review for sanity check

[moon161]

I've had an itch to watercool my CPU in the following fashion:

Water or alcohol phase change, with no compression-Heat boils fluid at the waterblock, and heat is rejected at a condensor at the same temp. Fluid/mist will drip down and refill the block. The whole system operates at a vacuum to depress the boiling point- no compressor is used. Nobody else seems to be doing anything like this, my guess is the unfashionably high operating temp (steady state design point of 120-130 F).

A look at the steam tables says operating pressure is .1 or .2 bar absolute pressure, so pulling the appropriate room temperature vacuum may be a trick. Denatured alcohol will mean less of a vacuum, but a potential steam explosion is bad enough, I'm not sure about adding ETOH to it =8-0X.

[Jack_The_Ripper]

Sounds like a heat-pipe to me. Zalman makes commercial ones for PC's. I remember seeing an article on someone's webpage on making a heat-pipe (dont have the link at the moment, if someone knows what im talking about feel free to post). You could always use propane or r134a from a car A/C refill kit instead of alcohol. If done right you can have a very decent silent machine.

[Enyin]

Interesting concept - i've wondered before if it would be possible to make a phase change system with water.
It would be cool if you could make a diagram of your system so we can see how you picture it.

The problems I had when trying to determine if it's possible or not -
-how do you produce a flow rate in the system? you mention no compression, but you're in a vacuum, so how do you get your 'refrigerant' to cycle
-water has great heat transfer capabilities so this type of system could probably produce a great btu/hr but in the end your operating temp is still 120-130f isn't it? Your cycle won't even start until your cpu or whatever heat sources you have bring your refrigerant up to that temp which is around 50c - a temp that i would feel is uncomfortably warm for my system. My phase change system is in pieces right now but with just my watercooling setup with a fan blowing over the res (in place of evaporator submerged in res) i'm at 87f (31c) under full load
-implications of pure water in the system - corrosion, lubrication of moving parts / pump / compressor, etc

It's an intriguing idea and i'd do it too if i could figure out how to make it work - new and unusual = fun project It seems like this type of concept would only really be suited for cooling something where the 120-130 operating temp is appropriate to what's being cooled

once i saw an old ammonia based air conditioner from the 40's - it had a gas fired burner in it and pumped chilled ethylene glycol into the house, i don't have clue how it worked but whatever principle it used might be useful here! anyone here know anything about that type of system?

[redleader]

-how do you produce a flow rate in the system? you mention no compression, but you're in a vacuum, so how do you get your 'refrigerant' to cycle

This is called convection. Its natural and in theory it will pump heat faster then any Eheim ever could. In practice its more complicated.

water has great heat transfer capabilities so this type of system could probably produce a great btu/hr but in the end your operating temp is still 120-130f isn't it? Your cycle won't even start until your cpu or whatever heat sources you have bring your refrigerant up to that temp which is around 50c - a temp that i would feel is uncomfortably warm for my system. My phase change system is in pieces right now but with just my watercooling setup with a fan blowing over the res (in place of evaporator submerged in res) i'm at 87f (31c) under full load

This is what your intuition says, however thats not how it works. Think when you get out of the shower, you don't have to be in a hot room for the water to evaporate and make you cold. In fact any time you are above the freezing point this effect will occur.

The operating range is actually quite broad, wider then with watercooling.

-implications of pure water in the system - corrosion, lubrication of moving parts / pump / compressor, etc

What moving parts? How would their be corrosion?

Heres a link to some on going heatpipe experiments:

http://arstechnica.infopop.net/OpenT...5&m=7250984465

once i saw an old ammonia based air conditioner from the 40's - it had a gas fired burner in it and pumped chilled ethylene glycol into the house, i don't have clue how it worked but whatever principle it used might be useful here! anyone here know anything about that type of system?

I finished my thermo final 4 hours ago, so what the heck. A heatpump can use any source of work to pump heat. It can be electrical, chemical, kinetic, potiential, even thermal (though you would need very special cirrcumstances before this will work).

My guess is that system was burning fuel to generate kinetic energy that was driving a compressor. But it could have been more exotic.

[moon161]

No moving parts, except maybe an external fan. Coolant boils, convects to the wall of the vessel, and condenses, giving off heat.

If you google 'heat pipe' , you can find a good how to
and

a mass produced one too .

My plan is to use a flow restrictor in one side of a coolant loop system, instead of a plain vessel. This will force liquid up the pip on the opposite side, developing a head which will force flow without a pump. Since forced convection gives a higher temperature gradient at the wall, this should make for better heat transfer out of the system.

The operating temperature range of the system is limited by the strength of the pressure vessel- too hot = too much pressure = boom. It will not boil dry- as the pressure increases, the boiling temp increases as does the proportion of liquid to vapor.