Watercooler-LITERALLY: idea.
 

O.k. So I was thinking and its common sense, that the coolest the water will get in a watercooled system is around the ambient temp the system is under. Right? So, in order to get the water cooled system colder you would have to chill the water that is in the system. Right again?

O.k., so why not use a watercooler for your water cooled setup? You know, the actual water cooler you drink out of, that sits at the corner of the office and makes for a conversation get-away. It will actually be a "Phase change/Watercooled Rig".

You can run it all day long, and its very quiet and very easy to find.

If I was gonna do it, I would gut it out and build a custom case for the guts so it doesn't actually look like a watercooler.

For those who don't really get it, the watercooler is a phase-change system that is cooling the water. The evaporater (or coil) is wrapped around the bowl that holds the water. All you really have to do is get the cooled water to your computer.

So, what'dya think?

.....or, someone already did it.

http://forums.overclockers.com.au/sh...hreadid=130201

and

http://forums.overclockers.com.au/sh...hreadid=137536

Someone already did it. Can't remember what site is was on. Probelm with them is they are generally not rated to handle the heat load modern CPU put out. You will end up with a system running about the same as if the rad was still there in place of the chiller.

If I recall someone was selling these on ebay a while back aswell...

Still an idea if you can find one that will have over 100watts of cooling power.

EDIT: bah, Cathar beat me to it.... :shrug: :D

Yeah, now that I think about it, It'll be hard for the coil to keep up with the warm water going into the bowl. It probably won't be able to chill it quick enough. Unless, you keep the water in the bowl longer length of time before it exits to the cpu. Like, having the water run through a copper coil wound up, inside of the bowl that has the cold water. ( I can picture it in my head, better than I can explain it.)

Spending more time in the bowl won't help if the coil is underpowered for the task.

Watercoolers have 1/12 HP compressors, which is generally enough for a modest chiller. You just have to upgrade the condenser and maybe tweak the cap tube. Lots of people have done it.

Personally I'd use a large compressor.

Problem with water coolers (and fridges and similar devices) is often their compressor is not 100% duty cycle - it's something you should definitely check before using one for a computer.

I have seen some aquarium water chillers for people that have big aquariums and exotic fish that only live in cold water (10C or so). I don't have any links immediately handy, but you usually find them in the same kinds of on-line stores that most of us browse for pumps.

They are powerful enough for a CPU cooler app, but they are priced around six hundred bucks and up. Kinda spendy to cool water. They are designed to be very, very quiet while they are running though.

I work as an Automotive Technician and specialize in automotive HVAC. Modifying the condenser or building a new one, or changing the capileri tube isn't hard work with the right knowledge and ability. I work part-time at a Rad shop and all we do is build, repair, and diagnose Heating and AirConditiong systems on cars extensively more than the average garage. Of course it differs with commercial and residential systems. But, as a Technician you learn one most important thing after a number of years in a trade and that is, anything can be fixed or repaired.

I was also thinking of using the low side of the watercooler and T-ing it off so one half goes to the coil for the bowl and the other half going through an actual automotive evaporator and that way you can chill the water and chill the air going through the computer case. Sort of like a car with front and rear A/C. Without the recirc option of course.

Then, I would for sure have to modify a condenser and change the compressor.

what'd you guys think of that idea?

......or has that one been done too.

Yes its been done. Making a case air tight and then chilling the air on the inside is one way to reduce condensation on multiphase cascade systems. My opinion is that its rather pointless if you're operateing well below -100C. Above that insulation is a more elegant solution (IMO of course).

Other then that your idea is pretty solid. AIM for a high heatload (couple hundred watts) when you size the cap tube so that you don't have to wait forever for the coolant to pull down, even if it makes the temps a bit higher.

What refrigerant are you thinking of? R-22 or R-404 would be great if you can get them.

Saw someone's rig on the net where he put the PC in an airtight chamber and used chilled nitrogen to cool it - just had a big heatsink on the CPU. Got it pretty cool without having to worry about condensation.

I'll probably use R-12. My boss has a couple bottles stashed away, for the classic cars that want to keep their A/C system in original order.

It'll take some work, and time (which, I can't seem to get enough of).

I use an aquarium water chiller and have found it works well for my needs. It is only 1500 BTU's but it kepps my mobile down to around 15 C overclocked to 222x12. I decided to use this over a TEC as I started thinking of my power bill. Got my first bill in the mail last week since running my chiller (hardwired to stay on all the time). I was pleased as my energy consumption was no more then it was at the same time last year. Although 15C is not the 0C I once dreamed of, I had to come to terms with other factors, such as noise and cash! :D At the moment I am still planing on making a chiller that will go colder, but I think for long term use I will stick with my chiller as it is quiet and doesn't break the bank. I have an AC unit and a fridge that I am currently working on. Sadly, we have had such a nice summer here in the PNW, that I have been using the AC unit to keep the Appt. cool. The fridge on the other hand is being gutted as we speek. :D

(!)

No, time in a given area in a closed loop is irrelevant, as it is constant no matter what the flowrate is.

Think about it.