Chip cooler launches liquid nitro at CPUs A US company that specializes in CPU-cooling systems has just released its latest weapon in the war against melty microprocessors: liquid nitrogen. Auburn, Washington-based Koolance offers a broad range of CPU coolers, but none that will make a serious overclocker's heart skip a beat more than the CPU-LN2 Liquid Nitrogen …
The liquid nitrogen is not sold as an over the counter product. The liquid nitrogen will still come in big(ish) canisters from gas supply houses.
This cooler is pre-packaged kit that lets users not have to build their own. It is not intended that this is used as a production system... it is for overclocking enthusiasts.
A couple things, Where the hell do you get Liquid Nitrogen anyway? The shipping would be very expensive. And I am sure that DHS/CIA/FBI/[insert government body here] will add you to their list in no time.
Of course the safety groups will get involved since there will more than likely be more than a few deaths / mutilation occurring from ether:
1) Body and -190 °C liquid (not pretty, happened at my University, lab partner no longer had a usable hand)
2) Rapidly expanding liquid in a sealed vessel, not to mention how explosive Nitrogen can be at room temp.
But given the size of the resevoir for the nitrogen on the unit, the difficulties in getting the nitrogen in some locations and the hazards involved in handling/using liquid nitrogen. I suspect that this is going to be a nice thing, assuming it lasts to much longer than the first lawsuit that gets filed when some 15 year old is found dead in his parents basement having mishandled or abused the liquid nitrogen.
1) You buy it from a gasses company. Liquid nitrogen is a well used industrial chemical. Uses include shrinking parts and seals for the tightest of interference fits and breaking fruit.
2) What sealed vessel? Where? I see a tube. And as for that well know explosive INERT gas nitrogen which makes up the lions share of the atmosphere?! Don't light that match!
Tom's Hardware did this back when they wrote good articles. Yawn.
@Iam Me - Huh? You decant liquid nitrogen into a thermos for practical use so it is easy to handle. You only need some thick leather gloves.
Jeez people! Where is your sense of fun. I suppose you're all part of the lobby that got all the good fireworks banned! Who needs a nanny state with adventurism like this??!!
Just get a dewar flask (I got mine off eBay, but that was before eBay sucked so bad) and then look up gas supply company. Take your flask in and that's it. They might be a little bemused to see somebody who doesn't look like a professional buy it, but as long as they see you have the right equipment it shouldn't be a problem.
What they fail to include -- though hopefully they mention it somewhere -- is that you need to put many layers of insulation around the chip. Towels, absorbent clothes, rubber pads, use 'em all. Otherwise, you'll get water condensing around your chip, on the mobo. And you will be extremely unhappy.
I see the greatest "danger" to this as the problem of cold nitrogen gas displacing the warmer oxygen at floor level as the cooler's nitrogen evaporates. With calm air (as in a basement) little Fluffy or Fido might soon be a lot less playful.
Come to think of it, any ankle-biters crawling around on the floor would experience the same issue and be far more likely to generate lawsuits.
Welders that use nitrogen/argon/CO2 have the same issues when working in enclosed spaces. (Not with Fluffy, but with themselves.)
But a small fan to make sure things get stirred up should be all it takes to mix the nitrogen with warmer air.
And I feel sorry for those of you who were (apparently) deprived of the opportunity to play around with cryogenic gases in chemistry class. Driving a rubber-band nail into a chunk of wood with a hammer made of mercury was the stunt I most remember. Yeah, we used to get to play with large volumes of mercury too.
When I worked at a university it was common practice to dip your bare hand in liquid nitrogen and throw a handfull at a colleague. As long as you don't hold your hand in for more than a few seconds it doesn't even hurt. The person you throw it at gets a bit cold and damp as the nitrogen condenses water from the air.
The liquid boils into a gas as soon as it comes near a warm hand so you're only in contact with the gas. The gas is still pretty cold though (-195 C) so will freeze you if you hold your hand in there long enough.
The main problem I can see with this kit is liquid nitrogen is good for getting things really cold, but it has a low latent heat figure so it doesn't take much energy to boil it away. I can't see this pulling much more heat from your CPU than a good water cooling system would.
Thanks for making me chuckle.
Firstly with your proposal to store LN2 in a sealed vessel. No-one else does. (Hint wiki -> Dewar)
Secondly for your descriptoin of room temp N as explosive. Its in the air, donut. Nitrogenous compounds are used in the manufacture of many explosives, but trust me, nitrogen doesnt go bang. They pack your crisps (yank-chips) in it cos it wont react with anything.
And finally for your brainy man icon. That made me laugh out loud.
Liquid Nitrogen is inert. It does not burn or support combustion. If your in a room which is 100% nitrogen you suffocate. It's also the stuff used to store sperm. You may have seen the long handled tongs they tend to use for handling this.
But it is *very* cold so special purpose clothing, face masks and gloves are strongly advised. The potentially big danger is expansion. All cryogens expands hugely when they boil, and they can boil fast. Nitrogen vapour density at normal atmospheric pressure is about 1/196 that of its liquid, and that's at -196. That ratio goes *up* with rising temperature. 1 drop of liquid N2 in a confined space at anywhere above -196c will boil. When it hits a container wall it will start to exert pressure. If the container can't hold it the walls will fail, explosively. Its not how much is trapped. Its how strong the walls and how hot does it get. This cooler presumably works by dropping LN2 onto the surface and venting the boiled gas to lower the average temperature. It should be a very small flow as Silicon based semiconductors work poorly at very low temperatures. That valve had better be *very* reliable. Flooding that surface will a full open valve should (without *very* extensive emergency venting) make quite an effective fragmentation grenade. I can live without this ever being demonstrated. Making sure even small amounts of cryogen does not get trapped in joints, valves etc is a key part of designing cryogenic plumbing.
I think its an expensive and potentially dangerous gimmick. It'd be interesting to find out how often you need to top up that flask as well. So expensive, may be dangerous and tedious to use as well. Still uses should be able to put LN2 handling experience on their CV's.
When handling liquid nitrogen, the key thing is to ensure that no liquid gets spilled into the cuff of your glove from above, or you will lose some flesh or a finger. You can safely pour liquid nitrogen right over your bare hands, provided you don't cup your hands and keep your fingers open - it forms a vapour layer and the liquid runs right off (yes, I have done this repeatedly). But get it caught in a glove, and it has nowhere to go, and pools in the fingers. End of said fingertips...
Way back in the late 70s, we thought that future computing would be done using Johsephson junction superconductors cooled with liquid helium...at 4.2 degrees Kelvin. There was even the requisite cover story in Scientific American that swore it - and I did a Science Congress project on it in 1981. But no one really wants to run a computer at 4.2K (I/O is a bitch), and the liquid gases (especially helium!) are expensive over time. I might just have to buy one of these in honor of those ancient prophecies...
LN2 cooling has been around for many years. No Nitrogen is not explosive at least not as an element. However a rapid expansion of any compressed gas in a container of insufficient strength to contain that expansion will result in an explosion, this is even true of noble gases that have very low levels of reactivity with other elements.
For those wanting to find out how the experts and world record overclockers use LN2 and other methods of extreme cooling check out :
http://www.xtremesystems.org/forums/forumdisplay.php?f=10
I was wondering when someone would come up with this. I dont think it will ever go commercial, but i would love to play around with this and see just how much i could get out of a processor. Hell, if you can use something colder than Liquid Nitrogen (Helium?) then maybe we could make a superconducting processor with infinite speed!!!!!
you'll spend hundreds to get a 50% speed boost from your processor, and then in 6 months a bog standard off the shelf processor will be 10% faster than that with no fancy stuff required.
liquid nitrogen should be kept for it's proper uses - such as throwing into swimming pools.
Admittedly, I never went sticking my hand into dewar flasks, but having worked at two universities with a ready supply of the stuff there's practically no end to the fun you can have - including the 'shattering banana' trick, which we usually did for the kids on open days and long before Heston Blumenthal made it trendy, and chucking a cupful of LN2 onto a warm floor and watching the droplets skid around and evaporate (think water on a hot griddle). Oh, and lest we forget, demonstrating the Meissner effect (again, another crowd-puller on open days)
Having said that, the only machines that were liquid cooled in any fashion back in those days were made by some bloke called Cray ...
Although I don't do it myself, anyone interested in extreme overclocking should have a look on bit-tech.net - got some interesting writeups on extreme overclocking there, along with the various weird and wonderful pots that folk use to get their CPUs (and GPUs) down to insanely low temperatures.
The one with the gloves in the pocket.
Go back to reading the Daily Mail, there are others just like you there, you can be together.
Seriously liquid Nitrogen is easy to get.
"there will more than likely be more than a few deaths / mutilation occurring"
Well come the revolution I hope you aren't on my side with your exploding Nitrogen.
OMG I'M BREATHING IT NOW!!!
Am I going to explode too????
Even Paris knows what Air is made of...
From the Toms Hardware Guide Intel P4 is 84W on 1.12 cm^2. That's 75W per Sq. .
Apollo re-entry heat shield design. 77 W Sq. cm.
However.
MEMS Handbook 2nd Ed (M. Gad-el_Hak) pg 10-29. Potential for cooling IC's by MEMS array of water spraying through boiling. 300 000W Sq. cm. This is not a typo.
You can guess what is in my (over sized) pockets.
"Hire some hot babe to keep your tank full while you game away and you would be the coolest geek in the world."
I suspect that hiring a hot babe to take to your LAN party is going to be quite enough, without the need for dry-ice style effects. Actually, dry-ice sounds a lot more sensible - the temperature range should be better, as well as the availability / handling.
@Christopher Ahrens looks like you went with the bang of thermos N2 filled.
@Steve well that would explain why people in bars are all eejits, the cause is in fact not alcohol but leaky N2 cannisters that lower the O2 level.
The only interesting part is you could inflate your tyres with the N2 leftover, bigger molecules hey.
/the hot air balloon membrane.
It's true that water could remove a lot of heat from a chip if the heat from the chip were enough to boil it to steam. But that's because of the latent heat from the phase change, and requires that the chip run hot enough to boil water. One wants the chip to run cooler than that to perform well.
Ice, being a solid, doesn't do convection well, so it, while also capable of carrying a lot of heat away by melting, is not really a good alternative.
I would think that the best way to maintain a chip at a specific, desired temperature is to use copious amounts of a working fluid just slightly below that temperature. So a system which brings large amounts of rapidly flowing chilled water very close to the chip, and then cooling the water which the chip had warmed elsewhere would seem to be best.
"...because of the latent heat from the phase change"
The MEMS proposal relies exactly on that.
"..the best way to maintain a chip at a specific, desired temperature is to use copious amounts of a working fluid just slightly below that temperature. "
I'd agree that there is probably a temperature where chips work best. The point at which the free electron concentration (the conductivity) is high enough for the chip to operate properly but low enough to avoid any thermal damage. However IIRC modern high pin count chips are mounted face down in the package. All the heat generating action is happening in the top 0.01mm of a chip. That heat then has to diffuse up through the wafer thicknes (0.3-0.6mm total. 0.6mm is a 12" SOA wafer) then the package lid. So if that fluid temp is *too* close to the ideal circuit temp the heat will not be picked up by it as there will be no temperature difference. Untill the chip overheats and stabilises at a temp that will move the heat into the fluid.
I am wary of open loop cooling systems. OTOH a heat pipe allows you to take a fluid and dial in whatever boiling temp you want, within limits. Water at 0.1 atm for example would boil at c45C. Its tricky to find a substitute that's both non-toxic (no methanol), non-flammable (so not butane, propane etc) and not a CFC. The thermophysical properties site for the US NIST http://webbook.nist.gov/chemistry/fluid/ can be quite interesting for searching. You'll be looking at saturation properties of fluid
NEVER EVER EVER put liquid nitrogen in a thermos AND SCREW THE LID ON. Thermos are not perfect and eventually the expanded gas will explode your thermos as the pressure (hence temprature, hence pressure...) increases. Put it in the thermos with the lid loosly balenced on top. This is by far the biggest danger of LN2.
Ln2 is pretty safe I spilled it all over the place cooling a CCD in a microscope at work. It looked remarkably similar to this device a loose plastic T shape dropped into the hole on top. You fill up the dewar while the chip sits in the bottom, DO not put the top in yet otherwise when it vents it will fire the piece of plastic and send it ricchetting around the lab. Also do not let your 14yo work experience "shadow" look down the hole at this point. All of a sudden a whoosh of "steam" shoots out the top as the first amount of LN2 cools the dewar, then it is safe to fill up and drop the cap in but this again just sits somtimes pushing up as the slowly boiling nitrogen vents up past it.
This happened the week before I started working at a university in New Zealand:
TA walked into ground floor mechanical workshop with a thermos she couldn't get the lid off.
Tech noticed it was cold, asked what was in it, blanched slightly when told "Liquid nitrogen" and ran out the door as quickly as possible, ending with a nice overarm throw into the middle of a duckpond outside.
2-3 minutes later, thermos went boom. Resulting browny-green splashmarks on nearby buildings reached up to the 6th floors. (This was pointed out to all staff handling LN2 as an illustration of how not to handle the stuff.)
The mechanical workshop was in the base of one of those buildings. What might have happened if the tech hadn't been a quick thinker doesn't bear thinking about.
Bang, because there's no Brainiac sign.
Well the dewar on my microscope looked about the same size and in a room temp (approx 20C) lab I had too fill it twice in an 8 hour work day (although had plenty left at end of day). I would say allowing for expansion and loss during filling it used about 0.5 litre per filling (less in the afternoon if it was still cool). However this was to cool a CCD so was not really generating as much heat as a processor would.
The reason they stick Liquid nitrogen on CPU's isn't to remove the heat per sae, it's to remove the resistance.
Basically NO electrical resistance =sort of super conducting silicon = extremely high CPU cycles.
Hence (I think stupid) over clockers like to see how FAST they can push a chip before it goes bang or they create a new record or something.
e.g. DX 486 works at petaflop speeds for 9 seconds etc.
.. Liquid nitrogen offers questionable benefits over other methods, is aimed at over-clockers and not the majority of pc owners, is only dangerous if used by non-experts (majority of the worlds population) - hmm...
I'm guessing its not the best idea ever!
Paris, cos she's the hottest icon available (the penguin is a close second)
DHD
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
