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

New Idea for hybrid submersion cooling system.

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Droolguy

Registered
Joined
May 6, 2010
Location
Bend, OR
Ok,

So I have been playing with the idea of building a submersion PC for a while now trying to figure out the best route.

I have an idea but I am not sure how well it will work, theoretically I think it will rock from the little bit I know but I need some second opinions.

Anyway, here goes.

-------------------------------------------------------------------------
-------------------------------------------------------------------------
1. I wan't to submerge a PC in FC-72 due to its 56 degree boiling point.

2. I would then put some high pressure waterpumps in the system and pump the FC-72 into submerged components with waterblocks still on them but with no return line since the computer is already submerged.

3. The FC-72 would vaporize when pumped through a hot GPU/CPU waterblock and the gasses would be released into the cooler liquid surrounding the components. The expansion would pull a great amount of heat from the components due to a phase transition.

4. The surrounding liquid would cool the vapor FC-72, as well as the extra space in the container would act as an expansion chamber.

5. Could add a refrigerated line to the case to further promote cooling (FC-72 doesn't turn into jello untill -90C so its perfectly safe).
-------------------------------------------------------------------------
-------------------------------------------------------------------------


So thats as far as I have gotten so far, any help would be appreciated, trying to build an almost completely self-sustaining supercooling system.

Thanks again everyone.

P.S. Hi! Looking forward to being a member here at Overclockers.
 
so you would have a constant flow of FC-72 to the block(s)? and it wouldn't be a loop? o_O
 
That's a very interesting concept, it's certainly a possibility.

As long as the FC-72 doesn't eat anything, i don't see why it wouldn't work.

Last time i checked that stuff was brutally expensive though.
 
You're not going to see any phase transition, not even with a high-flow pump. Even if it did the concept is also missing a fundamental part of any phased cooling system, a separate section in the loop to dump waste heat.
 
There will be a separate refridgerated (phase cooled) line to pull the heat out of the whole enclosure, not just specific components.

Basically a phase cooler would be cooling the liquid in the tank to be used in another open loop semi-phase cooler. With the added benefit of being completely submerged for total coverage of all components.

I am having difficulty finding someone who can build an air tight enclosure that will handle the pressure caused by the FC-72 vaporizing.

I was planning on using a 40-50 gal industrial aquarium (1 inch thick walls) that is half-full, but they are telling me that the acrylic won't hold up to constant sub zero temps on the phase cooling line... I'll probably end up using one anyway though.

I could just take the waterblocks off completely and cool the bare card, it would be much simpler and cheaper. I just don't know how well the vapor will displace itself in that situation. I suppose I could try that first.
 
Well, the guy who was supposed to be getting me the FC-72 fell through, does anyone know of a company I can buy from? Its not easy to find.

Looking for 10-20 gallons.

Will post a mock up of the design I am planning to use for the whole system tonight.
 
I had a similar idea but with mineral oil. I still plan on trying it one day but first I need a job. XD

Only problem I see here is keeping it airtight. You would either lose FC-72 through boiling or blow the case apart. You would have to have a very strong airtight seal. I would take a fish tank and build a steal frame around it to hold it together.
 
Pretty sure fluorinert is like $40/gallon, isn't it?

I'd go welded steel on the enclosure, then you just need a decent welder and some plate.
If you go ape with it it'll hold hundreds to thousands of PSI, should be fine.
 
Only problem I see here is keeping it airtight. You would either lose FC-72 through boiling or blow the case apart. You would have to have a very strong airtight seal. I would take a fish tank and build a steal frame around it to hold it together.

Right now I have a 24X24X18 1-inch thich acrylic seamless (one piece) aquarium on standby with an airtight top, it should hold up to the pressure, or it'll be an EPIC fail. Either way will be great on camera.

Pretty sure fluorinert is like $40/gallon, isn't it?

I'd go welded steel on the enclosure, then you just need a decent welder and some plate.
If you go ape with it it'll hold hundreds to thousands of PSI, should be fine.

If I get it at under $100/gal I would be estatic since it is not a consumer product, its an industrial one that can be quite hard to find.

As for the welded steel it would conduct the heat perfectly and be strong but I don't know if it would seep impurities into the Fluorinet and cause it to be conductive.
 
As for the welded steel it would conduct the heat perfectly and be strong but I don't know if it would seep impurities into the Fluorinet and cause it to be conductive.

That's a good point. I think you should build a steel box frame to go around the aquarium. Then it will hold hundreds of psi. Maybe even thousands. OR the side would blow out in a Kodak moment.
 
So after a little more research into Fluorinert... its going to have to be a metal case.

Fluorinert will cause plastics to become brittle over time.

Yes this includes the components of the computer, but they won't be supporting massive amounts of pressure.

Still trying to find a new supply of FC-72... no luck yet.
 
I had a similar idea but with mineral oil. I still plan on trying it one day but first I need a job. XD

Only problem I see here is keeping it airtight. You would either lose FC-72 through boiling or blow the case apart. You would have to have a very strong airtight seal. I would take a fish tank and build a steal frame around it to hold it together.

Not trying to shoot down your idea, but I literally just scraped this project after 3 months of work. I thought the idea of using waterblocks on components that would feed into the main pool would work great, but what I realized is that any chemical that isn't electrically conductive and would still provide some level of cooling for the blocks is far more viscous than any standard pump (and by that I mean a water pump or any pump that home depot would sell).

I double checked with alot of Chem Engineer professors at my school looking for a possible solution, but there simply isn't a non-conductive chemical that is pump friendly (unless you want to use distilled water with an electrolyte filter, which is expensive).
 
than any standard pump (and by that I mean a water pump or any pump that home depot would sell).

I will ONE DAY do this, no funds/time right now. But I was planning on using a pump my uncle has at his shop. It's basically for pumping oil out of the grates and into barrels. If it will pump dirty shop oil it'll pump mineral oil.
 
I double checked with alot of Chem Engineer professors at my school looking for a possible solution, but there simply isn't a non-conductive chemical that is pump friendly (unless you want to use distilled water with an electrolyte filter, which is expensive).

Fluorinert basically flows just like water, it's just twice as heavy.

It is completely compatible with any water pump you can think of.

Unless those engineers work with super-computers, high powered lasers, or hematology, I doubt any of them would have ever even heard of it.
 
Anyone have an idea how to keep the case pressurized and get the cables out?

I was thinking about stripping the outer casing off of all the wire groups and running them through a small silicone filled tube.

Think it will work?

EDIT:
Reading more into this, I think I will move to FC-770.

Much lower cold tolerace (-127c) and it will not vaporize at realistic computer temperatures (95c) so I don't have to worry about the case pressurizing.
 
Last edited:
if something went wrong with the CPU, it can hit 95c pretty quickly .. how fast would it vaporize?
 
As pressure builds the boiling point rises, as long as the case can hold a PSI or two it should't be an issue. To be sure you'd need the vapor pressure/temp chart for it, but that should be available.
 
I will ONE DAY do this, no funds/time right now. But I was planning on using a pump my uncle has at his shop. It's basically for pumping oil out of the grates and into barrels. If it will pump dirty shop oil it'll pump mineral oil.

Interesting, how quite is it / what are the dimensions? More power to ya if you found a working pump.
 
Back