 |
April 09, 2013, 01:57:07 PM |
|
Phase-change cooling leveraging low boiling point, dielectric fluids has something that has always piqued my interest.
Of course, the design challenges of a) housing b) fluid conservation, reclamation, and compression/condensation and c) minimizing design complexity and fluid usage for cost reduction are not trivial.
I'll be blunt. BFL's prototype power usage has me concerned. To a point where I might make a cloud chamber leveraging phase change and a heat exchanger.
Cooling studies based in CPU have pointed out that a FCBGA package, used by the BFL ASIC, while it radiates heat away from the board better, still the chip itself acts as a heat sink for the remainder of board components. Whatever is going on with BFL's board components, if BFL is making true statements, is causing some serious heat and that heat will follow the path of least resistance. If the chips are the path of least resistance - no bueno.
So, to entirely abandon hopes that BFL's design would properly manage thermal loads of all components on their board I have begun to design a DIY cloud chamber.
To restate, the goals are:
1. DIY
2. Simple - no chip blocks - semi-immersion method.
3. Safe - no caustic or flammable fluids
4. Modest Novec, dielectric fluid, usage (it's expensive. $700 / Liter)
5. Energy efficient. Gas-liquid to air heat exchanger.
The most non trivial part to this is to ensure that the chamber is utterly air tight. The proposition of using Novec means that special precautions for maintenance access, and design of the maintenance hatch, could be tricky.
Anyone interested in keeping an eye on this?
|
