PDA

View Full Version : submersion cooling


falconsforge
03-01-2001, 03:09 AM
Wondering if anyone has played with submersion cooling
What sort of liquids seem to work best, minimal residue ect
3m's fluorinart looks perfect but way to expensive,
Idea's comments
alan

Cryonosis
03-01-2001, 01:27 PM
hehe. . .I love submersion cooling. . .but would never use it. I tweak with my hardware too much and wouldn't be able to if I submersed it.

I have a couple ideas for it. . all of them include moving the flourinert around but in different ways. so here we go:

1. have the flourinert enter the chamber where the computer and compnents are and then have a drain at the opposite end so that the flourinert moves across the components. between the drain and the intake you can put in a cooling setup to chill the flourinert.

2. place the computer components and flourinert into a freezer chest and use high power HSF combos on all of your critical components (like if you were air cooling). the flourinert would get really cold from the chest and be moved around by the fans.

3. create a large whirlpool effect parallel to the board with fans and then have peltiers cool the liquid at one end. Would be difficult with additional components in computer.

well anyway. . .3m is about the only way to go to get good stuff for use in total subersive cooling. great Idea, so much so OCZ is currently working on making it comercially available. but get this, you have to order the complete system and be content with never modifying the hardware again as the components are sealed in to prevent leakage. I don't like it much since you can't really play with/upgrade the hardware once you finalize your first revision as you have to drain the liquid and lose a bunch to evaperation while you let your motherboard dry.

well anyway, still having fun,

------------------
©××××ק===============
»»»»»»»»Cryonosis«««««««««
©××××ק===============
moderator for Broken Pixel http://www.brokenpixel.com
webmaster of the Kingdom of Ys http://cryonosis.8m.com

falconsforge
03-01-2001, 06:09 PM
Tweaking the gear shouldnt be to much of a problem, just pull the board out and plug unplug away. Messy maybe, Yes its a hassle if you use a sealed system with flourinert but there are a lot of other dielectic fluids out there.
Mineral oil, silicon oil, and a new one made from a vegetable oil, that shouldnt need a completely sealed system.

Cryonosis
03-02-2001, 02:11 PM
the big problem I see with using those materials, oil and silicon based, is the thin coating that will form on your motherboard and components. . . unfortunatly I can only speculate the problems that could occur, such as connectors getting a dielectric coating and not working right. also I think the thermal qualities of oil and silicon based materials is rather poor for use as an alternative coolant.

------------------
©××××ק===============
»»»»»»»»Cryonosis«««««««««
©××××ק===============
moderator for Broken Pixel http://www.brokenpixel.com
webmaster of the Kingdom of Ys http://cryonosis.8m.com

Freakyfrank
03-03-2001, 05:07 AM
the prob with submersion is that u cool every component, including the things that should NOT be cooled.. (bios etc)

also u'r dealing with a lot more wattage to cool.

it would be nice if it worked.. but its a bit roundabout.

Cryonosis
03-03-2001, 11:29 PM
actually total passive submersion is the best known method for cooling a computer. current supercomputers utilize this method for cooling their components.

and I'm interested in knowing why you wouldn't want to cool the Bios. conventional science would dictate cooling would invite less internal resistance within the chip providing a more stable bios.

------------------
©××××ק===============
»»»»»»»»Cryonosis«««««««««
©××××ק===============
moderator for Broken Pixel http://www.brokenpixel.com
webmaster of the Kingdom of Ys http://cryonosis.8m.com

Freakyfrank
03-04-2001, 10:13 AM
i'n not sure it was the bios or other components..
and i don't know at which temp..

we all know that guy with fluorent (and liquid nitrogen)

and components on his mobo broke when he cooled the mobo.. (it could be the condensators or the bios or etc)
but i dont know the link and temps.. http://www.procooling.com/ubb/frown.gif

and these supercomps are build for these circumstances..
but no probs when u stay above the -50C..

well it could not be the bios http://www.procooling.com/ubb/smile.gif

------------------
OC'ing is my middlename

Cryonosis
03-04-2001, 05:18 PM
actually when he slowly brought the tempurature down to -150 degrees the componenets actually functioned properly. . .the only problem he ran into was that the flourinert kept turning to gel from being so cold. It's just when you cool it to rapidly that you run into problems.



------------------
©××××ק===============
»»»»»»»»Cryonosis«««««««««
©××××ק===============
moderator for Broken Pixel http://www.brokenpixel.com
webmaster of the Kingdom of Ys http://cryonosis.8m.com

falconsforge
03-04-2001, 05:51 PM
So what else can be used without spending $500 a gallon on flourinert, besides being way to spendy its also a shitty green house gas.

Cryonosis
03-05-2001, 12:13 AM
kiddo you need to do some reading before your ranting. Flourinert is an inert alternative to flourochlorocarbons. THERE IS NO REACTION BETWEEN FLOURINERT AND OZONE.

now then, there aren't any other real alternatives to Flourinert, most diaelectric liquids have poor thermal conductivity which is a major problem when your trying to cool components.

------------------
©××××ק===============
»»»»»»»»Cryonosis«««««««««
©××××ק===============
moderator for Broken Pixel http://www.brokenpixel.com
webmaster of the Kingdom of Ys http://cryonosis.8m.com

Alives
03-05-2001, 02:05 AM
there used to be a website for a guy called dr. ffreeze. He had a special styrofoam cooler with his computer in submersion cooling using mineral oil and a window AC unit to cool the liquid. The link i have to the website is down.

falconsforge
03-05-2001, 02:32 AM
Hardy a rant, you are correct when you say that Fluorinert has no interaction with ozone but that does not mean it is not a green house gas.
This is a direct cut from 3m's product specifications.

"Fluorinert liquid FC-40, a perfluorocarbon (PFC), has a high global warming potential
and a long atmospheric lifetime. As such, it should be carefully managed so as to
minimize emissions."

Im not dissing Fluorinert, it looks like a great product, I find it hard to believe its $500 a gallon great but hell if i had the bucks id probably buy it.

The question is, what is the next best thing?
Idealy it would be a lot cheaper, not require a completly sealed system, to facilitate tweaking, have at least better thermal conductivity than air and minimal residual.

Cryonosis
03-05-2001, 10:21 AM
that's a good question. . .the only other items I can think of would be mineral water and oil/silicon based liquids. All of them are poor thermal conductors when compared to water and the oil/silicon based liquids leave a nice coat on anything you put in them. I think that's why flourinert is so expensive. . .there isn't really anything else like it. I'm sure there's some no-name company somewhere that makes a similiar product for half the price or less but I don't know of any.

------------------
©××××ק===============
»»»»»»»»Cryonosis«««««««««
©××××ק===============
moderator for Broken Pixel http://www.brokenpixel.com
webmaster of the Kingdom of Ys http://cryonosis.8m.com

Freakyfrank
03-05-2001, 04:41 PM
Originally posted by Cryonosis:
actually when he slowly brought the tempurature down to -150 degrees the componenets actually functioned properly.


well, eventually the mobo broke http://www.procooling.com/ubb/smile.gif

falconsforge
03-07-2001, 05:15 PM
Well I have been harrassing some industrial chemists in the hope of finding some alternatives, so far best suggestions are solvent style hydrocarbon dilectics (should minimise residual), isoparrafins are the other suggestion to date, they are checking out some specifications for me and I should have some details soon. ECM (electro cutting machine?) fluid is another option, it has good heat transferance properties but a little higher residual than id like.
All have good viscosity at around 0c and run out at a couple of bucks a litre.
Will let you know what the chemists come up with.
Saw a review of the OCZ immersion cooled rig,
they have gone for a mineral oil as well, no detail on what type thou. Apparently they plan to bring out a DIY kit as well in a couple of months.

Freakyfrank
03-08-2001, 03:28 PM
ok some advice for ya:

most dielectric fluids (when pure) become somewhat conducting when minerals and/or salts are dissolved.

pure water's conductivity is nigligible.
but you don't want normal tapwater on ur mobo..

grep
03-17-2001, 08:54 PM
Pure deionized water is used in hi power radio amplifiers and Power companies use it to wash hi voltage insulators on live systems. However one thing that must be considered is that the mineral contamination of pure water is caused by the motherboard itself. Water is a universal solvent and will eventually disolve anything. The power companies use a total loss system, and the Radio Amps have to change the water constantly to prevent contamination.

Ron

Joe
03-17-2001, 09:27 PM
Yep De-Ionized water can be used for a short while but once it picks up minerals, and essentially ionizes, its as good as tap water.

fasterstill
04-14-2001, 11:55 AM
did total submersion a couple of years ago.used a mineral oil used for transformers called diala.works but very messy . Bios chips don't like sub zero temps it throws up multiple boots at start up and can loop for several miniutes before starting windows.Also found problems with the oil swelling some components after several weeks of immersion. :) if anyone wants more details just ask!

p[ara]d0x
04-14-2001, 07:18 PM
I remember reading an article on OCZ's peltier-chilled mineral oil submersion system. Can't find it now tho. If you could find a waterproof case it would be fairly easy to do. Just pump the oil through a radiator and/or a few peltiers to cool it. Only thing I wouldn't really care for with an oil-immersed system is changing something on it. I getting oil all over myself :D

captain
09-25-2001, 04:20 PM
I love to try submersion my self. But I would not cool the liquid to subzero temperatures. I would be happy with 3 degrees celsius.
The reason I want to submerge my system is to avoid condensation problems.

If I submerge a computer even at room tempurature I can put 2 172watt peltiers on 1 172watt peltier and get the processor temperature down to about -25 to -30 degrees.
There is no need to cool all compnents down to subzero temperatures. I can also put pieltier on the chip and graphic card without worring about condensation.

Dix Dogfight
10-18-2001, 03:18 PM
Hey captain why don't you build a airtight case with watercooling on all critical components. Place the radiator outside the case.
Place a lot of humidityabsorbing salt in the case and voila, you don't have to worry about condensation on your peltiers.