Awesome new cooling idea, and it works this time!!!!!!!

J

joeswm8

New Member
i got a new idea!! we learned today in chem that the vapor pressure of water must be equal to the external pressure in order to boil. The vapor pressure of water at 35C is 44.2 mmHg, or about 6% of the regular atmospheric pressure around us. using a closed loop only consisting of a block and a simple condensation rad connected by tubing, we have water filled in the block and up both sides of the tubing with the rad at the top of the case so the block is at the bottom of the loop and enclosed with water at all times. With a T-line and a one way valve, the pressure is lowered using a micro pump to around 44.2 mmHG, or to preference, in order to get the water to boil at a specified temperature. When the CPU is then loaded, the water will boil at 35C and stay a that temp while absorbing loads of energy. The boiled water will then condense in the upper radiator structure and effectively direct the water back into the block, creating a continuous process with lots of cooling potential. If water boils at 35C, it stays at its boiling point until all of the water is boiled away, however if there is a continuous source, the system will stay under 40C at all times.

I believe this idea has lots of potential and is easy to carry out. Your thoughts?
 
you do realize thats a D-type boiler with no safety mechanism right? the cpu is the burner, the block is the feedwater drum, the upper radiator structure is the steam drum, and your condensate return is, well condensate return(look at the diagram here: http://en.wikipedia.org/wiki/Water-tube_boiler) Vapor at 35C is the same as vapor at 100C, its vapor. without control of the heat generated by the cpu you will only continue to gain vapor. by placing this vapor in a vacuum you make the critcal temperature of the vapor linear to its boiling point and once critical temp is reached you will essentially have created a bomb.

Boiler safeties are used to ensure this type of situation does not happen(city water make-up, low water cut off, high pressure cut off, etc, etc.)

In the situation you described you are using the CPU as a burner. This is a burner that will get hotter and hotter, increasing the vapor in the entire system. Water as a vapor occupies about 1600 times more space(don't quote me on this I'm going by memory) then water as liquid. Where is this space going to be when it eventually reaches its critical point?
 
I know it sounds like I'm killing every idea you've had and you probably don't like me right about now but I;m just being a devil's advocate. I've gone to school for a while to learn this stuff.

The best way to get rid of the heat generated by a cpu(real hot, real fast) is by first absorbing that heat quickly and alleviating the absorbed heat even quicker. Try thinking about it differently. Most people think of using A/C in the summer time as cooling the room. Trained professionals know it as removing the heat from the room.
 
the pressure would leave the system through the one way valve. However, the water would all be gone in a short time. But if the water vaporizes and then condenses soon enough, there should only be a little section of higher pressure that can be accounted for.
 
joeswm8 said:
But if the water vaporizes and then condenses soon enough, there should only be a little section of higher pressure that can be accounted for.
Click to expand...

Pressure is created throughout a system. The only way you are going to contain pressure is by isolating it and not allowing anymore in or any of it out. Your water will eventually disappear enough where the water and vapor will become the same temp and the rest of the water will flash into vapor, skipping the actual phase change. This is where your explosion will come from. The lower the boiling point of water, the easier it becomes to create this type of situation.
 
Last edited:
The basic concept you described is a heat pipe, which has the advantage of being a closed cycle.
 
i have revised the idea. The would be a three or four way barb thing, one to the block, one to the micro pump, one to a resavior, and one to nothing. The mpump and out would have one way valves that let pressure out and the res barb would have a valve to refill. It seems the only problem with this is the pressure, and upon containing this correctly, it is possible.
 
joeswm8 said:
i have revised the idea. The would be a three or four way barb thing, one to the block, one to the micro pump, one to a resavior, and one to nothing. The mpump and out would have one way valves that let pressure out and the res barb would have a valve to refill. It seems the only problem with this is the pressure, and upon containing this correctly, it is possible.
Click to expand...

pneumatic lines and P/E devices throughout to control everything with the computer having control over the P/E devices to tell the sensors what to do at certain points on the slope/y intercept. But my guess would be that as soon as the vapor under the a vacuum leaves that vacuum it would become superheated and expand at the rate that it would under atmospheric pressure.

At this point, and even before this point anyone would be better off paying $50 for a good heatsink and fan.
 
Last edited:
i tend to disagree. A Valve leading to out of the case causes no harm, and its not a 100% vacuum either. And the gas would not be superheated, its the same temp inside and out
 
Last edited:
if you don't have a 100% vacuum then how are you going to keep it boiling at a low temp? And you're right. I went about this backwards, venting vapor in vacuum will cause the water vapor to condense. But then that leaves another problem, how do you continously have a vacuum?
 
yeah! got you on this one! touché lol

for water on any liquid to boil, the vapor pressure of the water must be greater than the external surrounding pressure. If i want to have the water boil at 40C and essentially always stay at that temp even during full load (because when boiling the temperature of water stays the same):

vapor pressure of water at 40C: 55.3 mmHg

therefore the vacuum must take out enough gas to lower the pressure to 55.3 mmHg, allowing the water to boil at 40C. And say I pump this pressure out and keep the vacuum on a very low speed to keep up with the vapor pressure created by the water, it would only have to maintain 55.3 mmHg which is much much easier to do than a full vacuum


Initial Drawing Stage (1 of 2 possible setups, not the one with a copper condenser, will make that one soon):

Boiling_Cooling_Idea.bmp
 
Last edited:
1) At 40 C and 50 W of heat generated by the CPU, you’ll need to pump about 24 liters per minute through the vacuum pump (assuming you can the system down to 55 torr in the first place). That’s not impossible, but most roughing pumps I’ve worked with can’t do this (and these are expensive pumps).

2) A heat pipe still does the exact same thing (evaporation of water/ethanol) in a closed cycle with no need for pumps. The water won’t be as cool as if it were at room temperature as is the case in your setup, but the heat is primarily going into the latent heat of vaporization – latent heat of vaporization @ 40C = 2.4 MJ/kg, enthalpy change from 25 to 40 C = 62.7 kJ/kg or about 3% of the heat of vaporization.

At this point, and even before this point anyone would be better off paying $50 for a good heatsink and fan
Click to expand...
Agreed.
 
I don't like it. I'm telling you there's something very dangerous about a setup where you have a constantly running heat source for boiling water.

Let me do a little bit of reading and I'll get back to you.
 
I don't like it. I'm telling you there's something very dangerous about a setup where you have a constantly running heat source for boiling water.
Click to expand...
If it were water boiling at atmospheric pressure then yes, I would agree with you. In that case the danger comes from the potential to build up pressure, and if the apparatus isn’t designed to ASME standards then you have a bomb or a missile (the results of which I’ve witnessed) on your hands. In this case, however, you have a negative pressure differential; the worst I could see going wrong is an implosion due to poor materials/assembly or the vacuum pump blowing up from being overworked.
 
exactly!! its not over pressure and 100C guys!

its very low pressure (often not dangerous) very low temp (40C) bldengineer is thinking the other way. And why would I need a pump? The water sitting on the CPU absorbing the energy would not pass 40C because of the principle of boiling. The water stays at the boiling temp until it all has turned to gas. No pump, why would I need one anyway? Only a res letting new water in every so often, and the vacuum maintaining the pressure in the tube. Very simple
 
How about making a man made Diamond heatsink? Diamond conducts heat the best out of anything. Then the Diamond leads out the case and is a one way into a pool of liquid Nitrogen. Costly, but worth every penny....No, I'm serious.
 
Wow...I've learned so much about cooling in just this post! I'm confused and intreged! Continue on with the debate :D You just might be on to something joeswm8.
 
The water stays at the boiling temp until it all has turned to gas.
Click to expand...
Well, it would be continuously turning into gas. Think of a boiling pot of water. This is the reason the vacuum pump would have to work so hard - to get rid of the water vapor in order to maintain the pressure.
No pump, why would I need one anyway? Only a res letting new water in every so often, and the vacuum maintaining the pressure in the tube. Very simple
Click to expand...
A vacuum is a pump, and your drawing says "pump" right on it.
How about making a man made Diamond heatsink? Diamond conducts heat the best out of anything. Then the Diamond leads out the case and is a one way into a pool of liquid Nitrogen. Costly, but worth every penny....No, I'm serious.
Click to expand...
If you're going to go all out, why not use LHe3 and sapphire?
 
You must log in or register to reply here.
Back
Top