The best way you get sub zero is by phase change where you evaporate a fluid on either the CPU block or in a reservoir and make a water chiller.

The other way people get sub zero for 24/7 is by using thermoelectric coolers, which essentially pump heat away from one of their faces and move the heat of what you want to cool, plus a whole load of their own heat to the hot side, which then needs to be cooled.

Full mineral oil submersion is an interesting workaround though, my only real complaint would be, the mineral oil will act very strongly (a lot more than air) as a heat transfer medium so you will want to insulate the cold loop a lot more.

Perhaps a good thing to try might be converting an AC unit, submerge the evap and make an oil chiller instead of a water chiller.

 

Yes the conventional methods to achieve sub zero cooling do produce Condensation tho which from my understanding is a pain.

Yes I was drunk when I wrote this and missed alot of key points.

Though as long as the loop can stay at a low temperature or Phase change, Cold/Heat exchange into Mineral oil should be tolerable. Though I do wonder what would happen if you went Sub-Zero on Mineral oil and have an extensive loop system to keep at cooled, with Thermoelectric coolers or Phase changers all chilling the Mineral oil loop.

Though I'm sure having every component on your board running at Sub Zero may have its issues =)

 

There's no way to avoid condensation once you drop temperatures below the current dew point. All you can do is proof the setup to disallow the outside environment from interacting with crucial components that could be damaged by water.

If you want to use a water chiller and want to avoid condensation I would get an adjustable thermostat so that you can tweak that to your dew point.

 

Quote:

Originally Posted by HarryBahzack;14339519
There's no way to avoid condensation once you drop temperatures below the current dew point. All you can do is proof the setup to disallow the outside environment from interacting with crucial components that could be damaged by water.

If you want to use a water chiller and want to avoid condensation I would get an adjustable thermostat so that you can tweak that to your dew point.

The dew point is irrelevant if the whole setup is in the liquid phase anyway.

If the whole rig is submerged in oil the only place condensation will form is on the outside of the tank/case where it doesn't matter anyway.

Quote:

Originally Posted by Neokolzia;14339476
Though I'm sure having every component on your board running at Sub Zero may have its issues =)

Fair point, would probably be best to aim for the highest temp you can keep everything working. I'd probably look at between -20 and 0oC but I'd frankly be happy with around 10oC. (assuming I was using a converted AC unit)

 

Quote:

Originally Posted by Crabid The dew point is irrelevant if the whole setup is in the liquid phase anyway. If the whole rig is submerged in oil the only place condensation will form is on the outside of the tank/case where it doesn't matter anyway.

To submerge the PC just for the sake of avoiding condensation when much cleaner and just as effective methods exist is stupid.

I took as much credence to his submerging points as I did his continual use of dry ice.

 

Quote:

Originally Posted by HarryBahzack To submerge the PC just for the sake of avoiding condensation when much cleaner and just as effective methods exist is stupid. I took as much credence to his submerging points as I did his continual use of dry ice.

To be honest, I quite like the idea of having a fully submerged build and AC unit to cool it all. Assuming you don't get issues from freezing components it will offer very good cooling for the motherboard and RAM.

And hell, half the stuff in this part of the forum are insane impractical ideas. The more wacky ideas that actually see the light of day (of at least a glimpse of light through a basement curtain

) the better imo.

 

Seeing as how I have a mineral oil submerged system, I think I have some input that could be valuable.

Not every part of the system is submerged. You'll still have some parts that are exposed to the air. This is mainly the PCI bracket area and mobo I/O area. You do not want these parts submerged as well because the oil will wick up the cables and get everywhere. Mineral oil is not fun to clean up, and changing anything inside the system becomes a complete and utter mess.

While condensation would be greatly reduced by submerging the system, it wouldn't be eliminated. You'll still have condensation forming on the bracket of the video card and on your USB, ethernet, firewire, and any other ports on your mobo, since you won't want the cables that plug into this area to come into contact with the oil for cleanliness reasons. Also, mineral oil is less dense than water, so if any water does get inside your system, it will sink to the bottom. You should be able to see the fluid interface at the bottom when this happens. If you think you see something that looks like water in the oil, drop in some food coloring to see if it mixes with the mystery fluid. Food coloring is water soluble and will not mix with the oil.

 

Quote:

Originally Posted by HarryBahzack;14340144
To submerge the PC just for the sake of avoiding condensation when much cleaner and just as effective methods exist is stupid.

I took as much credence to his submerging points as I did his continual use of dry ice.

Well this is the only method I'm aware of (new one I suggested) that can offer submerging a PC fully in Sub-zero Liquid.

And I would have Cables and everything inside the setup, USB etc. (if I were to do it, and run those through Silicone, (Air tight submersion environment with airtight access seal.)

I would probably run everything with biggest and baddest Air Heatsinks that are possible to maximize heat transfer to components.

And yes it is possible that everything will wick up the cables but this could be easily combat'd with conventional methods to attack condensation, at a much smaller scale.

________

Now apparently 10C is the lowest useful temp of Mineral oil before it gets too thick.

Halocarbon 0.8 Oil apparently has a gel point of -100C, So... I see potential =o All the same stuff non-corrosive, non-conductive, etc.

 

Annnnd...Then there's the fact that not all chips in a system like cold. Hard drives certainly don't. Sound chips won't. Nics certainly won't etc etc. As well as any other chip on the motherboard other than say the cpu NB and SB.

 

Quote:

Originally Posted by Elloquin Annnnd...Then there's the fact that not all chips in a system like cold. Hard drives certainly don't. Sound chips won't. Nics certainly won't etc etc. As well as any other chip on the motherboard other than say the cpu NB and SB.

In current tech even the sandy bridge CPUs don't like being frozen.

A lot of people on phase change have apparently been getting similar results to watercooling.

 

It would be a nice thought though if every component was capable of running at -100C, =) that would be a cooling enviroment.

 

^ and expensive to run.aswell as near impossible with current tech.

 

Quote:

Originally Posted by MightyMission;14383182
^ and expensive to run.aswell as near impossible with current tech.

Could just toss the computer in the -150C room some of these new High-tech Resorts have.

Cryogenic chamber therapy, roughly about -150C =), now obviously expensive but hey... all in theory.

 

the only way i see who could work without creating a mess would be to create a case to house the pc

and make it air tight ... and then vacuum the case once your setup inside the case is prepped ... no air in = no condensation

voila

+ its allready been done in the past on xtremesystem iirc

 

Quote:

Originally Posted by Fr0sty the only way i see who could work without creating a mess would be to create a case to house the pc and make it air tight ... and then vacuum the case once your setup inside the case is prepped ... no air in = no condensation voila + its allready been done in the past on xtremesystem iirc

Linky please? AFAIK; creating a full vacuum would be damn near impossible for us mere mortals. The pressures that would be put on the case would be so large that said case would probably have to be made of bulletproof Lexan or something like that...

I too have thought about this whole issue and the one thing that keeps coming to mind is that it would be much easier to replace the air in the case (lots of humidity) with some dry/inert gas that won't create condensation...

So one could simply pump helium or nitrogen (in their gas form) into the case and push all the regular air out of it... That would negate the condensation problem and would be much easier to do than creating a full vacuum.

 

just google "goliath"

 

Quote:

Originally Posted by CL3P20;14732472
just google "goliath"

Heh... I just found that thread over at XS. Impressive is the least of it, but it does not really apply to what the OP wanted to do here. Goliath dealt with condensation by simply cooling the air inside the case to the same temperature as the liquid running through the loop; therefore, negating dew-point and condensation.

The OP here expressed that he wanted to have the components inside a case where the air is vacuumed out...

The issues I see with this method:
- Creating an absolute/perfect vacuum is impossible; at least for normal folk like ourselves. Best we can hope for is a partial vacuum.
- Going with a partial vacuum still leaves air particles inside the enclosure and therefore a chance for condensation.

I'm not trying to shut the OP's idea down... I'm actually very interested in seeing something like this happen as it would be the perfect way for subzero benching; I just don't see it as a very viable alternative... I think cooling the air in the enclosure (like in Goliath), or replacing the air with an inert/dry gas is a much easier way of achieving the same result (no condensation).

EDIT: I just got to page 14 on that thread and found out he did vacuum out the air, but he also re-pressurized the chamber with Nitrogen.

 

You wouldn't need a perfect vacuum. Lowering the pressure lowers the dew point, so you only need to get sufficient negative pressure to prevent condensation at the lowest temperature that the air will come into contact with.

Also, having an entirely mineral oil submersed PC (horizontal or vertical), rear brackets and all isn't too much of an issue - if you use short extension leads up to the air pocket then any wicked mineral oil comes out of the sockets in the air pocket and flows back into the oil below it. The cords you plug into the extensions won't wick, because they're not submerged.

The way I see this working is to have a sub-zero (phase-changed) CPU, but submerge the whole motherboard + block. It doesn't matter if the oil turns to gel around the coldest points, because you don't want it to conduct heat to those points (because a gel does not flow, and oil / gel has a high specific heat capacity, but mediocre thermal conduction). It therefore actually helps to insulate your phase change (or appropriate alternative) design, which increases efficiency, as well as preventing condensation on those points.

Your case will already need to be water-tight on 5 sides, so it wouldn't be too much of an issue to make the whole thing air tight. It's been done before, in Goliath. After all, you only need 2 cables to actually come out (assuming no optical drive). Power and display. Those can be sealed in, and the rest can be done wirelessly...

This way, it should be possible to have a sub-zero CPU (yes, ok, sandy bridge doesn't react well to it but others have done in the past and may in the future) but the rest of the components would just be oil-cooled. If the oil gets too hot, then a supplementary cooling system can be added which isn't sub-zero, so as to avoid the gel problem.

My knowledge of phase-change cooling is practically non-existent, so I don't know the constraints of the process - someone would have to work that out themselves.
I do have a question though. What happens to a hard drive in a low pressure environment?

PS. Sorry for the thread bump...

 

Doesn't oil coagulate at sub zero temps?

 

Quote:

Originally Posted by Rampage Jackson Doesn't oil coagulate at sub zero temps?

Yes...

 

Look into a Slush box!

 

Slush box? My first thought of "slush box" is one of these:

To prevent the oil from coagulating? Not a bad idea, but that would result in the same consistency as a slush puppy. Lumps. Maybe lumps would still work...

The only other slush box I can find is an automatic transmission vehicle, which I guess wasn't what you meant...

EDIT: I've completely forgotten how to do image tags >.<
Further EDIT: There we go...

 

Quote:

Originally Posted by zsolmanz;15002065
Slush box?

Actually; "slush box" is a term used in the overclocking/benching world quite often...
Here's an example of one:

http://www.overclock.net/dry-ice-liquid-nitrogen/715668-dice-isopropyl-alcohol-slushbox-version-ii-4.html

 

http://forums.extremeoverclocking.com/showthread.php?t=101902

I keep a few of these in the basement for various things and stuffs

 

Quote:

Originally Posted by xXSebaSXx;15008977
Actually; "slush box" is a term used in the overclocking/benching world quite often...

That's cool - thanks for the heads up