Water to Water heat exchanger?

I'm building a water chiller for my computer, and I need some suggestion on how to build a water to water heat exchanger. I have two rads in my system as of now, but the ambient temperatures right around it are way too high (it's in "heat bubble" under my desk, in a room that already stays 5c above the rest of the place) resulting in it going over 40C on my heat spreader. Unfortunately, I cant do any thing permanently attached to my system to remedy this because my system MUST remain portable and able to run on just the 80 and 120mm rad's that are in it.

So, here's what I've come up with... A water chiller system based off the
$90 Easy Phase Change Cooling! thread that uses it's own coolant loop to cool the main loop inside my computer by way of a water-water heat exchanger. Think the way a two look nuclear reactor works, and you'll get the idea.

So, now the part I need help with... What's the best way to go about building a heat exchanger that will not take up much room (2" tall and wide, by 6" long MAX, preferably smaller), not restrict the main loop's flow much, and I can build out of modified pluming supplies (or things I can find locally) or even order? I don’t care about the chilled loops restriction, as this is all that loop is cooling. I'm also going to ether add valves or quick connects to the outside of the computer so I can disconnect this system easily for transport and use without the chiller.

Here are some ideas I've come up with so far....

1) Main loop flows through a copper pipe inside of a pipe (plastic) with the chilled water flowing through.
a) Main loop water flows through a single 3/8" copper tube inside a PVC chilled water tube.
b) Main loop water flows through two 1/4" copper tubes inside of a PVC chilled water tube.

2) A set of bent copper tubes (two 1/4", or four 1/8" in an inter-woven zig zag) goes inside of a PCV pipe.
a) Chilled water goes through tubes, main water flows outside.
b) Chilled water goes outside, main water flows through tubes

3) Main loop water flows through a pipe, with a copper coil inside containing the chilled water. This was my original idea, but unfortunately I could not find any copper tubing locally that will bend tightly enough to fit inside 1.25" pipe without kinking and cutting down on flow majorly.

4) (complex) A series of copper tubes run between two disk inside of a PVC pipe. These disk seal off the center of the pipe from the ends, but the ends are connected by the tubes. Then the main loop water can be run through the ends and into the tubing, while chilled water is pumped through the center.

Any other ideas, or comments on the ideas I've come up with so far? Has any one actually built one of these before maybe?

1a might be your easiest to construct solution, and is commonly referred to as a counterflow heat exchanger.

These are pretty common in engineering heat transfer texts and such, and you should be able to calculate the heat transfer etc.. if you use this method.

Just the first few links off google for some example/background:
http://www.taftan.com/thermodynamics/COUNTER.HTM
http://www.engineersedge.com/heat_exchanger/counter_flow.htm
http://www.me.wustl.edu/ME/labs/thermal/me372b5.htm

By aligning the direction of the flowing fluid, it becomes a parallel flow exchanger
http://www.taftan.com/thermodynamics/PARALLEL.HTM

and some of your other ideas are crossflow exchangers
http://www.taftan.com/thermodynamics/CROSS.HTM

Just some stuff you might watn to read up on. Anyway, good luck on your project. Post progress and results of course.

This might be too bulky for you, but if you could find a very small car radiator with transmission oil cooler plates in one of the tanks, you could mount that on the side of your case as in Voigts "Black and Blue" system and use the tranny cooler to tranfer heat to your chiller.

If I did 1a, would I need to do any thing to increse surface area of the main loop's line inside the heat exchanger? Or would a plane old copper tube 6" long do the trick?

There's really no way to tell for sure unless you know a lot of information about both loops (like flow rates, temps, etc...)


Anyway I forgot it was only 6" long, so I would go with 2 or 3 smaller copper tubes inside the one PVC tube, should definately help to double or triple the surface area.

Then the only way to know for sure is to try it and see. And be cautious and monitor temps of everything along the way.

Hum... The shell type (the one with the internal plugs) is looking better to me now. I can probably do it with epoxy and milk jug caps inside of 1.5" PVC. Now I need a low cost ($40 or so) pump....