• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Has anybody tried this?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
Status
Not open for further replies.

B-Hizzle

Registered
Joined
Jul 12, 2012
Location
Missouri
Hey, I need plans for my future extreme liquid cooled computer. I was thinking you get another air conditioner to keep your parts lower than 20°C (68°F) off and if you can manage it on. The reason for this is to maximize conductivity of metals. You will have to modify your exhaust, use 1 small fan at a very low RPM at the highest point. Now I recommend you some 9x140mm or bigger radiators that will work with Fluorinert and you hook every part in your computer that generates any heat with homemade silver (or any metal with a higher thermal conductivity) waterblocks with very thin selenium (or anything with a lower thermal conductivity) heat guards wrapping them. This is so the waterblocks don't get heated by ambient temperature, but it also means your metals aren't at a 20°C from ambient temperature. To get around this you must liquid cool the heat guards or use a thin liquid cooled separate waterblock connected to main waterblocks made of something like aluminum, beryllium, copper, gold, magnesium, or anything lower than your main waterblock's thermal conductivity but not too low instead of a selenium heatguard. That part needs major expirimentation, I can't do it yet because I'm still building my PCs. Also make sure your waterblocks allow maximum flow, fitting 1/2in (very) thickwall tubing or better. And you can make the waterblocks based off brandname ones, I suggest Koolance or EK, made in a 3D editor, and printed with a 3D etcher or something of the sort.
Now the fun part (finally)
Get or make a high pressure air or CO2 pump (buy a giant empty propane tank and attempt to get it filled with CO2 :)). Get the coldest variant of Fluorinert, a really good intercooler, and a substance like liquid nitrogen (somebody try pure oxygen at -218°C) to cool the Fluorinert without freezing it. Nitrogen may be too cold. If it does freeze add product like acetone, antifreeze, or WD40 and pump fine pulverized solid Fluorinert.
Set your liquid cooling up like this:
DIAGRAM.jpg
(this is an old pic from http://www.octools.com/fluorinert/ by OCTools Staff) instead of submerging your motherboard, simply set up normal liquid cooling (with your beast parts) but just after the pump insert your cooling tank and intercooler out side of your case.
This way I would think you could get an efficient, quiet if you are really good, and you don't need to change it every time your start up unlike DIce. Also this way you may be able to drop passed -183°C and superconduct multiple items in your motherboard, maybe it will get us 1THz without overclocking, it may be really bad and just break it or parts could repel and break solders and stuff.
I recommend you do this on junkers not your i7s. ;) I really hope this works, if it does TELL US ABOUT THE SUPERCONDUCTIVE MOTHERBOARD! If not I hope it's a good way or breaks DIce.
 
I'll try looking for a formula for ≈1K. I think thats where we all draw the line and say, "I would rather the temperature to go up, PLEASE be stable!" as we cross our fingers in hope we don't freeze our computer and everything around it in place for eternity. :)
 
I'm confused as to the point?

  • First and foremost, motherboards don't really do so well at that kind of temperature. Voltage controllers, MOSFETs and pretty much any number of ICs wouldn't function like CPUs do at LN2 temperatures. That's partially why we freeze CPUs, RAM and GPUs directly.Everything you list freezes at LN2 temperatures.
  • All of the materials you list will freeze at LN2 temperatures.
  • Is this is a 24/7 proposition? If so, how would you keep LN2 supplied?
  • If you do as you said and run standard water blocks, what liquid do you propose that won't be frozen as soon as it enters the LN2 chamber? ....what intercooler and/or tubing do you propose that won't shatter at LN2 temperatures?
Basically, I applaud your gumption, but there is no way this could even come close to working.
 
I'm confused as to the point?

  • First and foremost, motherboards don't really do so well at that kind of temperature. Voltage controllers, MOSFETs and pretty much any number of ICs wouldn't function like CPUs do at LN2 temperatures. That's partially why we freeze CPUs, RAM and GPUs directly.Everything you list freezes at LN2 temperatures.
Then freeze it with solid Fluorinert, not DIce. Much colder unless DIce is as low as you should go.
  • All of the materials you list will freeze at LN2 temperatures.
That is okay as long as your pumps can push it, I recommend turning solid Fluorinert into an extremely fine powder by freezing it all, making a capped reservoir with fans to keep it in that state (keep it off the fans). It's like a snow maker.
  • Is this is a 24/7 proposition? If so, how would you keep LN2 supplied?
Yes. You probably will have to keep your cooler running to avoid freezing of its parts. The LN2 or oxygen at -218°C is only cooling the intercooler. Get another reservoir and suspend the intercooler inside and have a n enter and exit point keeping the reservoir air and thermal tight. It may have to be replaced at some point though.

  • If you do as you said and run standard water blocks, what liquid do you propose that won't be frozen as soon as it enters the LN2 chamber? ....what intercooler and/or tubing do you propose that won't shatter at LN2 temperatures?
Fluorinert will freeze but try it, it's really once liquid, now pulverized solid extreme cooling :). Waterblocks aren't very standard, I don't know how a standard sized silver waterblock will hold up though. Of all things I didn't think of tubing, if you keep the Fluorinert in a constant very fine pulverized state the intercooler shouldn't blow, I suggest you use metal piping though, flexible may work.

The main problems are the extreme temperature and keeping the Fluorinert in a fine pulverized solid state. There must be a constant flow of the Fluorinert dust and tubes must be big so it's more spaced out and less likely to stick together. Dry ice or something that can withstand the temperatures and is very slick may have to be added to the Fluorinert dust. Fluorinert must be at your supercooler's temperature when its pulverized or it will stick in the intercooler, and the supercooler chamber must be very close, not touching, to the Fluorinert reservoir and pump(s) to avoid temperature loss and clumping. More pumps may have to be added to allow flow and reduce clumping.
You might go through a lot of LN2 reservoirs or other supercoolers and a lot of intercoolers and pipes/tubes.
That got me thinking a bit more, thanks!
 
How do you plan to pump solid fluorinert?

If you're serious about this, buy a cascade. That'll give you -100 to -150c 24/7.
Build/buy a large enough cascade and you can have -200c 24/7.
 
How do you plan to pump solid fluorinert?

If you're serious about this, buy a cascade. That'll give you -100 to -150c 24/7.
Build/buy a large enough cascade and you can have -200c 24/7.
The solid Fluorinert must be a dust basically. And you may need a lubricant like DIce and a lot of custom high pressure C02 pumps.
 
Sounds easy to do, sustainable, and affordable...

Well the hard part is making it at a temperature very close to what your supercoolant would put it to so it doesn't clump during freezing and refreezing. A little extra LN2 may be needed for this.
 
http://www.ebay.com/itm/3M-FC-40-FL...982?pt=LH_DefaultDomain_0&hash=item19c73b3696

dang that stuff is not cheap at all

wonder if conundrum can hook me up with some unobtanium cheaper
Yeah well it's pretty awesome and fun to play with, as in this case. You should really only need enough to fill your pipes a few times maybe 3 1.5L if you are efficient. Less you have the less likely it is for you to get a blockage then explosion but the less performance you will get. I would use a gal if I did it the real deal, a lot less for tests. Maybe we could use unobtanium heatsinks?:D
 
Dust, in general, does not pump.
Unless you have some sort of special dust pump I'm not aware of.
 
Dust, in general, does not pump.
Unless you have some sort of special dust pump I'm not aware of.

Well when I say dust, I mean a well pulverized. Fluorinert is very heavy too. It will be like a really thick snow form a snow blower that weighs 25% more. I'd say it's worth a try if anybody has a cupfull.
 
I think the first thing you should do is figure out how you're going to pump a powder, any powder, through a cooling loop.
 
I'm confused as to the point?

  • First and foremost, motherboards don't really do so well at that kind of temperature. Voltage controllers, MOSFETs and pretty much any number of ICs wouldn't function like CPUs do at LN2 temperatures. That's partially why we freeze CPUs, RAM and GPUs directly.Everything you list freezes at LN2 temperatures.
  • All of the materials you list will freeze at LN2 temperatures.
  • Is this is a 24/7 proposition? If so, how would you keep LN2 supplied?
  • If you do as you said and run standard water blocks, what liquid do you propose that won't be frozen as soon as it enters the LN2 chamber? ....what intercooler and/or tubing do you propose that won't shatter at LN2 temperatures?
Basically, I applaud your gumption, but there is no way this could even come close to working.

Chil and Intersil controllers have special models that will work subzero (-40c) but generally the PWM controllers all list operating temps from 0-85C.

that is just one thing out of hunderd on the board that wont like being frozen. You are spot on sir :thup:


How do you plan to pump solid fluorinert?

If you're serious about this, buy a cascade. That'll give you -100 to -150c 24/7.
Build/buy a large enough cascade and you can have -200c 24/7.

I havent seen the latter but the former will work well and only cost you a grand or so plus electric :) I wouldn't get one over the internet though (to many issues I have seen), would want a local guy to build it.


EDIT: At OP..

Kudos for thinking outside the box. That is a good trait and wotrth fostering.

Your idea is too complex, requires expensive components and machinery and is not a sustainable process like a closed water loop. Not saying don't keep thinking about it, but pretty much everything you suiggested has been said 100 times before in one fashion or another. From fridges to AC units to LN2 reservoirs. LN2 is crazy 'spensive at most places and wit ha high end system churning through 3-5L an hour with a pot which release hot gas back into atmosphere.. your LN2 bath would become an overpriced LN2 cooker and pressure chamber pretty quickly.

Your best bets at this time are.... Cascade as already mentioned for low sub-zero temps, single stage for ambient or lower temps, or 24/7 capable with ability to go subzero a glycol based "open" loop that you can drop some Ice/Dice into a large reservoir when you want to get a little sub-ambient action going on. Or wait till winter and throw your Rad outside. Even cheaper!

I like your ideas of ways to integrate insulation into the blocks though. thats pretty unique.
 
Last edited:
I think the first thing you should do is figure out how you're going to pump a powder, any powder, through a cooling loop.
Well instead of a Co2 pump get a high power blower, prepare to buy a lot of huge CO2 canisters. A siphon? A vacuum? Make it a paste somehow?
 
Go for it hizzle...

I'm interested in seeing you get this started... ignore the haters. :p
Lol, I'm trying to get the supplies. I'm a bit caught up in a build though, might take me a month or two. And the haters are helping quite a bit, there are many flaws and they are pointing them out so I can get the most logical fix. It may be best to wait until I fix the kinks. Thanks though!

EDIT: At OP..

Kudos for thinking outside the box. That is a good trait and wotrth fostering.

Your idea is too complex, requires expensive components and machinery and is not a sustainable process like a closed water loop. Not saying don't keep thinking about it, but pretty much everything you suiggested has been said 100 times before in one fashion or another. From fridges to AC units to LN2 reservoirs. LN2 is crazy 'spensive at most places and wit ha high end system churning through 3-5L an hour with a pot which release hot gas back into atmosphere.. your LN2 bath would become an overpriced LN2 cooker and pressure chamber pretty quickly.

Your best bets at this time are.... Cascade as already mentioned for low sub-zero temps, single stage for ambient or lower temps, or 24/7 capable with ability to go subzero a glycol based "open" loop that you can drop some Ice/Dice into a large reservoir when you want to get a little sub-ambient action going on. Or wait till winter and throw your Rad outside. Even cheaper!

I like your ideas of ways to integrate insulation into the blocks though. thats pretty unique.
Thanks! To avoid heating the LN2 it must be cool! I must add shutable channel Y valves. After the first good run switch to LN2 tank 2 and before that run it through a DIce chamber. After a good run and the LN2 switched the Fluorinert may cool down. To minimize the effect of the hot Fluorinert I will add a DIce casing around the reservoir. And since its continuous I can add LN2 tank 1 back with a new tank, open both channels, and minimize the Fluorinert's effect further. LN tank 1 on run 2 can cool all the little things like GPU and RAM and maybe the north bridge, tank 2 can cool the motherboard and CPU, then connect back up to be split off and give even heat to both tanks aside from small variables like distance. I can also have a wider section after merging back up to reduce clogging.

You guys are great at finding whats wrong lol. :)


PWM controllers... I know crap about motherboard components what is is?
And what are these Cascades lol? Are they a cooling manufacturer?
 
Last edited:
Status
Not open for further replies.
Back