post #11 of 11
Thread Starter 
Yeah, sorry Skeeter, I seem to have partly misread your intention, I thought you intended to use water as the working bad...was early AM when I responded. Think the rest of what I said stands good though.

If you are going to use a suitable working fluid to submerge and cool the radiator by evaporation, then why bother with a radiator? Simply choose a working fluid such as Novec 7000 and submerge the whole mobo. That way you get direct cooling of the cpu/gpu (and other mobo components) which would be much more efficient than cooling the rad as you've suggested as it cuts out all the steps each of which has its own temp delta. So to outline the mobo submersion system heat transfer would simply be heat spreader/working fluid while the cooled water loop would be heatspreader/tim/waterblock/water/radiator/working fluid

Although I can see one advantage of submerging the would allow you to use a cheaper working fluid, one that didn't have to be compatible with mobo submersion.....but then again why even bother with submerging the rad?....just direct tec cool the water loop.

I'm not sure if you have envisaged this working as a heat pipe or not? If you wished it to work as a heat pipe then the box would have to be sealed and evacuated of non volatile gases so that the working fluid is in thermal equilibrium with its vapor phase. In that case the box would have to withstand pressure differences possibly both positive and negative as the workload and cooling altered the internal pressure.

This way the tec's would alter the temp of the working fluid due to the liquid/vapor/pressure/temp equilibrium and so the working fluid temp would go below the atmospheric boiling point of the working fluid.

If you did not have this arrangement in mind and the box was held at atmospheric pressure ( very much like the new build I'm working on) then the liquid/vapor equilibrium would be broken and below the boiling point of the working fluid then there would be little energy exchange between the tec's and the working fluid. So your minimum working fluid temp would be around the atmospheric boiling point of your working fluid. (unless you were to pump the working fluid over the tec's. )

When you first described this I presumed (perhaps wrongly) that you would be going for the latter system where pressures would be atmospheric and using a working fluid with a boiling point just above normal ambient temps......this would be the easiest system to make by far.....but you would not get working fluid temps much below the working fluid atmospheric boiling point (unless you were to actively pump the working fluid over the tecs ) and therefore your water loop temps would be around normal ambient temperatures.

This is what led to my comment that there would be little value in doing this just to achieve around ambient temps which can adequately be obtained with more rad/air flow. Obviously if you were going with the former (more difficult) build then you could achieve sub-ambient temps and there would be more value in it.

I think with the lack of full detail I may have jumped a couple of steps ahead of where you were at with your suggestion (perhaps in the wrong direction) as I've thought out quite a bit of stuff relating to this as its similar in someways to my next build....but tbh the only way this would be of advantage would be for submersing the mobo......otherwise just direct tec cool the water loop.
Edited by technogiant - 4/17/13 at 5:56am