Thoughts on Refrigeration
Thoughts on Refrigeration
Refrigeration cooling... I've been considering this.
I'm aware there are two main issues with refrigeration: condensation and compressor power.
Whenever you lower a component's temperature below the ambient level, you raise the chance of condensation. Even if you apply cooling directly to the chip, it might still lead to moisture forming on the back of the chip or the board. But I believe I have a way forward.
A real freezer. Most contemporary freezers are self-defrosting. Their concept relies on a heater element that melts frost from the cooling part. The resulting water is then removed. This same method works with dehumidifiers, removing most humidity from the air inside a case. Combined with a sealed design, you can essentially eliminate all moisture from the interior of a computer case, allowing operation at sub-zero temps without condensation despite temperature swings.
An airtight setup would need to be perfectly sealed. The motherboard would have to be fully isolated in a hermetically sealed unit. I imagined routing all cables through insulated walls to the rest of the case, which would contain a second board with peripherals. This would function as a freezer inside the case, holding only the motherboard and directly attached parts. All wiring would pass through the walls and remain sealed.
Once the motherboard, GPU, memory, and wiring are placed in the freezer, it might take several hours to clear the humidity before operating the computer. Additionally, the compressor would need to be powerful enough to manage the heat from the board, CPU, GPU, and memory—something akin to an HVAC or automotive A/C unit, not a typical freezer. Everything else should remain outside the freezer, ensuring it stays sealed after installation.
I imagine an automotive A/C system with a robust motor attached. It would likely require a dedicated box outside the case, equipped with insulated hoses leading into the freezer. This unit would also draw significant power and generate considerable heat. If your goal is to reach refrigeration levels, you'd likely overlook the electricity costs.
Noise is another challenge. Reducing it without compromising ventilation—essential for a compressor—seems difficult. The heat from a strong compressor could warm a small room even in winter, demanding ample airflow that conflicts with noise reduction. This explains why most HVAC systems are placed outside.
Water drainage would function similarly to a regular fridge, with containment and evaporation.
The method of cooling chips directly is problematic because humidity must first be removed. I favor a radiator design with a glycol pump that only activates after humidity levels drop. This would require a humidity sensor to control the pump or valve. A radiator also cools chips, the air, and the board together. Since glycol absorbs more heat than air, it offers better thermal stability during short overloads.
TL;DR
So far, my plan includes a fully sealed, thermally controlled freezer paired with a dehumidifier system. It blends features of an HVAC unit and a standard fridge. All connections would be protected via walls, and a sensor-driven cooling setup would handle humidity.
This is still quite theoretical. What are your opinions and suggestions?
The 'PC in a refrigerator' idea has been done numerous times....and each time with less than positive success. Your best bet would be to use the compressor for sub-ambient cooling via chilled liquid. There's even a section of the watercooling sticky that addresses this. There's also a thread that addresses oil immersion cooling. But hey, you're the expert, let's see what you build and prove us all wrong. Edit: Updated with appropriate links....
PC in a fridge or freezer:
This question asks about the quickest method to destroy a fridge or freezer. The appliance isn’t designed to handle continuous heat inside its casing, so the performance won’t last long. It will shut down prematurely.
The main issue here is...
To maintain components at a certain temperature? Why is this important? What exact improvement does it provide? To enable overclocking beyond a certain threshold? Is the overclock just for that purpose or does it offer other benefits?
I’ve seen various opinions online.
Most people say it was an interesting experiment, but they won’t do it again.
The resulting noise would be quite loud.
:lol:. Is that a better idea?
If this shows how much thought you put into a new engineering project, I think this thread isn’t really for you.:lol:
No insult meant. Not everyone likes this kind of engineering work.
USAFRet
The goal is to offer much more effective cooling than a regular ambient water cooler can achieve by using much colder water.
The reason behind this depends on the person using the system. I’m just thinking about it as an intellectual challenge.
And you’re correct, as I mentioned; there’s a solid reason why HVAC units are placed outside the house.
Even if everything else works out, heat and noise could still be a major problem.
Hmm, it's curious why folks keep chasing these unusual cooling methods when all the investment could have gone into a better processor and standard air or water cooling would suffice. If you're someone who needs to push chip speeds extremely high, just go for conventional big air or water cooling—no need to waste time with the rest.
Why, thank you Stubbies. Why do folks keep trying to build guitar after guitar after guitar? It offers a cool mental puzzle. It fulfills the urge to shape things in your own style. Pursuing hobbies brings joy.
My idea is just about that—imagining better mouse traps, such as this PC freezer. To fix something once others gave up.
The aim is to achieve much more effective cooling than what a typical ambient water cooler offers by employing water that is considerably colder. The decision on this lies with the user, as I’m considering it purely from a theoretical standpoint. You’re correct; there’s a valid explanation for why HVAC systems are placed outside buildings. Even if everything else works out, heat and noise could still be major obstacles. "Much better cooling" – do existing consumer products require it? Through this setup, you might actually lower the operating temperature, but does that translate to improved performance? Current parts are built to function within existing cooling limits, whether air or liquid. Would you notice a noticeable difference at 20°C compared to 65°C? Usually not. Would it work better at -3°C? How much improvement and would it justify the extra power use and noise? By what standard should we judge performance? If your goal is to see a working PC with a CPU running at -3°C, go ahead. For tasks like speeding up large Excel files or boosting gaming FPS, investing in a superior system makes sense. Now, if you could push the system down to -100°C, you might see a real boost—but only about 8 seconds of it, like a car engine running on nitro and LOX. You’d have to rebuild it every night.
Phase change cooling has been around for some time, especially in CPU cooling, though it faces challenges like condensation and limited real-world effectiveness.
You could try an experimental setup, but I’d prefer a chilled oil submersion system first. :lol:
And honestly, I wouldn’t bother with the simpler, possibly more practical, and painful submersion cooling either.
I discovered a compressor capable of delivering up to 2030 BTU at 7 deg. C., with a noise level of 48Db. This equates to roughly 595W, allowing it to keep a motherboard steady at 7 deg. C. with up to 595W of heat removal. Exceeding that capacity would be impractical.
I could potentially lower the noise to around 40dBA, similar to a high-speed fan, but 600W seems insufficient for pairing with an O/Cd motherboard or GPU, particularly given the thermal constraints of a freezer-like environment.