"Ok so, how cold is the "Shot of coolant" going to be? Keep in mind the processor would continue to generate heat so just dumping it in ice water for a few seconds is only going to bring it right back to it's original hot spot a few seconds later."
Question is what is the specific heat capacity and the latent heat of evaporation of the fluid. It's capable of "absorbing" significant calories in evaporating. How quickly the core returns to dangerously high levels is a question of thermal inertia and how much heat (energy) you are dumping into the system. http://en.wikipedia.org/wiki/Heat_capacity
Scroll down a bit to the discussion on Heat Capacities. With this sort of intermittent super cooling scheme we'd like to increase the thermal inertia (mass) of the system - ostensibly by placing something having properly engineered thermophysical constansts in intimate contact with the surface (top and bottom?) of the CPU. This will act as a buffer and help moderate wild temperature swings (either way). I think it's all do-able I'm not a big gamer tho - my system is doing A.O.K. thermally. I applied better compound and give it the every 5 months cleanout .. that seems to do the trick .. perhaps I will go the distance and come up with some kind of evaporative cooling rig. It's just not a burning requirement (pardon the pun) - at this juncture.
I gave a better reply earlier but it didn't get posted for some reason .. Apply dielectric grease http://www.google.ca/search?hl=en&so...=f&oq=&aqi=g10
everywhere - to seal out atmospheric condensation (addressing the condensation objection) and as for how fast the coolant gets consumed .. that's largely a function of how hard you throttle your CPU.
I envision this system giving short coolant blasts (too much is not necessarily a good thing as it cools via evaporation) .. but lets suggest the system delivers 2-second long bursts of coolant .. if you'd clicked on those links you'd be the expert on how cold it can cool .. which is actually a misnomer as the true temperature depends on the conductivity of the substrate .. more germaine to discuss how many calories are required to evaporate a given quantity of coolant. The evaporation is driven by partial pressure changes and the cooling effect is ultimately a secondary effect .. stemming from the latent heat of vaporization http://en.wikipedia.org/wiki/Latent_...f_vaporization
So in discussing things related to cooling efficiency its more appropriate (germaine) to discuss heat flux than temperature.
Suffice to suggest I'm of the mind that there is ample heat removal capacity associated with the vaporization of this chiller fluid and that given a suitably sized/designed intermediately-located thermal mass having the appropriate thermal characteristics .. not too conductive and not too insulative .. we could have a system that is stable. The "duty cycle" would vary according to how much processing power you require. i.e. the interval between the 2 second coolant blasts would vary based on sensor readings. It's my impression that normally the core temperature fluctuates pretty drastically and that ideally we would have a heat removal scheme that would respond without too much lag .. i.e. maintain a steady operating temperature - regardless of CPU load. OK enuff.
Build it - and they will come.Edited by DoubleDensity - 10/8/09 at 2:52pm