The only thing I have a problem with water chiller is the pump, and the fluid. And that I mean how the pump handle the cold in higher viscosity fluid... If going realy low that is, also more problems with insulation.
Thats why direct die evaporator is the way I'm gonna go.
I like that idea bowman about the small chiller, I was thinking something similar just in a shape of one big block like a heat exchanger, so unused fluid from cpu to this evap which will have chanells milles on both side one for evap. one for water.
What is the moste efficient way to controll the freon fluid flow through the system, if somehow there won't be enough fluid from cpu evap to cool water for gpu and nb to let say 10°C, but to complicate even more
to let me boost the cooling for short amount of time.
Because playing with optimal lenght of capilary tube is not realy practical for me because I don't have the equiment for refilling.
So I would like to make unit that have some cooling left and not running at it's limit. So if I would use thermostatic expansion valve how much performance would I actualy lose, my goal is to keep cpu in the range of 120-150W bellow zero at full load.