Out of case refrigerator
Out of case refrigerator
So let's imagine I have a box larger than 12 cubic inches that I connected a tubing from the CPU into. Inside this box there would be a CLC refrigerant system—pump, radiator, fan, and a condensation trap for the hose part inside the box. Maybe a radiator sandwich with a fan in between two radiators (one for the refrigerant side and one for the CPU side).
The concept is that warm water from the CPU block enters through a hose into the regular radiator. The fan then blows cool air from the refrigerant side to the CPU side, after which the water is sent back to the CPU. Condensation would stay separate from the main case.
There are additional details I think need clarification. I believe using LN2 is somewhat similar?
Please expand on this. If this had been tried, links would be appreciated.
Thanks!
DbDBlackJack :
This is purely theoretical. I aim to achieve the coldest possible liquid for the CPU block without any condensation, no matter the cost or difficulty.
In short, can we really do it? Lol
I've been thinking about this a while: whenever they go below ambient temperature, condensation becomes an inevitable issue.
If you're looking to push things to the extreme and ignore expenses, consider building an immersed PC. The whole setup would be submerged in a tank filled with a suitable coolant. This coolant would be chilled via a cooling system, just like you mentioned. Everything would be contained in a tank so that moisture forming on the tank wouldn't interfere...
DbDBlackJack :
So lets say I have a >12 cubic inch box that I ran tubing from.the cpu into. Inside this box would be a clc refrigerant system: pump, rad, fan, and condensation trap(for the bit of the hose inside the box). Perhaps a radiator sandwich with a fan in between 2 rads (1 for the refrigerant side and 1 for the cpu side).
The idea is warm water from the cpu block enters via hose into the normal radiator, the fan blows cool air from the refrigerant radiator to the cpu rad, then the water is discharged back to the cpu. Condensation would be kept separate from the main case.
There would be a little more to it. I think LN2 is somewhat similar?
Please expand on this.
If this had been tried, links are appreciated.
Thanks!
Edit: grammar
The goal of this is to deliver sub-ambient liquid to the CPU water block (else why do it) so that means warm(er), moist ambient air will condense out it's moisture onto the water block and hoses leading to it. You have to account for that inevitable condensation (unless you live in high-dry desert area) on those parts because it will migrate to the motherboard and into/under the CPU socket.
So if the fluid moving to the CPU is below room temperature, it can lead to condensation forming on the block or hose? I understand how it works. If the surrounding temperature is around 25°C and you send in 23°C liquid, what would cause condensation? I’d appreciate any educational resources you have on this topic. It seems similar to reducing ambient temperature through central air conditioning.
DbDBlackJack:
If the fluid moving to the CPU is below ambient temperature, it might lead to condensation forming on the block or hose? I understand how that could happen. If the surrounding air is around 25°C and you introduce 23°C liquid, what would cause it to condense? If you have any educational resources on this topic, I’d love to review them. From what I see, it appears similar to reducing ambient temperature through a central AC system. There are several solid Wikipedia entries on dew point and psychrometry that could help clarify things. Many variables come into play, and I haven’t fully grasped the details—my knowledge comes from observing condensation at the right humidity and pressure levels. For instance, even with a small difference like 2°C, moisture can condense, especially when humidity is high and atmospheric pressure is stable. You might be able to manage this by directing warm air away from the VRM or using a fan. I recommend obtaining a psychometric chart for your area to identify conditions where condensation is likely. But honestly, I’m surprised by the idea of using an external phase-change cooler to achieve such low temperatures. If you’re only cooling by 2°C, it doesn’t seem worthwhile from my perspective. Since you’re near ambient, perhaps focusing on a more efficient radiator instead would be better. An all-copper 360 radiator (or two) could supply cool liquid under any conditions without the added expense and complexity of an external system.
EDIT: For further reading, here’s a useful Wikipedia link: https://en.wikipedia.org/wiki/Psychrometrics
Hypothetically, is it possible to achieve the lowest temperature for a liquid near the CPU block without forming condensation? After all, whatever the cost or difficulty.
DbDBlackJack:
This is purely theoretical. My goal is to achieve the coldest possible liquid temperature for the CPU block without forming condensation, no matter the cost or difficulty.
In short, can we actually do it? Haha
I've been reading about this situation for a while: whenever they go below ambient, condensation becomes an issue.
If you're willing to go all out and ignore expenses, consider building an immersed PC. The whole setup would be submerged in a tank filled with a suitable coolant. This fluid would be cooled via a cooling system, just like you mentioned. The entire assembly would be enclosed so that moisture forming on the tank wouldn't interfere with the internal components. The coolant needs good dielectric properties, safety for electronics and plastics, and must stay pumpable at low temperatures—something like a Flourinert might work? I'm not sure.
In my concept, the chilled fluid would be drawn from the chiller into the CPU and GPU areas, then released directly into the tank, likely near the VRM. The return path, probably at the top of the tank, would pass through a filter before heading back to the chiller's heat exchanger to reach the coolest temperature the fluid can handle. It would definitely be an engineering challenge to perfect all the details, but it could be a great project.
Enjoy! Immersed PC builds are interesting to follow... there are some YouTube videos about them. I'm not sure I've seen any that actually tried sub-ambient cooling yet, so if you're interested, there might be an opportunity for your project!
Ahh, now it makes sense. It doesn't matter what the ambient is; if you go cooler than that then you will condense—I'm sure there are other factors involved, but mostly for the most part—and that condensation would spread across the hose back to the CPU. Thank you so much!
You'd need to make sure the liquid stays warm enough. I'll do some research. Thanks again!
DbDBlackJack shared some concerns about condensation spreading across the hose back to the CPU. On the CPU there’s a water block that could drip onto the socket, between the pins and the motherboard, leading to corrosion and mold formation if moisture is present. It’s important to note that even with ambient cooling fluids, you can manage this without a power-hungry phase change system. If you’re serious about using phase change, there’s a video worth watching for more details.
Let's be precise here. Water cooling isn't about cooling itself; it's about transferring heat. Since water acts as an insulator, it merely relocates thermal energy, similar to a heat container. Removing the fan from your CLC leaves you with a large, ineffective heatsink. The water won't reach 20 minutes before becoming warmer than the surrounding air, and then it actually adds heat to your CPU.
LN2 offers phase change cooling. The actual cooling happens through the evaporation of LN2. This process quickly raises the temperature of the nearby air, making condensation likely. As you observe, transferring cooler air into a much warmer environment is only feasible if you manage to expel the expanding N2 through insulation that maintains ambient temperatures on the other side. Ideally, you'd use thick, double-walled pipes with another fluid to handle the wide temperature swings.
For maximum cooling efficiency, consider starting with solid components if possible. Many people overlook this, but water-based kits are preferable due to direct application (you can purchase blocks for VRMs, chipsets, and RAM) and their inherent heat capacity, rather than relying solely on cooling performance. Today's top coolers remain air-cooled solutions (compared to CLC contenders).