When I built the HTPC back in November, I bought my first heat pipe-based CPU cooler and was very impressed with its efficiency: with the CPU at idle and the heatsink fan going full bore the CPU temperature is actually below ambient. I didn't know much about heat pipes so I read up on them and was shocked to find that, even though they are nothing more than an evacuated tube with a bit of fluid inside, they can conduct heat considerably more efficiently than an equivalent cross-section of solid copper.
Anyway, Bob got bored this weekend and built a giant heatpipe just to see how well it would work. The construction is extremely simple; it's just a 10' length of ½" copper tube with a cap soldered on to one end and a ball valve soldered onto the other. Some foam insulation was placed around the piple to make it more comfortable to hold when the pipe gets extremely hot or cold, but doesn't have much effect on the functionality.
Bob holding snow onto the upper end of the heat pipe. Ball valve in the foreground.
High-tech heat pipes have a wicking material on the inner surface of the tube to expedite the transport of the liquid-phase coolant back to the hot end of the pipe. A budget heat pipe like Bob's doesn't require a wicking layer, however; it just relies on the bouyancy differential between the liquid and gas phase coolant (which in this case is water) to get the liquid back to the hot end. For this reason, the BobPipe will only work if the hot end is lower than the cold end, though our experimentation demonstrated that an angle of about 2° off horizontal was plenty.
The key is to get the pipe reasonably evacuated, which it turns out can be done without anything fancy. We opened the ball valve and poured in a small quantity of water (about 3 tablespoons worth). Then, the ball valve was closed and the pipe tilted so that all of the water flowed to the non-valve end of the pipe. This end was heated with a blowtorch until the water was boiling. When the water was superheated and the pressure inside the heat pipe was at several atmospheres, the ball valve was cracked opened a bit a some of the steam let off (but not all). The remaining steam, when it re-condenses, will result in an internal pressure of well under 1 atm and increased efficiency of the heat pipe..
After heating the lower end of the pipe for several minutes with a blowtorch, to the point that it would severely burn me if I touched it, Bob held a handfull of snow onto the upper end for about 10 seconds (the snow was melting at an astounding rate) and already I was able to comfortably touch the spot where I had just had the blowtorch. Soon the low end of the pipe was cold to the touch. It was really pretty cool.
Total cost: ~$15
