I offer a serious concern that the wide temperature swings caused by freon cooling a processor IC will cause failures because of the expansion/contraction cycles as the system is used. A commercially made heat pipe system would be much better for this application.

 

I offer a serious concern that the wide temperature swings caused by freon cooling a processor IC will cause failures because of the expansion/contraction cycles as the system is used. A commercially made heat pipe system would be much better for this application.

Wide temperature of the processor or the refrigerant? And what are some reasons this will cause the faliure? I don't quite understand your concern

 

Wide temperature of the processor or the refrigerant? And what are some reasons this will cause the faliure? I don't quite understand your concern

Extreme temperature changes of the processor being cooled by the refrigerant In addition, if it were a good idea it would already be in practice. In addition, the plumbing to connect the high pressure refrigerant lines will be a royal aggravation inside any computer housing. AND the size of the compressor and condensor will make the system rather large.

 

3.5kw is massively overkill for a processor. The outdoor unit will produce a lot of liquid and without sufficient heat gain to boil it off (even with an accumulator fitted) the liquid will flood back to the compressor.

Use less refrigerant? If you undercharge the system the compressor will overheat as it relies on a stream of cool vapour entering the suction port.

What might work is to fit the indoor unit somewhere convenient, pipe and wire it up but run the suction pipe through the processor heat exchanger on the way back. You could then control the duty by slowing or even switching the fans off. The indoor thermistor could be resited onto the processor or somewhere near it that's warm.

The outdoor unit needs the indoor unit to work anyway nowadays, each unit is packed with electronics.

Personally I would look at using a small fridge. Remove the evaporator and use 1/4 pipe to extend the liquid line and capillary into your processor heat exchanger. And again 1/4 pipe back to the compressor (insulated) fit a small accumulator and it will need a bit more refrigerant adding. If the compressor gets too hot fit an axial fan to blow on it.

Good luck, I think you're going to need it.

 

3.5kw is massively overkill for a processor. The outdoor unit will produce a lot of liquid and without sufficient heat gain to boil it off (even with an accumulator fitted) the liquid will flood back to the compressor.

Use less refrigerant? If you undercharge the system the compressor will overheat as it relies on a stream of cool vapour entering the suction port.

What might work is to fit the indoor unit somewhere convenient, pipe and wire it up but run the suction pipe through the processor heat exchanger on the way back. You could then control the duty by slowing or even switching the fans off. The indoor thermistor could be resited onto the processor or somewhere near it that's warm.

The outdoor unit needs the indoor unit to work anyway nowadays, each unit is packed with electronics.

Personally I would look at using a small fridge. Remove the evaporator and use 1/4 pipe to extend the liquid line and capillary into your processor heat exchanger. And again 1/4 pipe back to the compressor (insulated) fit a small accumulator and it will need a bit more refrigerant adding. If the compressor gets too hot fit an axial fan to blow on it.

Good luck, I think you're going to need it.

That's why Drc_567 offered me a good idea to this problem. To install a reservoir to catch the liquid and only let the gas pass into compressor. I'm going to just use enough refrigerant pressure to have the hot high pressure side hotter than the outside ambient air. That way the cool side will stay cool, although it will make the system less efficient.

As for the outdoor unit needing the indoor unit because of propietary technology tied to each other, the unit I'm planning on buying has all the connections coming out of the outdoor compressor unit. I will just not use the conenction for the indoor unit and then wire the compressor to run with a relay on an Arduino and have it controlled by a thermal probe in the glycol reservoir.

I heard that refrigerator compressors aren't fit to handle continuous heat input I think. If you can give me some good ideas on using refrigerator parts I'll be glad to consider it.

 

Extreme temperature changes of the processor being cooled by the refrigerant In addition, if it were a good idea it would already be in practice. In addition, the plumbing to connect the high pressure refrigerant lines will be a royal aggravation inside any computer housing. AND the size of the compressor and condensor will make the system rather large.

The compressor is going to run continuously. The temperature changes of the processor shouldn't effect the system because of how relatively small the amount of heat is being removed compared to the cooling potential of the AC system. As for the plumbing and tubing work, I don't plan on putting any of the AC parts inside the case and just gonna run the 2 pipes necessary to have the cooling done to the inside of the PC tower.

 

While I have to admire your tenacity and wish you the best of luck I played this same game about 12 to 15 years ago. I was then active in the old Maximum PC Comm Port and their forums. Among may projects were dragging home an old water fountain ripped from the plant during renovations. Built a reservoir and added a circ pump so what I had was water cooling extreme, I also had condensation extreme and not just from my cooling lines but the entire CPU cooling block and heat sink assembly. Doesn't take much when the block hits dew point of ambient air, even in low humidity environments.

I then decided to do some test, what remains of images can be found here. I wanted to see just how much heat the system generated. The system was dual Xeon processors running at 100% all cores in both processors. The HSF assemblies were an optional configuration. The bottom of the linked page reflects my temperatures. The point is, no additional fancy cooling was required.

My neighbor bought one of those large in room air conditioners which you vent to the outside. You may want to look into something like that and just feed cold and dry air into you case. Use any quality HSF which uses heat pipes to remove the heat from the CPU. I can tell you that while it sounds good trying to pump refrigerant through a CPU cooling block is a futile mess but again I encourage you to try as you will come away with some lessons learned.

You may want to run a few test similar to the link I provided so you at least know where you are and how much heat you actually want to remove?

Ron

 

As I said, an accumulator then, but it probably won't work. The duty will be so low that the back pressure will be in a vacuum, pressure and temperature are relative. look up pressure temperature charts for R410a, the suction line will have a very sharp frost on it (-70°c) but the outlet at the top of the accumulator will be warm and the compressor will overheat... Actually it won't overheat as such, the thermistor on the compressor discharge will tell the microprocessor it's too hot,
shut the compressor down and you'll get an error code on the remote display.

Anyway how much power does your processor use? For your 3.5kw cooling (1.1kw electrical) AC unit to be matched to cancel out an electrical heating load your heat gain would need to be in the region of 3.5kw (electrical )

Let us know how you get on

 

I see now. But I still want to try this. I know this can be very risky but I'm going to be doing this on a old $40 desktop PC and do some extensive experimenting on that before I actually impliment it.

For the condensation problem I have already provided my solution earlier in this forum of using active heating on the pipes. Anyway if all goes wrong I guess the worst I can have is a burnt out compressor and a cheap broken computer.

Also I have changed my plan earlier in this forum when I said I'm not going to have the pressurized gas run over the cpu. Instead I now plan on cooling using a evaporator submerged into a box filled with glycol and running the glycol through the cpu block

 

Also in the case that 3.5 kw is overkill for the processor, can I reduce the cooling by running the compressor at a lower voltage or something of that sort?

 

I see now. But I still want to try this. I know this can be very risky but I'm going to be doing this on a old $40 desktop PC and do some extensive experimenting on that before I actually impliment it.

For the condensation problem I have already provided my solution earlier in this forum of using active heating on the pipes. Anyway if all goes wrong I guess the worst I can have is a burnt out compressor and a cheap broken computer.

Also I have changed my plan earlier in this forum when I said I'm not going to have the pressurized gas run over the cpu. Instead I now plan on cooling using a evaporator submerged into a box filled with glycol and running the glycol through the cpu block

Like I mentioned, by all means, have at it. There is always a better mouse trap and you come away from projects like this with lessons learned. Take pictures and post back how things go. Following the water in my chiller experiment I went with glycerol. Machined my CPU block at work. My lines weren't as bad of a problem as the CPU heat exchanger block for condensate. Anyway, again wish you the best of luck.

Ron

 

Than

Like I mentioned, by all means, have at it. There is always a better mouse trap and you come away from projects like this with lessons learned. Take pictures and post back how things go. Following the water in my chiller experiment I went with glycerol. Machined my CPU block at work. My lines weren't as bad of a problem as the CPU heat exchanger block for condensate. Anyway, again wish you the best of luck.

Ron

Thank you and I will try to keep everyone updated on how it goes. Also where can I get glycerol?

 

Than

Thank you and I will try to keep everyone updated on how it goes. Also where can I get glycerol?

I just used common anti freeze (ethylene glycol).

Initially my "computer room" was a home office space which was home to my wife's computer, my computer and my workstation computer which I linked to previously. I finally put in a window AC unit for the heat of summers which kept the room not only cold but reduced the humidity. Eventually we did major home renovation and went with whole house air conditioning. I also had several UPS units and added a whole house home generator system. I gave the window unit to my neighbor and they still use it today.

Following all of my CPU cooling experiments, based on the link I provided earlier, I figured out that while my things were fun to experiment with the added CPU cooling wasn't really necessary. My only really high end beast was the workstation PC. But it was all fun and interesting.

Common anti-freeze aside there are other oils and coolants which work real well but can also get pricey. Many are used in ice baths for temperature calibration.

Ron

 

A simple and easy way to approximate the heat from the computer electrical stuff is to use the power input figures, because in most electrical things, ultimately most of the power winds up as heat. For cooling the computer, not just the CPU, but ALL of the components, you could also use one of the the packaged dehumidifiers and simply tap off some of that cold dry air from between the cold evaporator and the hot condenser. Then just use a flexible air tube to connect it to the computer case. Even a cheap dehumidifier would work, and they are made for constant duty. It would eliminate all of the plumbing work and all of the condensation issues inside the computer housing, and it would certainly be cheaper than an AC unit.

 

A simple and easy way to approximate the heat from the computer electrical stuff is to use the power input figures, because in most electrical things, ultimately most of the power winds up as heat. For cooling the computer, not just the CPU, but ALL of the components, you could also use one of the the packaged dehumidifiers and simply tap off some of that cold dry air from between the cold evaporator and the hot condenser. Then just use a flexible air tube to connect it to the computer case. Even a cheap dehumidifier would work, and they are made for constant duty. It would eliminate all of the plumbing work and all of the condensation issues inside the computer housing, and it would certainly be cheaper than an AC unit.

That gets into another interesting world. Somewhere in all the junk I have piled up around here is a few of those Kill-A-Watt meters. I would give the workstation things to do in order to run up the CPU usage and monitor the overall AC power consumed while looking at the inside temperatures. It was interesting until it wore off.

I did, at one point place a simple 12 VDC computer fan in some flexible duct and channeled the dry cold air (about 40 degrees F.) into the computers which worked out pretty well. Anyway, what was interesting and made perfect sense was the more work I gave the processors to do the higher the power consumption and the higher the generated subsequent heat. While all pretty logical it was entertaining to chart things and record the data.

Ron

 

Also in the case that 3.5 kw is overkill for the processor, can I reduce the cooling by running the compressor at a lower voltage or something of that sort?

No you can't reduce the voltage to the compressor to reduce the speed. AC motors are asynchronous to the mains frequency... It would slow down but then would pull more current and overheat.

Anyway you shouldn't have to if you're now going to cool a bath of water. Put a thermostat in it set to 40°f and the compressor will switch off.

 

No you can't reduce the voltage to the compressor to reduce the speed. AC motors are asynchronous to the mains frequency... It would slow down but then would pull more current and overheat.

Anyway you shouldn't have to if you're now going to cool a bath of water. Put a thermostat in it set to 40°f and the compressor will switch off.

If I want to go lower than 40f can I just lower the pressure inside the system and put a reservoir to collect the unevaporated refrigerant and instead of using the built in thermostat to controlled the duty of the compressor, instead use an Arduino and a separate thermal probe to stop the Compressor at a different temp?

 

If I want to go lower than 40f can I just lower the pressure inside the system and put a reservoir to collect the unevaporated refrigerant and instead of using the built in thermostat to controlled the duty of the compressor, instead use an Arduino and a separate thermal probe to stop the Compressor at a different temp?

You can't lower the pressure in the system easily on an AC system, they aren't like refrigeration systems. The metering device is usually capillary and fixed. All the components are matched and calibrated for an evaporating temperature of about 5c (that's why the evaps don't ice up) The pressure in the system will fall naturally as the temperature falls.
The best way to get to a lower temperature is to simply run the compressor longer and see what it will pull down to. Use antifreeze of course. It's uncharted territory, it may work OK. Brewery beer coolers work in a similar way but with coils of 3/8" copper as the evaporator coiled round the inside edge of the water bath.
You won't need to fit an accumulator if you stick with the original evaporator.

 

Actually the brewery drinks coolers don't use antifreeze. They build up an 'ice bank' round the evaporator. The thermostat phail is only a few cm away from the evaporator and is set if I remember correctly, at - 2°c so the middle of the tank is water at very close to freezing. Some coolers have pumps on to circulate the cold water up to the bar top and back to keep the drink cold in the pipework.. Also some brand displays on the bar are connected to make them condensate and look attractive.

We used to call the small under counter chillers "Flash coolers" they would be ideal for your application.