Oh that might be the reason why these never got out, they made a CPU heat sink that managed to violate every RoHS regulation at once.

Please reconsider and look if you can find a tech museum or somebody who wants the cooler in their collection. Is by far one on the most advanced cooler around, and the amount of engineering put in it seems ridiculous, I mean you don't see everyday an magnetohydrodynamic pump!

See if they can drain the and clean the cooler. It's an interesting piece to keep on the shelf once it's safe.

​@@savclaudiu2133they're not that rare, he doesn't need to do that.

If you do scrap the cooler, could you send the controller/PSU to either: Kaizer Power Electronics (Denmark) or The Post Apocalyptic Inventor (Köln Germany) They like having this sort of thing to teardown/repurpose.

Asus or msi mobo & overclock management

😅

RTX 480?

Ah yes, the RTX 480, easily the gpu of all time.

I'd say the GTX295 instead, which is essentially 2 GTX480 chips running on the same PCB. Recipe for smoke, essentially.

There are PWR for some reason

"So unless you have a nuclear reactor in your PC... you don't need one..." so the perfect cooler for a 13900K then

You ever see a 4090! That's pretty close buddy!

@@rossclutterbuck1060 I came to say exactly what you said.

>Implying modern processors don’t produce the same thermal output of nuclear reactors. I wasn’t born yesterday, bub.

Exactly my thoughts! As unsafe as this thing can be, it is a rare example of truly thinking outside the box, and it should be remembered for that.

he probably doesnt care about others xD

@@gelo1238 Or cares enough to destroy it and not be a burden on his mind and anyone else's. Either could be true. 😃

@@gelo1238 yeah he is a war criminal

​@@nazgu1It's not really unsafe, no more than any LiPo battery pack fully charged. Even if you managed to rupture one of the pipes (which wouldn't be that easy at all, dropping it would barely bend them) it's unlikely the sudium-potassium alloy would actually catch fire (unless you tore it apart and then drop it in water).

Also NaK is much much cheaper to make, as sodium and potassium can easily found in sea water, while gallium can only be found in aluminum ore with average concentration of ~18 ppm

yeah but such alloys are liquid to 785 degrees celsius, and they are used to cooldown something to 100-200 degrees probably not pc to 30 degrees. same as car optimum working temp to burn less fuel is 90 degrees, so still 10 less before in worse case scenario when you have to fill radiator with bottled water. did that many times in company with shared cars. either i will put my drinking water in radiator or wait till engine will cooldown. cos i didnt had coolant

@@tomaszszupryczynski5453 , bullshit. Na-K eutectics starts from -11,4 °C.

I like how you said Natrium.

NaK alloy was used as a reactor coolant at the UK's Fast Breeder Reactor at Dounreay....

You’re correct. The wires are going into a coil around the block, basically half a transformer. That’s why it’s pulling 30amps, it’s basically a short circuit. I’m kind of surprised/impressed that it wasn’t so electricly noisy that it didn’t make a hum or hiss on the audio recording.

@@leafydialupking1 There is no coil, the field is produced by permanent magnets which are held by that clamp thing (which is actually shielding and guiding the flux more efficiently) on top. Making an electromagnet for this application would be very inefficient and at 30 amps it would require more power than the CPU its cooling. As for noise, this is a DC application, so the only noise would come from the power converter board

"Liquid Railgun" would've been a much better name for this cooler :)

@@WouterVerbruggen I would think for it to be a Lorentz force pump that the wires would be perpendicular to the magnetic shield, since the current needs to move perpendicular to the magnetic field and parallel to the direction of flow. And in that case the inverter module must include a current clamp since the pump would have very low resistance.

@@leafydialupking1 The current flows throught the liquid metal from one side of the channel to the other, the lorentz force pushes on the fluid directly. So with the liquid flow as reference, the magnetic field is top to bottom, current from left to right and thus force in the direction of the flow. The electical "short" is the low resistance of the liquid metal, which I would guestimate to be in the 100 nOhm range so you'd need some expensive equipement to measure that directly.

Ah that explains how some of them made it to the retail market. Thanks!

There's thinking outside the box, and then there's releasing something to the public that could cause them permanent skin damage, lung damage, environmental damage to wildlife... Most postal services won't even let you deliver products with NaK in them as it's that much of a hazard, more often than not you need to order them from qualified suppliers. Yes technically the unit is sealed and the same could be said for batteries... But this is a liquid that can leak if those pipes aren't solid.

shouldn't have ever gone past the lab though LOL I'm sure it'd be a decent type of cooling for very specialized applications

@@sinephase IIRC this kind of cooling (on a much larger scale) is already used in some nuclear reactors? I guess in situations where non-metals would boil, and you can't allow gallium on-site because it could corrode critical components, NaK starts to become a reasonable option as a heat-transfer media?

@@05Matz Right, I know it was used in the "Fast Breeder Reactor" at Dounreay in Scotland.

@@05Matz from what I've been reading a eutectic alloy has worse viscosity as well. They didn't choose sodium for nothing.

Can you explain the second problem it solves? I don't quite understand.

@@queueeeee9000 there it is hahaha read it again xd

@@queueeeee9000 (sorry for a wall of text) Boundary layer effect, when you pump a fluid along a smooth tube, you will get different flow speeds in the tube depending on how close to a tube inner wall you measure. Gas/Water/liq. metal flow creates a drag resistance against tube inner surface. Slow flow near inner surface compared to middle. That sort of creates a layering effect, liquid near wall prefers to travel along the wall. In this liquid metal design, the liquid metal is only getting warmed by the conducted heat transfer from external source at the inner surface of tube. Then, already warm liquid continues to travel slowly along the inner tube while in the center of the tube "cold" liquid rushes past the entire hot area without picking up any heat. Only when the flow is disrupted by a rough surfaces, bend or a shift in inner diameter will there be promotion of mixing the hot "wall fluid" and cold "center fluid". It is possible they have rough surface inside tube to break up these layering effects, impossible to tell without inspection or manufacturer information. I'd guess no. A heatpipe, as used in PC cooling, has different densities inside, one density in middle, other near inner tube surface. Along inner tube wall there are small (sintered) structures that are sparse enough for high volumes of media to pass per second. The opposite is true of the central structure, a denser wick like structure that has an affinity for adhering to liquid: Wick effect. By combining these two structural differences (with low pressure water) it provides several advantages: It promotes high flow of media near inner surface where the heat source is, the sintered sparse structure provides good mixing of media because of the chaotic ways it has to flow. Also because of the phase change from liquid to gas more energy is taken up without heating the water that much. The turning of water into high volumes of gas it will leave the heatpipe hot zone much much faster than a liquid would. Liquid creeps slower in wick towards hot zone, gets hot, turns to gas and zips off to cool area of heatpipe. The volume flow of gas is probably around 1600 times faster compared to the liquid speed in wick, if their respective surface area in tube are equal. As for an approximate real world flow speed or water amount/sec in heatpipes I have no idea, never done any such calculations nor looked it up.

@@Paxmax thank you very much for the explanation. That's incredible. I had no idea how much actually goes on inside the pipes. Incredible. How does the wick in the center, stay in the center?

@@queueeeee9000 its a compressed pellets structure that surrounds the denser center structured "wick".

I don't think you are correct on thermal conductivity. At least, partially. You either need to look at the whole system in a more abstract way, or look at the heat transfer, surface conditions, and flow inside the tubes. Looking at the whole system, assuming ideal mixing of coolant in the tubes, you have mass flow and delta T. Mass flow (not volume flow) is a product of flow velocity and density, and delta T is governed by mass flow, heat input, and specific heat (assuming no phase change). Liquid metals have a much lower specific heat, so you need a much higher mass flow. With mercury and gallium based liquid metals, it is somewhat offset by a higher density than water, bringing the volumetric flow to reasonable levels. However, eutectic NaK is less dense than water, so you need a much higher volumetric flow to match the performance of water as a coolant. If you look at the surface conditions and the flow, you want a high surface surface area, and enough turbulence to mix the fluid. With enough surface area and turbulence, the thermal conductivity of the fluid basically doesn't matter. Hence very fine microfins and jet plates on water blocks. In round tubes, as seen here, its quite possible the improved thermal conductivity of NaK helps performance, unless there are some form of heat-pipe-esque surface on the inside of the tubes on the waterblock side. Also, pretty sure the 'pump' is a magnetohydrodynamic pump, based on what appears to be a giant electromagnet, and very high current. I am by no means an expert on that kind of pump, but my understanding is its basically a 'railgun' for fluid. Now, I though it had to be really long to get meaningful flow, so I could be wrong, but I don't see why you would need 30+A of presumably low voltage to power a normal pump.

How is it pointless? Compare it to a regular water cooler with a pump, which it is similar to.

Were overclocked CPUs of the day capable of operating above 100C? That would be my first guess as to what would possess somebody to miniaturize NaK reactor cooling for PCs, but there still should have been some kind of... oil or something they could have come up with in between! @@LiveType

​@@cdburner5911 Of course, thermal conductivity is not the only important part, but in this case, I still believe it is quite important - you have to take into account that water would require more fiddly pipes (surface area) and a different pump to push through those fiddly pipes (since the overall volume is much smaller than 'normal' water cooling and you are explicitly not relying on the latent heat of phase change to transfer heat away from the coldplate) which means a lot higher pressure, which would require thicker pipes (...). This makes it not obvious whether NaK would be worse with those specific constraints before you test it - you may observe that AIOs require a completely different setup. Assuming the pump in this setup is the limiting factor (looking at its power usage, that's a fair assumption) as long as you want to have a compact MHD pump with no moving parts, water would be a worse coolant - besides everything else, your pump limits mass flow, so lower density as long as the pipes sizes work, is not an issue. It's maybe even beneficial, as it gives you larger pipes in the fin stack, which (assuming sufficient thermal mixing of the coolant) would increase heat transfer to the fin stack (up to a point). And that all is besides the fact you would probably need to use saltwater or other mixture to even use MHD, which would increase corrosion, be even less efficient etc. Of course, it turns out that it's cheaper to make small and reasonably efficient propeller pump and have radiator with large internal surface area of the water pipes (that is then slow moving and readily transferring heat). Note that this no longer fits within the footprint of a standard CPU radiator (I have not seen any AIO that fits into standard cpu cooler footprint) and for AIO with comparable performance would most likely want 2x120mm... Which kinda means, this design is not _wrong_ per se, just their assumptions on what are the important constraints were flawed - it turns out violating standard cpu cooler footprint is not such a huge deal, small propeller pumps are cheap and reliable and completely besides that, heat pipes _work well_. Still - was that whole approach worth the effort? Not really, as we can see. But it sure was interesting!

It's not pointless if you compare it to (liquid) water cooling. Then again it is, because heat capacity was never a problem in water cooling in the first place.

It has no moving parts, but it requires a fan attached to it that does.

regular heat pipes don't have a pump at all they rely entirely on the capillary effect of the sintered copper inside (or copper groves/copper wire mesh if you have a very cheap cooler).

It's also an order of magnitude less efficient though...

Interesting

30A pump

what for, its not heatpipe, and you know how water cooling works when pump breaks, there is no flow so extreme temp on cpu. heatpipes arent full so water evaporates and then at end coolsdown and returns on side to bottom to repeat process. in water cooling its based on circulation, you need pump to push hot water out and push cold in

@@tomaszszupryczynski5453 the liquid metal has a very high thermal conductivity, and unlike water coolers, the tubing on this thing is copper and attached to the fins.

@@blar2112 NaK has lower thermal conductivity than aluminum by an order of magnitude. It would make for a horrible cooler if it's not actively pumped. Tubing material is irrelevant in the context of comparing this cooler with a water cooler too, since the actual water-to-radiator interface is still metal to metal. The channels that the water goes through in radiators is metal.

@@cardboardsnail the amount of metal in contact with the cpu in a water cooler is tiny, plus it has no direct exchange interface with air. This thing is full metal in direct contact with the cpu and air, i bet without the pump it could handle 80W~ no problem, something impossible with a water cooler.

gonna be far worse, like having solid "pipes" and some small convection effects

It's also possible that it required a much greater heat output than what cpu's we're putting out back then. Thermalrights giant coolers had that problem with idle temps and low power cpu's.

aha i see what you mean,its possible the heat wasnt enought to liquify the "wax" in the pipes so there wasnt good enought circulation i think there where some ASUS cooling systems for GPU that had that issue@@sirmonkey1985

@@sirmonkey1985 heatpipes have a limit, once you overpower them they stop working (all working fluid evaporated and not able to condense). and this is the only case where i see this NaK stupidity shine because it can't be overpowered as it's not based on phase change of the working liquid.

blinded by marketing

Y...yeah but ... but NaK ... NaK has lower density than water (it floats) ... meaning even less heat capacity...

​@FTreba It only "floats" as it's ripping the hydrogen and oxygen apart into gasses, it's almost like a liedenfrost effect.

@@Steevo69 Umm, that may very well be, but NaK IS actually about 15% less dense than water.

Try Lithium, as in batteries. Sodium/Potassium need an oxidant (air/water etc) to react. Charged Lithium batteries can let-go with just heat run-away, oxidant not needed!

anhydrous ammonia, hydrogen sulfide, cesium, rubidium...there's a lot to choose from...

@@AKAtheA the limit is knowledge.

@@AKAtheA Ammonia would be the obvious (as in usable) one. It's still used in big commercial refrigeration units.

Thanks, now I'm thinking of heat pipes filled with Hydrazine -.-'

$200 for a block? Insanity. I ran a custom loop for years but got tired of maintenance and went back to air a few years ago. I had my CPU and GPU cooled and spend just over $350 for EVERYTHING

IKR?

yeah they are a fire risk not an explosion risk (without some assistence like locking them in an airtight container to make pipe bombs)

And brainiac on discovery channel showed how just small drop of lithium destroys whole bath thub. ... and thunderf00t and periodic tables did videos showing that alkaline metals in wather is not that a big deal afterall...

so it is like the motor has a single winding with very high current, that is why it needs a beefy power brick.

Do you have a butane torch? Do you have draino/ concentrated soap? This is no more dangerous or toxic than either. Unless put in a closed pressure cooker with water and somehow broke it, there isn't an explosion danger. The chemical burn risk is because it reacts with water to produce hydroxides, also known as soap. I'm sure you are terrified of your dishwasher too.

​@@danielrouw2593everything you mentioned people know to handle with care. But a PC cooler filled with a dangerous substance that looks at a glance like any other safe PC cooler is a disaster waiting to happen. Very very strange to defend this useless dangerous product but you do you 👍

@@RikiHorizon i love how needing to do something intentionally idiotic is now the manufacturers fault. I just love the plastic guards around ujoints on industrial drive shafts that make servicing them impossible, because people like you ignore all the big red triangles and touch spinning things. Keep fighting the concept of personal responsibility and defending nanny states holding your hand. You do you.

Now that we have the snarkiness out of the way, apply your logic to each of the products that people "know to handle with care" when they first came to market. We would have zero innovation. Humans would still be in the stone age.

@@danielrouw2593 I don't think anyone was advocating for making this illegal. Also, leaks and accidents happen, NaK is much more reactive than you give it credit for. If the air is humid it'll burst into flames all on its own. What we're saying is that the company who made this was dumb for making it because while it might have the potential to slightly outperform a heatpipe, a heatpipe is simpler, more robust, safer, cheaper, easier to manufacture, stores better, and doesn't present a disposal hazard. Drano and concentrated soap are what NaK violently reacts to form, after which it begins to react the way you're expecting it to. People like you underestimating the dangers of unfamiliar substances are actually one of the main drivers of nanny state regulation, because you get yourself killed mixing gasoline and styrofoam and your family sues the companies who made them. They then lobby for tighter regulations on substances to lower their legal costs.

...and the engineers were from the nuclear industry.

Mercury's volumetric heat capacity* is 1.89 J cm^-3 K^-1, less than half that of water, though maybe the high thermal conductivity makes up for that? * isobaric, at 25 degrees C. Since mercury is much denser than water, I assume this is more salient than mercury's *mass* heat capacity, but the physics involved is way past my pay grade.

Moving all that mass would require much more power compared to water.

@@arandomcomp2427 True, though water pumps used for CPU cooling are typically 5-15 W. Mercury's density is 13.6 * that of water, so you might need quite a powerful pump (maybe 100W?) to get it flowing, but after that you'd just need enough power to overcome friction losses.

There is nothing moving inside to break. The "pump" is just two electrodes on both sides of the pipe and a pair of magnets perpendicular to them (top and bottom). When the current is flowing through the NaK it generates a magnetic field that interacts with the magnetic field of the magnets and, through a Lorentz force, pushes the liquid along the tube.

@@MikrySoft The pump might not break, but we are talking a pretty viscous "liquid" being looped and numerous metals are involved. All sorts of things can go wrong.

What kind of repair shop throws away your entire PC?

I cant remember their name.. but it was through insurance. So they decided to not repair it and give money instead.

It's sodium-potassium alloy, NaK. Its liquid down to -12c, and its viscosity doesn't change much at these temps. There's a LOT of reasons that is not ideal in a residential consumer product. It reacts with water, just like sodium and potassium. It will also react to atmospheric oxygen and can just ignite. Fun stuff.

@@BadHaddy It's basically a bomb.

Don't exaggerate, it's not going to _explode_.@@spankeyfish But if it gets broken open, the stuff inside will heat up as it corrodes from contact with the air, more-so the more moisture it absorbs from the atmosphere, and if it comes into contact with actual water it'll get VERY hot, and produce a bunch of hydrogen (that might just bust into flames) and strong bases, which sounds like a bad time. It's probably not dangerous to USE... but probably dangerous to throw away, because then you don't know what it's mixing with.

30amps!🤣lmao it just keeps getting worse! I'm gonna be completely honest, man.. this thing is so crazy, I have to respect it. This is the most audacious shit I've ever seen in PC cooling!

He mentions that the important bit is the conductivity between materials, not within the material, but doesn't go into details which is better there, the salt solution or water

They are very different cooling scenarios though. A nuclear reactor outputs a **lot** of heat compared to a computer. Maybe the higher thermal conductivity is wasted because there isn't enough heat gradient? (I'm no material scientist maybe that makes no sense :P) Also, water may not have a huge thermal conductivity, but remember that heat pipes are phase-change cooling, the fluid inside (water or otherwise) is absorbing HUGE amounts of heat when it evaporates, and releasing all of that heat when it condenses in the cold side of the pipe.

LMR and used on a nuclear reactor because the coolant does not absorb neutrons, and it can handle extremely hot temperatures at normal atmospheric pressure. I get worried about my AIO leaking, I can't imagine how I would deal with this cooler leaking!

Danke