- Sep 30, 2006
- 4,267
- 421
- 126
I am building 2 water-cooled quads and I plan to have an external water-chiller at some point in the future. I am trying to design a method where I can use quick-disconnects to attach the chiller to both computers. I want to be able to attach without turning off the computers, so I am thinking along the lines of a heat exchanger inside each computer. I have been looking at tube-in-tube as perhaps the best way to accomplish this. However, I have an idea for a simpler way to do this.
External configuration:
To the west is the water-chiller with a pump switch and an in and out line for each computer.
To the east is a computer with an in and out line, each with a quick-disconnect.
Computer configuration:
Inside each computer is a small manifold which we can visualize as a large X having top and bottom ports on the west and on the east.
Top west port receives cold water from the chiller through the quick-disconnect.
Top east port receives hot water from the computer circuit.
Bottom west port sends warm water to the chiller through the quick-disconnect.
Bottom east port sends warm water to the computer circuit.
Tube-in-Tube manifold:
The west top and bottom ports connect to the same pipe.
The east top and bottom ports connect to the chamber surrounding the chiller pipe.
The Rudy:
This manifold will not have the inner pipe. The cooling solutions from both sources will intermix.
This requires the chiller circuit and the computer circuit to have the same cooling solutions.
Of course, the chiller will have a tube-in-tube exchanger and pump for each external connection.
If the chiller output is always above the dew point, I will not have any condensation problems inside the computers.
The chiller could have a temperature controller to ensure this. (I haven't thought that out yet.)
What do you guys think?
EDIT:
Some janitorial closets use this method to add chemicals to the water in a precise ratio controlled by nozzles. I want to do the same thing, except I want to control the ratio by the solution temperature out of the bottom west port.
External configuration:
To the west is the water-chiller with a pump switch and an in and out line for each computer.
To the east is a computer with an in and out line, each with a quick-disconnect.
Computer configuration:
Inside each computer is a small manifold which we can visualize as a large X having top and bottom ports on the west and on the east.
Top west port receives cold water from the chiller through the quick-disconnect.
Top east port receives hot water from the computer circuit.
Bottom west port sends warm water to the chiller through the quick-disconnect.
Bottom east port sends warm water to the computer circuit.
Tube-in-Tube manifold:
The west top and bottom ports connect to the same pipe.
The east top and bottom ports connect to the chamber surrounding the chiller pipe.
The Rudy:
This manifold will not have the inner pipe. The cooling solutions from both sources will intermix.
This requires the chiller circuit and the computer circuit to have the same cooling solutions.
Of course, the chiller will have a tube-in-tube exchanger and pump for each external connection.
If the chiller output is always above the dew point, I will not have any condensation problems inside the computers.
The chiller could have a temperature controller to ensure this. (I haven't thought that out yet.)
What do you guys think?
EDIT:
Some janitorial closets use this method to add chemicals to the water in a precise ratio controlled by nozzles. I want to do the same thing, except I want to control the ratio by the solution temperature out of the bottom west port.
Facebook
Twitter