rossd
12-10-01, 01:42 AM
It occurred to me that I could use my air compressor to cool my CPU.
I'd run the output air from the compressor through a radiator to get it down to room temp, then through some insulation to the "cold zone".
In the cold zone, I would have an expansion valve. The air coming out of the valve would be even colder, and would go through a block on the CPU to bring the temp back up to the cold zone temp. This would absorb the heat from the CPU.
Lastly, the air would go back out through the insulation to vent into the room.
The inflow and outflow tubes would need to be run very close to form a counter-flow heat exchanger. That way, the compressed air arrives at the CPU cold and the expelled air exits to the room at room temperature.
From quick calculations, I can see two hitches:
1. I would need an adiabatic expander for the expansion valve. These appear to be very complex. If I didn't use an adiabatic expander, the work done by the expanding gas would (I guess) be converted to heat, which would more than equal the cooling capacity of the system.
2. My air compressor seems to be too small by about a factor of four.
Has anyone ever tried this method or heard of anyone trying it ? It would seem to be a practical way of cooling about 200W at -50C, if you have a big compressor (1.6L/s at 100psi, if my calculations are correct).
Thanks,
rossd.
I'd run the output air from the compressor through a radiator to get it down to room temp, then through some insulation to the "cold zone".
In the cold zone, I would have an expansion valve. The air coming out of the valve would be even colder, and would go through a block on the CPU to bring the temp back up to the cold zone temp. This would absorb the heat from the CPU.
Lastly, the air would go back out through the insulation to vent into the room.
The inflow and outflow tubes would need to be run very close to form a counter-flow heat exchanger. That way, the compressed air arrives at the CPU cold and the expelled air exits to the room at room temperature.
From quick calculations, I can see two hitches:
1. I would need an adiabatic expander for the expansion valve. These appear to be very complex. If I didn't use an adiabatic expander, the work done by the expanding gas would (I guess) be converted to heat, which would more than equal the cooling capacity of the system.
2. My air compressor seems to be too small by about a factor of four.
Has anyone ever tried this method or heard of anyone trying it ? It would seem to be a practical way of cooling about 200W at -50C, if you have a big compressor (1.6L/s at 100psi, if my calculations are correct).
Thanks,
rossd.