Liquid Metal CPU cooling.

Sounds like an interesting idea, seems a lot like heat pipes we use now, althought with some enhancements.

Hopefully someone with more knowledge will chime in here.
 
Ballz2TheWallz said:
ugh saw this years ago over in the OCC forums,or something similar
Click to expand...

Sorry to waste your time man. Just thought it looked cool.
 
i would expect to see direct die cooling take off before i see this as being terribly feasible.

yes, it does look neat, and i understand why you posted it, but the tech is...........unlikely to be feasible.

i kind of hope that in a few process generations, that we will be seeing a move such that instead of a straight up heatspreader, we'll see something more like a interated and sealed cooling jacket with inlet and outlet nozzels, allowing for closer contact between coolant fluid and the core.

then the old optimization process for increased surface area vs increased flow begins anew.
 
DFI Daishi said:
i would expect to see direct die cooling take off before i see this as being terribly feasible.

yes, it does look neat, and i understand why you posted it, but the tech is...........unlikely to be feasible.

i kind of hope that in a few process generations, that we will be seeing a move such that instead of a straight up heatspreader, we'll see something more like a interated and sealed cooling jacket with inlet and outlet nozzels, allowing for closer contact between coolant fluid and the core.
Click to expand...

That's if companies feel they want to heavily invest into watercooling or some sort of liquid cooling for the retail market. Seems unlikely, as the average person doesn't want to deal with the hassle of liquid cooling. Perhaps enthusiasts will, but I don't think that's feasible for the retail market.
 
DaLurker said:
That's if companies feel they want to heavily invest into watercooling or some sort of liquid cooling for the retail market. Seems unlikely, as the average person doesn't want to deal with the hassle of liquid cooling. Perhaps enthusiasts will, but I don't think that's feasible for the retail market.
Click to expand...

maybe it will be feasible, maybe it won't.

i'm not talking about watercooling the way that it is at present. i'm talking about something more akin to what apple runs with. it's not actually water, it's fully sealed, it doesn't require customer maintenance.

something that would be as safe and easy to install as a heatsink, and would be standard equipment.

a mature and easy to use technology.
 
Think that's cool...

These guys are my neighbour in the research park...

SprayCool Technology is the process of using liquid evaporation, or phase change, to cool electronics. A fine mist of coolant is sprayed onto electronic hot spots and immediately evaporates. The vapor is then captured and the heat is rejected as it circulates through a heat exchanger. This results in an extremely efficient method of cooling.
Click to expand...

http://www.spraycool.com/html/index.html
 
Is this "liquid metal" really a complete metal, or a sort of homo or heterogeneous mixture of metal atoms in another liquid? It might say somewhere on the site, but I couldn't find it. I will say that I did skim it though, because I have homework to do :(

Of course, I doubt they would want you to know exactly what is in the substance, and probably have a patent on it :)

I wonder how easy it would be to make something like this on your own, if you could get your hands on coolant. All you would need is some physics knowledge dealing with magnetic fields and electricity.

My only concern would be the magnetic fields generated by these silent pumps. Does anyone know if the magnitude of their fields would be more or less than the average aquarium pump?
 
DarkenReaper57 said:
Is this "liquid metal" really a complete metal, or a sort of homo or heterogeneous mixture of metal atoms in another liquid? It might say somewhere on the site, but I couldn't find it. I will say that I did skim it though, because I have homework to do :(

Of course, I doubt they would want you to know exactly what is in the substance, and probably have a patent on it :)

I wonder how easy it would be to make something like this on your own, if you could get your hands on coolant. All you would need is some physics knowledge dealing with magnetic fields and electricity.

My only concern would be the magnetic fields generated by these silent pumps. Does anyone know if the magnitude of their fields would be more or less than the average aquarium pump?
Click to expand...

DC current = constant magnetic field = no induced currents to worry about in other computer hardware.
 
DarkenReaper57 said:
Is this "liquid metal" really a complete metal, or a sort of homo or heterogeneous mixture of metal atoms in another liquid? It might say somewhere on the site, but I couldn't find it. I will say that I did skim it though, because I have homework to do :(
Click to expand...

That's really interesting. I was trying to think of what metal would be liquid at low temps, but it never occurred to me that a suspension of metal particles in some other fluid could also work. Sounds like a corrosion nightmare.... :D
 
This looks really cool....but here is where I think cooling will be going in the future.

1) We are moving away from MHZ towards processor efficiency.
2) Multi Cores will facilitate this.
3) Increaingly, the CPU is being used less and less (Gfx, on GPU, Sound on SPU, dedicated lan chip, dedicated storage controllers, even physics calcs in the future, etc. etc.)

Increasingly, we are systematically removing the bottlenecks of performance with enhance memory, lower latencies, wider data paths, faster storage interfaces, more mature programming...etc.

I really believe that we will start to see less and less heat dissipation in the coming months and years and ultimately, this will mean we will be comfortably in the realm of air or passive cooling. At least for processors, for gpu's i think it will remain a burden for some time. Who knows what will happen there. I would imagine simply better implementation of SLI.

M
 
zer0signal667 said:
That's really interesting. I was trying to think of what metal would be liquid at low temps, but it never occurred to me that a suspension of metal particles in some other fluid could also work. Sounds like a corrosion nightmare.... :D
Click to expand...

what metal? mercury comes to mind. cesium would be really close. not that you'd want to use either of those ............
 
Mysogonist said:
This looks really cool....but here is where I think cooling will be going in the future.

1) We are moving away from MHZ towards processor efficiency.
2) Multi Cores will facilitate this.
3) Increaingly, the CPU is being used less and less (Gfx, on GPU, Sound on SPU, dedicated lan chip, dedicated storage controllers, even physics calcs in the future, etc. etc.)

Increasingly, we are systematically removing the bottlenecks of performance with enhance memory, lower latencies, wider data paths, faster storage interfaces, more mature programming...etc.

I really believe that we will start to see less and less heat dissipation in the coming months and years and ultimately, this will mean we will be comfortably in the realm of air or passive cooling. At least for processors, for gpu's i think it will remain a burden for some time. Who knows what will happen there. I would imagine simply better implementation of SLI.

M
Click to expand...

this series of strategies rquires a lot of standardization within the industry, something that is not always to get different hardware vendors to agree on.

it also requires a lot of optimization on the part of game and other software developers, which seems to be lacking thus far...............
 
I agree with Mysogonist, from what I have heard heat will be less of a problem with the multi-core processors and as far as software ppl having to customize their work for dozens of different hardware vendors, it won't happen, all hardware vendors will have to make their particular sound or video card or whatever conform to a standardized interface protocol, the software ppl will just use their side of the interface API and let hardware vendors worry about the implementation on their hardware. If their implementation doesn't conform, then they won't sell many of their cards. As late as last year I was very excited about building a watercooled system but seeing as I'm putting a cheap 3200+ in the system I'm building now and planning to upgrade to a multicore next year, I'll stick with air cooling now. I don't think its worth the time and effort to invest in watercooling.
 
Here is a link to some metals that are liquid.

http://www.scitoys.com/scitoys/scitoys/thermo/liquid_metal/liquid_metal.html

In the photo above, I am holding two small vials of liquid metal. The vial on the right contains gallium, an element that melts at 29.76° Celsius (85.57° Fahrenheit). The vial on the left is an alloy that contains gallium, indium, and tin, and melts at -20° Celsius (-4° Fahrenheit). (Both are available in our catalog.)
Click to expand...

It looks interesting to me.
 
Somehow the idea of using liquid metal seems a little... strange maybe?

Off the hand, I know of three metals that are liquid at or slightly above room temp.

There is quicksilver: Toxic, highest vapor pressure found in metals, and a denger with some other metals. For example, put a drop of quicksilver on a piece of aluminium. Amalgamation will take place, a lens-shape of aluminium-quicksilver amalgam slowly sinking through the aluminium, with needle-shaped aluminium crystals forming on the other side. Needless to say, not good for your waterblock.

Then there's Rubidium. Reacts explosively with slightly moist air. Bad as well.
Same thing with Caesium.

These things all seem a tad dangerous to me.

Edit:
IndyJoe, those are some of the most dangerous alloys. Sodium and Potassium stay liquid at room temp if melted and mixed, but will react violently if dropped, which causes a dramatic increase in surface area. In the presence of oxidizers or CFCs the mix will detonate unpredictably.

Not nice either.
 
Mercury is also a metal that's liquid at a variety of temps, just has very toxic effects and is not fun to clean up when you spill it. Stupid dots everywhere! But since they say non-toxic and environmentally friendly, it can't be mercury.
 
Only a guess but it might be this stuff.
http://www1.mscdirect.com/CGI/NNPDFF?PMPAGE=1758


This one shows its content and specific heat:
http://www.flexbar.com/PDF/18_19.pdf

Edit: on second thought, it would kinda suck if your coolant solidified halfway through your heat exchanger. I dont have a solution for that.

Personally I always wanted to get some of this stuff and mess around with it. ;)

I didnt read the application closely but the material they are using would have to be ferros (magnetic), in order for the 'pump' to work?
 
IndyJoe said:
Here is a link to some metals that are liquid.

It looks interesting to me.
Click to expand...

Both of the materials you listed are extremely expensive.
 
I think there's very little chance this is actually a liquid metal,; as nothing fits the required temperature regime and is 'non-toxic'. My guess would be colloidal, or micro-encapsulated particles in water. Gotta be big enough that exerting force on them drags the liquid, so a metal chelate is probably out unless the concentration is very high.
 
I'm surprises no one has mentioned here that drivers in loudspeakers have been cooled in similar fashion for quite some time now...Ferrofluid cooling has been around for ages...

MD
 
MD_Willington said:
I'm surprises no one has mentioned here that drivers in loudspeakers have been cooled in similar fashion for quite some time now...Ferrofluid cooling has been around for ages...

MD
Click to expand...

Yep, and the magnet on the speaker suspends the ferrous liquid in the voicecoil gap on its own- no need for encapsulation, etc. Unfortunately, it isn't very efficient, but in a loudspeaker, "noise" wouldn't be an issue so most of them use forced-air cooling. In a computer solution, I can see ferro-fluid being VERY useful. I think I have some experimentation to do.
 
I can hear it now

**In best Ahh-nuld voice**

My CPU is a neural-net processor, a lerning computer, cooled by the t-1000: Liquid Metal!
 
You must log in or register to reply here.
Back
Top