solid and liquid material is to low. To imagine more easily: For example, if you put water between two very thin layers metal, it will stick there(completly). But if this layer (of liquid water) is to thick, it will just ran out, if you arrange it vertically. So i think, that’s the principle why it can leak. And also be mentioned if liquid get hotter, it is normally also more viscous, which means it needs a even more thin layer to not leak out.

And if you are soo worried about it, then you probably don’t be technically proficient enough. If you can’t invent strategies to prevent leakage by yourself, maybe you should not apply it then. It’s to risky then i guess.

So if you need a step-by step turtorial for super-noobs, you should not apply it i guess. Sure, everyone needs to know how to apply and how much and so on and watch/read guides, but if you need it foolproof, then i guess it’s not the right thing for you. I would not risk it then.

15

PS5 owners also seem to have issues related to LM as well. Not exactly leaking out or “drying up”, but more like dry spots, causing oxidation in some parts of the die and/or HS plate.

16

I dont know, why this happens (does not happen normally on vertical-mounts on desktops), but he does way too much liquid metal on the DIE. Man it is like a sea on it. Glad there is this foam around. Would probably be a good idea to use such foams, when using luqid metal.

17

I belief oxidisation occurs because of liquid metal based on gallium. It will also move about more if the contact pressure is not strong enough. Laptops or consoles from the looks of things do not have strong pressure, they never have done.

18

But does the Frameworklaptop 16" has liquid metal on the CPU and on the gpu (the internal one) and on the gpu of the expansion-bay graphics card?

Because i discover right now the honeywell tpm 7950 thermal pad. Watch on youtube the video from LTT (and other vids). It seems to be same as good as liquid metal while not beeing that difficult to apply (and risk of damage) AND holding long (because it does not dry out like thermal paste).

If one of these compenents does not have liquid metal, this would maybe be a great alternative to liquid metal and its risks. Or maybe for a eGPU, which you can use with the laptop.

19

Well… no, it does actually have a shelf life and a degradation period, it just happens to be in the years, 1-3 years depending on the temperatures it runs at solidly and how long the system operates a year in hours.

Eventually, even non-conductive thermal pastes like Arctic will degrade, the performance degradation is gradual and most people simply don’t notice it or write it off on other issues. Liquid Metal wears out much faster than other silicone/epoxy/polyurethane/acrylate, some degrade rapidly due to overclocking for instance. Liquid Metal in an extremely hot system might need to be replaced as often as 1 year to get full performance.

It never stops working fully, but it does degrade. That is… just a fact of TIMs nature. Pumpout will shed TIM as the surfaces expand and contract at different rates since most IHS and coolers are different materials (IHS often are plated with nickel, coolers are bare copper) so some liquid metal will always escape. Powering down your machine and powering it back up again is one of the worst things for liquid metal, and will degrade it pretty rapidly.

Though the polymer pastes can actually separate over time, this really only happens at EXTREME temperatures. So, in general, liquid metal is vastly shorter-lived than its non-conductive contemporaries. Unfortunately, anyone here saying “Running liquid metal, 9 years in now!” are either uninformed or just not noticing the heat increasing. There are absolutely ZERO ways to get around pump-out, it is INVARIABLE when two surfaces meet with an intermediary between it. Period. It’s a fact of physics.

20

Unfortunately, Linus is… not an expert. Honeywell pads are good, but they are not as good as liquid metals or even silicone/urethane compounds. It’s usually several degrees higher than either, 5+ above most silicone.

But they are EXTREMELY convenient, he did get that right… but honestly, they do actually get damaged between applications so it’s in this weird spot of the more you take it out and use it for other systems it gets damaged, compressed in odd ways, frayed and such.

Linus is… a paid advertiser. A lot of people are fans but the man will push anything for the right sponsorship. He has a huge company with a lot of mouths to feed and rents to be paid. Those pads just are not on par with any of the compound solutions. To be frank compound solutions really aren’t that hard to use either, the real issue was back when people had all sorts of crazy ideas on how to applicate to things, but nowadays we are fairly well off in terms of information and guides out there for the application. For instance, Ryzen being a 2-3mm dot in the center, and Intel having a 5 dot with 2mm dots in an X pattern with a dot at each point of the X and a central dot in the middle. But even then you can just apply a nice pea-sized blob in the middle and get some spillage when you squish and be fine.

So long as it’s not conducive of course. Which is where it gets tricky, but so long as you are thorough with your cleanup during and after application, you generally should have zero issues. The conductivity is only a problem once voltage is applied.

If you are worried about liquid metal… just grab some Arctic 4 or a competitor’s alternative! Thermal Grizzly, Cooler Master, and even Noctua is in the paste game now! Of course, if you want to be ABSOLUTELY OVERKILL you can get Prolima which competes absolutely on par with most liquid metals out there. Like, better than Kryonaut and sometimes cheaper, too? It’s suspended zinc so it really conducts well. Arcit is on the lower end of conductivity but 8.5wmk is still good enough, but Prolima will outperform while Prolima is right up there with most liquid metal.

There are some liquid metals, though, like CLP which cost $27 a syringe and outperforms most everything. All these will work fine, Prolima though is just fantastic though. Spreads nicely, non-conductive, performs in line with cheaper liquid metals… lasts years longer than liquid metal. Doesn’t suffer from pumpout anywhere near as bad as liquid.

Anyone who says Liquid Metal is a 15c drop, it’s a load of it. Remember thermal pastes are only meant as a gap filler. The difference at max is about 5-8celsius for SEVERELY poor pastes and top-end pastes.