Mini fridge liquid cooling pc

[GraySenshi]

I know if you directly put a pc in a fridge it will eventually heat up the box and just keep getting hotter. But what if you put a radiator on the door and had it pulling air from the fridge and export it out and threw the rad

 

[firefoxx04]

Probably not a good idea.

 

[InvalidError]

Fridge compressors aren't designed to pump 50W worth of heat. They are designed to efficiently cool little more than insulation losses.

If you've ever had a new fridge or left one unplugged for a while, you'd know that an empty fridge takes hours to get down to temperature.

 

[beoir]

I know if you directly put a pc in a fridge it will eventually heat up the box and just keep getting hotter. But what if you put a radiator on the door and had it pulling air from the fridge and export it out and threw the rad

You have to worry about condensate if you're using the Fridge's heat exchanger even with a fan/radiator combo. If you turn it off you'd go below the dew point and get water collection when turned back on (very bad).
You may not get a uniform coolant flow across the component surfaces. The compressor would be cooling the whole fridge as opposed to directing the heat to the radiator. I imagine this would cause the fans and compressor to work overtime.

If you're going to have a liquid cooled sealed system in addition...You'd still have to worry about condensate on the rest of the components which also rust your liquid cooling components assuming any of those components are metal.

You may want to kick this over to the dedicated overclocker forums, as they can give you a much more detailed answer.

 

[USAFRet]

I know if you directly put a pc in a fridge it will eventually heat up the box and just keep getting hotter. But what if you put a radiator on the door and had it pulling air from the fridge and export it out and threw the rad

Ah, this again.

"PC in a mini fridge, AKA What is the fastest way to kill a mini fridge?"

If you're pulling the air out of the fridge, that has to be replaced by other air.
This can only come from the room....much hotter than inside the fridge.
Fridge components are not made for a 100% ON cycle. It will soon die.

 

[dudio]

The biggest problem is the threat of condensation (think moist, think water, think electricity).
The cool air from the fridge actually is not the threat, it's the warm air that has moisture potential.
Cold air hitting warm doesn't cause condensation; warm air on cool causes it.
The problem is you will have warm air in the PC that will contact a cooled surface close to the refrigerated air. I wouldn't trust it to not happen at some point.

 

[InvalidError]

The cool air from the fridge actually is not the threat, it's the warm air that has moisture potential.

Not really an issue here as the small amount of cooling provided by the fridge will be nowhere near sufficient to cool any surfaces within the PC below dew point. No dew point, no condensation.

 

[dudio]

The cool air from the fridge actually is not the threat, it's the warm air that has moisture potential.

Not really an issue here as the small amount of cooling provided by the fridge will be nowhere near sufficient to cool any surfaces within the PC below dew point. No dew point, no condensation.

I guess it depends on proximity, but I understand what you're saying.

 

[InvalidError]

I guess it depends on proximity, but I understand what you're saying.

Proximity doesn't matter. Household fridges have nowhere near enough cooling capacity to cool air to dew point when you leave the door open. If any condensation happens, it'll be on the fridge's evaporator coil. Everything else will be too warm to form condensation even before accounting for the CPU, VRM, chipset, etc. that put out far more heat than what the fridge can cope with, which means that most components/PCBs will remain well above dew point even if you replace the fridge door with the PC case.

 

[GraySenshi]

So basically it wouldn't work because it's not strong enough to keep cool when you leave the door open. Witch having a cutout for a rad blowing all the cool air out of the fridge.

 

[USAFRet]

So basically it wouldn't work because it's not strong enough to keep cool when you leave the door open. Witch having a cutout for a rad blowing all the cool air out of the fridge.

Here's a test:
Find a cheap or free mini fridge.
Plug it in, and leave the door open.
How long does it last?

 

[ccampy]

You'd need a large walk type fridge to have the cooling capacity to compensate for the heat a pc would kick out even then would be working quite hard

 

[rubix_1011]

This is where users searching the forum would find this topic being covered many, many times over the past several years. Not just this forum, but all liquid cooling forums.

 

[InvalidError]

Long story short: get a room AC.

 

[USAFRet]

And lastly:

Even if you were to build this, and it did not kill the fridge or kill the components via condensation, or drive you crazy sitting right next to a fridge running 24/7, or cause your electric bill to skyrocket.....

...to what purpose?
A cooler CPU?
So what?

Unless you are out on the ragged edge of OC and voltage, and the CPU is actually throttling due to overheating...reducing the CPU temp from 60C (current cooling) to 30C (fridge cooling) does exactly squat for performance.

 

[rubix_1011]

^ Agreed.

You'd make better use of the compressor by using it to chill a glycol mixture to use for the cooling loop. Far more efficient thermal transfer.

 

[InvalidError]

You'd make better use of the compressor by using it to chill a glycol mixture to use for the cooling loop. Far more efficient thermal transfer.

Even if you went custom loop and put the radiator inside the fridge, I doubt the fridge compressor would be able to keep up even with a 50W CPU. You'd likely need one radiator outside the fridge for the "first stage" cooling on the hot line to dump most of the heat above room temperature out before the secondary radiator inside the fridge for that extra 3-4C below ambient.

 

[rubix_1011]

Not what I was meaning by this, but that was my fault.

I meant, disassembling the unit and submerging the cooling coils in coolant and sealing that container, essentially acting as a large reservoir. Then, use this coolant to pump through the loop like normal watercooling.

This is done very often.

Edit: this is an example. (I Know this is a window A/C unit, but the concept is similar)

[video="https://www.youtube.com/watch?v=zxdydeDxyf4"][/video]

 

[InvalidError]

Not what I was meaning by this, but that was my fault.

Immersion heat exchange between the evaporator and liquid loop doesn't change the fact that a fridge compressor is not capable of handling that much heat. That's why people mod room ACs for chilled liquid loops, not fridges.

 

[rubix_1011]

Agree/disagree. Yes, the thermal load isn't meant to be handled 100% use, but the intent is for the compressor to cycle and allow the specific heat of the water to act as a thermal buffer for the compressor in both heat and cool cycles. A fridge compressor isn't going to provide as much cooling BTU as an a/c unit, you are absolutely correct.

 

[InvalidError]

Agree/disagree. Yes, the thermal load isn't meant to be handled 100% use, but the intent is for the compressor to cycle and allow the specific heat of the water to act as a thermal buffer for the compressor in both heat and cool cycles.

An under-sized compressor is not going to get to cycle or "buffer" the load when the load exceeds its cooling capacity. It'll operate at 100% duty cycle to cool the liquid mere degrees below ambient at best. Worst case, the heat output exceeds the phase-change cooling capacity and the cold side ends up warmer than ambient, leaving you worse off than a regular radiator.

 

[beoir]

If you're going to fully submerse it, I'd do an evaporative cooling technique. It would be a much easier beast to tackle. Lots of videos out there. I think I may've even run across one on Tom's using a full Ethylene Glycol solution.