Slashdot Mirror
Want a Cool and Quiet PC? Dunk it in Oil
The Last Gunslinger writes "Tom's Hardware Guide has published an article (complete with video) showing how they employed their own approach to the liquid cooled computer. To offset the loss of normal airflow around their Athlon FX-55 and GeForce 6800Ultra, the mad scientists in the lab decided to fill the case up with 8 gallons of cooking oil. The oil temperature leveled off at a comfy 104F during benchmarking operations intended to tax both the CPU and GPU to their limits. Interestingly enough, they first attempted this operation using deionized water. It worked for 5 minutes before developing short circuits...but the hardware was amazingly undamaged." Slashdot has covered similar projects in the past but it was neat to see the differences in oil and the look at capacitance around the CPU pins.
Not sure why you'd want to do this. The benefits (effective, silent cooling) are more than negated by the drawbacks.
For example, if you get water into the system you could fry your machine. Its not that difficult, especially if its not sealed too well. Another example being if the sealing were to catastrophically fail, you'd have 8 gallons of cooking oil that wanted out, and if you weren't at home could very well destroy the board.
Think you're going to try to take this thing to a LAN party? Good luck. Better wear one of those muscle belts to be able to lift and carry it. And better make sure those seals are extra tight. How's the buff Asian guy next to you going to feel when he and his machine are doused in cooking oil?
Job? I don't have time to get a job! Who will sit around and bitch about being broke and unemployed then?
I'm sure in hindsight it's a dumb thing to try, but sometimes you can get unexpected results. I think there was probbably enough garbage on the motherboard to provide enough ions to establish a current. I wonder what would have happened if they had rinsed the motherboard first.
AccountKiller
Well, it was done by the Cray II, in about 1983 (okay, that was with Fluorinert, not vegetable oil, but anyway), so it's not exactly a new idea or anything.
4 2&cid=14422980 (This is the article about the new Corsair watercooling rig)
It's interesting that this came up as an article, because in another thread I'd been discussing it yesterday:
http://hardware.slashdot.org/comments.pl?sid=1733
I think that we're going to see more stuff like this in the future. I don't think vegetable oil is where it's going to be though -- there are a lot better liquids that you can use, which conduct heat far more effectively. I found a place within a few minutes of googling that is willing to sell anyone 1gal or 5gal jugs of light white mineral oil (a petroleum product) for relatively cheap, in various viscosities. I think that would make a lot more sense than using some sort of organic oil that's going to go bad.
And if you were going to use it in anything serious, you'd really want to get 3M Fluorinert. It's expensive as hell, but it's designed for exactly this purpose.
"Ladies and gentlemen, my killbot features Lotus Notes and a machine gun. It is the finest available."
I don't know, the last place I worked at had a LASER welder that used De-ionized water flowing over the flash lamp (~400 volts) to keep the bulb at a set temperature. We would buy distilled water from the grocery store and change the water about once every 3 months. What probably made the difference though was that there was de-ionizing resin in a chamber that the water would flow through on its way to the Flash Lamp. It was really expensive if I remember correctly. I don't know much about it, but it consisted of really tiny plastic like beads about .5mm in diameter that also had to be changed at the same time as the water.
At Los Alamos National Lab, an early star wars prottype, the Beam aboard a rocket program launched a sub orbital sattelite that had electronic dissipating lots of heat for a short interval. Fans don't work well in space. And weight was a premium. The solution was to fill it with parafin. The parafin not only conducted the heat as a solid/liquid but it also has a phase change from solid to liquid which until the transition was 100% liquid clamped the electronics at the melting temperature of the wax. This required no circulation pumps.
Of course once it all melt then you are back to the steady state conduction of liquid parafin. But if you've ever made candles then you know that melting 8 gallons of wax on a stove burner can take a long time. If you can make that last say 12 hours--a work day-- and then let it cool down overnight you might never melt it all (or have two computers and play ping pong: one always cooling while the the other is heating).
Some drink at the fountain of knowledge. Others just gargle.
You've got to watch the thermal range, if you're wanting to do extreme cooling OR run really hot hardware. Some of 3M's synthetic liquids are excellent for this type of project - well, they would be but only a handful of enthusiasts have ever been able to afford them.
Finally, although you only need to extract the amount of heat being put into a total emersion system, you've got to cycle through most/all of the liquid in a reasonably short period of time. You shouldn't rely on the heat simply transferring through the liquid. Besides, if you do that, some regions will be hotter than you'd like and others colder, even if the average is just fine. The average doesn't matter, because no component will see the average.
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
That's a pretty interesting idea. Normally when we think of phase-change cooling it's liquid to gas and vice versa, but solid to liquid phase change is certainly an option too.
What I wonder about though, is whether in a conventional (atmospheric) application, you would end up with voids in the parafin (or other material with low melting point) as it heated and cooled. Obviously this would be a bad thing and could lead to overheating of the chips. I don't know much about the physical properties of parafin -- does it expand and contract as it heats and cools? If so then it seems like it could easily form voids around the chips.
I once worked with a liquid, some sort of long-chain polymer, that had a freezing point of around 40F. If you chilled the whole thing slightly below it's freezing point, that might be able to work in much the same way. Provided of course that it's a dielectric.
"Ladies and gentlemen, my killbot features Lotus Notes and a machine gun. It is the finest available."
You probably don't want something too flammable, but if you can seal well enough to keep water out isopropanol is relatively nonvolatile and nontoxic, it's just that alcohols tend to absorb water. (Another option is propylene glycol, the stuff used in nontoxic marine antifreeze.) A more off the wall option is a suitable molecular weight paraffin. High quality lamp oil is almost odorless and not particularly flammable in bulk. The question really would be how fast convection currents can move in the different options.
Pining for the fjords