Subzero PC cooling ideas/questions

I do understand the principles behind refrigeration. I've even fixed a few fridges and air conditioners in my time. But I'll readily admit that I'm no refrigeration engineer (and my mechanical knowledge kinda sucks a little, too).

My question is whether a subzero cooling system can be implemented within a PC chassis without the use of a compressor. No compressor at all. It seems to me that if the coolant pressure is sufficiently low then it will not require as much "push" back from the cold side (condenser, radiatior, etc) towards the hot side (evaporator, CPU cooler, etc) - perhaps gravity acting on a sufficiently large liquid-state coolant mass in the reservoir would be enough.

I understand that no cooling system can take more heat away than is produced and thus can at best reach equilibrium with the ambient (cannot cool below ambient) unless more energy (some electrical power, etc) is added to the system. I guess that's the real difference between active refrigeration and passive heatpipes. It's also why TEC cooling can be spectacular for spot-cooling but overall it's judged to be energy-intensive and inefficient. There are always thermal inefficiencies and losses, anyhow, I understand that.

But is no-compressor subzero phase-change cooling viable, or is the whole notion fundamentally flawed? I suppose the liquid coolant reservoir might need to be excessive large and not really workable for PC applications, but if it is workable then what sorts of scales would be involved?

Some time ago there was the Vapochill Micro. It wasn't sub-zero, but it was phase change, and didn't quite perform up to its promises.

There isn't a way to get sub-ambient without the input of energy. That either means you need some source of heat (to make cold) that isn't your CPU/GPU, or you need a compressor.

Oooh! Random piece of information from post-apocalyptic book on rebuilding civilization is relevant!

There is such a thing as what is called an "Einstein refrigerator" (guess where it got its name). Rather than operating on the principal of mechanically compressing and expanding gas, it operates by exploiting the difference in boiling points and gaseous pressures of a few different chemicals. It still requires a heat source, but it doesn't really matter what that heat source is (self-regulating ceramic heater coils spring to mind as a good candidate). The actual mechanism is a bit complicated, and I'm a bit intoxicated, so I'll just leave you with this picture and link, courtesy of The Great And Powerful Wikipedia May It Live Forever.

https://en.wikipedia.org/wiki/Einstein_refrigerator

Also, the patent:
ENORMOUS LINK

Not exactly what you described, but it will, I think, achieve the same goal (silent phase change with no moving parts).

Also, since I know someone will ask, here's the book.
http://the-knowledge.org/en-gb/

A cooling system to cool a cooling system? At a glance it seems unworkable, lol. And it does seem to use an electrical/mechanical generator, which sort of defeats the purpose, no? I was kinda hoping on a cooling approach where the only "extra" energy fed into the system comes from fans across the rads, let gravity do the heavy lifting. Might be unpossible.

Multi-stage passive cooling, some combination of vapor chambers and heatpipes and radiators, seems most workable in my mind.

My main concern with TEC is that once electrical power is terminated the "hot" and "cold" sides of the plate thermally equalize - the moment it's turned off, half the heat sitting in the radiator is quickly pumped right back onto the CPU. It seems there's no such thing as a one-way thermal conductor, analogous to an electrical diode.

Quote:

---

Originally Posted by Konrad

A cooling system to cool a cooling system? At a glance it seems unworkable, lol. And it does seem to use an electrical/mechanical generator, which sort of defeats the purpose, no?

---

What do you think the radiator on the back of your fridge is? All refrigeration systems work by extracting heat from one place, moving it, and releasing it somewhere else.

No electrical/mechanical generator; all it needs to operate is a heat source.

Here's a more succinct description of how it works, courtesy of Wired.

Quote:

---

Ammonia gas is released into a chamber with liquid butane in it. This reduces the boiling point of the butane, causing it to evaporate, and draw in energy from the environment, cooling the area outside the evaporator. The mix of gases pass through to a condenser filled with water. The ammonia dissolves into the water, and the butane condenses into liquid, which sits atop the water-ammonia mixture. The butane runs back into the evaporator, a heat source is used to drive the ammonia back into gas, and the ammonia heads to the evaporator to begin the cycle again.

---

The system is very workable indeed; it just lost favor back when it was developed because less toxic freon coolants were developed at the same time, and the inventors lost interested in pushing it commercially. The patent rights were sold to the Electro Lux corporation.

Interestingly, the design has seen a resurgence of interest recently with an eye to bringing cheaper and long-lived refrigerators to the developing world.

Quote:

---

Originally Posted by Konrad

I was kinda hoping on a cooling approach where the only "extra" energy fed into the system comes from fans across the rads, let gravity do the heavy lifting. Might be unpossible.

---

Radiation cooling can never result in a lower temperature than the ambient temperature of the environment in which the radiator resides.

To address your original question details, the change in pressure is where the temperature differential comes from; if you were to run a traditional refrigeration system at a lower pressure you would simply get less cooling.

A cooling system is the same as any other thermodynamic system: energy is required in order to do work (moving heat around), and the amount of work that can be done (amount of heat that can be moved) is directly proportional to the amount of energy that you put into the system.

The compressor-driven refrigeration system works by dramatically changing the pressure (and thus the density and evaporation point) of a fluid. Moving the fluid back and forth between these two pressures allows heat to be moved from one place (the evaporator) to another (the condenser). A cool liquid is pumped into the evaporator, where it expands into a vacuum, thereby drawing heat from its environment. This heated gas is then pumped into the condenser where it is compressed into a liquid, thereby creating much more heat. This liquid must then be cooled, after which it enters the cycle again.

The Einstein-Szilard system exploits the nature of three different fluids (water, ammonia, and butane) to move heat from one point to another. The same basic concept (changing the evaporation point of the solution) is used, but the change is accomplished by mixing and separating dissimilar fluids rather than changing the pressure of a single fluid. This process of mixing and separating is initiated by raising the temperature of one of the solutions.

TECs operate using the Peltier effect, whereby certain materials have been found to create a temperature differential across themselves when a current is passed through them. While very compact and silent, TECs tend to be extremely inefficient, creating significantly more heat than they absorb. Because of this, they tend to be used only when space is a strong concern and there are sufficient capabilities to dissipate the extra heat elsewhere.

There are two other methods of which I am aware for removing heat from an environment: evaporation and chemical.

The former can be seen in NO2 cooling systems, and simply consists of taking a very cold substance and introducing it to a heat source. The cold substance absorbs heat from the heat from the heat source, and releases it by changing the state of some of its volume.

The later is interesting, but likely no more realistic for a prolonged cooling system. There are certain chemical reactions which absorb heat from their environment (endothermic reactions). If you could rig up a system that would feed the reactants of such a reaction into a heatblock and simultaneously remove the product of the reaction from that heatblock, you could in theory creating a cooling system using such reactions. However, I have never heard of anyone creating such a system, likely because it would be a very expensive system to operate, requiring a constant supply of reactants.

Chemically reactive cooling might be viable if a perfect (and reasonably non-hazardous, non-corrosive, non-toxic, non-violent) substance existed. Endothermic reaction at the hot side, exothermic reaction at the cold side, analogous to phase change cooling but perhaps not able to carry as much heat away. Pretty hard to beat the energies involved in phase change, lol. And a compressor would still be involved, I think. Heat pipes still win, I think.

Looking at your no-moving-parts cooler details. I think it might actually be workable ... provided I use a really tall cylinder. Like maybe about 6-inch diameter and 6 feet tall, able to hold enough bulk volume for gravity-potential-energy purposes. Maybe a good industrial smoke-stack kinda theme calls out to me. Assuming the cooler approach works, lol.