SFF Thermal Solution: Mineral Oil?

[zovc]

Yeah, I think one of the main draws to mineral oil is that you only really need pumps and radiators, and maybe tubing to get the mineral oil where you need. I'm not sure if water blocks are more efficient than heatsinks, but I would assume that an air cooler-style design would probably dump heat into the mineral oil better than just a "passive" water block.

The vapourised fluid is captured in the loop and passed through a radiator (air-air heat exchanger) where it condenses and flows back down into the tank.

So--with this kind of setup--do you even need fans or a pump?

 

[jtd871]

@zovc, read this: https://www.pugetsystems.com/submerged.php and then come back and ask follow-up questions.

This is an old page - they don't sell these kits anymore due to the aforementioned patent troll.

 

[zovc]

Not only did I link to that page, but my last question was not about mineral oil. I was asking EdZ about Novec and other immersion fluids that he mentioned.

 

[EdZ]

So--with this kind of setup--do you even need fans or a pump?

Pump: if you very carefully design the container, tubing and radiator geometry to prevent vapour being trapped as it tries to rise, or fluid pooling after it condenses, then you can operate without a pump. Depending on radiator design, you may end up needing a small pump to move condensed fluid from a reservoir back to the main tank just out of practicality (i.e. to avoid having a system that is a metre tall 10cm wide column).

Fans: like with passive watercooling, it depends on the size of your radiator and the thermal power you need to dissipate. The calculation is slightly easier, as you don't need to worry about the thermal efficiency of the radiator changing with the temperature difference in the fluid: for unpressurised phase-change, the temperature is effectively constant (once the system has warmed up to stabilise at the boiling point of the working fluid), and you only need to worry about the fixed enthalpy of vaporisation. But the short answer is the same was with watercooling: big rad = no fans, smaller rad = fans.

 

[zovc]

Mineral Oil, or perhaps one of these phase-change fluids you're opening my eyes to, is making me reexamine an old design I was starting to consider. Alluded to in these two threads:

Can case panels serve as heatsinks without a "block" and "pipes" connecting them to components? (I think we can agree mineral oil is probably a better plan than this. Haha)
Large, thin radiators? (Thread includes some pictures of a 3D concept component layout.)

Basically, I'm wondering if we could make this kind of design work without needing to use extra space for a pump/tubing/fans. The notion definitely makes me think that we could make a really thermally efficient, tiny case.