Watercooling setup question...

[engjohn]

Take a look at the pic and tell me what you think. Do you think it will work or will the temp equalize...

The pelt is 80w runnin off a 15v 12a PS and gets mighty cold...
The CPU is a Celeron2 800 @ 1066 (133FSB) it puts out about 36.4 Watts.

I kinda want to use a chiller instead of directly using the pelt and the coldplate for condensation issues.

Thanks

 

[AmIdYfReAk]

kinda of a stupid question.. but...... whats goin to keep the water moving?

 

[Colin]

You are going to have to slow the water down through the chiller to transfer the heat. Not very practical. There is some good information on chillers at overclockwatercool.com. Note that I do not recommend their services.

 

[engjohn]

AmIdYfReAk (Jun 21, 2001 08:10 p.m.):
kinda of a stupid question.. but...... whats goin to keep the water moving?

Sorry, Of course I have a pump in the reservoir....

 

[engjohn]

Colin (Jun 21, 2001 08:15 p.m.):
You are going to have to slow the water down through the chiller to transfer the heat. Not very practical. There is some good information on chillers at overclockwatercool.com. Note that I do not recommend their services.

How about this second Idea, or am I still barking up the wrong tree here?

 

[TT120]

Y'know, I had the very same idea as your first one. I didnt have another waterblock so I set up a test rig like your second idea except with an Alpha PAL 6035 instead of the pep. The water didnt get that much cooler and I figured it was because the water was going too fast through the waterblock on the pelt. I needed to figure out how to either slow down the water throught the block or make the layer thinner like a pasteurization process done to milk only in reverse. instead of heating the milk (water) I would be cooling it. Haven't quite figured out how to do it yet but I'm workin on it.

 

[William]

its easy to slow down the water, sord of, just use larger pipes.

 

[Colin]

I played with a pelt, Amazing block and a Pep66 one evening. As I suggested, a trip to Chip Eckert's Site will make the issues involved obvious. Use a radiator, bong or bottled water cooler.

 

[cjtune]

I think that the only reason one might use this setup is to displace the problem of condensation elsewhere rather than at the CPU where most ppl place their TECs. Otherwise, heat transfer by convection (any fluid, be it water or air) is less effective than conduction (soild-to-solid interface). Why do you think you need up to 1/4 m^2 of area in the form of radiator or heatsink fins to throw away heat that was conducted away from just 1 cm^2 of exposed CPU core area?

 

[Spewn]

I hate how many times I have to say this but: You cannot slow down the water at any point in your watercooling setup without slowing down the water at EVERY point. That's final, no if's and's or but's. With that out of the way, you STILL run the risk of condensation even with that setup. Assuming your radiator is efficient enough to cool the water to ambient levels, the pelt will(idealy) drop it BELOW ambient. That will cause water to condense out of the air. This will only occur at the secondary water block, but it will still occur. The main problem with this setup is the efficiency of it though. As far as my understanding of peltiers goes, they're designed to be heat pumps, not magical icy hot pads. The water doesn't have that much heat for the pelt to pump away, so it won't work as efficiently as it would if it was directly on the cpu(at least, I would think as much). The other thing is, you'll need to take a lot more energy out of that room-temperature water to get it to cool down a few degrees than you'd have to take away from your cpu, given water's VERY high specific heat capacity. In my opinion, it isn't worth the effort to try and chill the water with a peltier(or the added cost, 2 waterblocks = expensive), but it's your time, and your money. If you do end up doing this, make sure you post your results

 

[cjtune]

cjtune (Jun 22, 2001 09:15 a.m.):
I think that the only reason one might use this setup is to displace the problem of condensation elsewhere rather than at the CPU where most ppl place their TECs.quote]

Yeah, Spewn, you're right. However slightly the water is chilled, there is still a chance of cendensation at the CPU.

 

[cjtune]

It'll be easier for you guys to predict condensation by reading off values from what is known as a Psychrometric chart. It's a chart relating values of dry bulb temp, wet bulb temp, air density, enthalpy (energy), specific humidity ratio, relative humidity together at a set pressure. Ppl working in the heating, refridgeration, and air-cond maint. and design business usually have this chart -see if you know anybody there. By setting two values (eg. room/dry bulb temperature and an estimated/measured value of humidity), you can determine the dew-point temperature (temp at which condensates form) of the air. If anybody needs to know, room air at 25'C and 60% relative humidity will form condensate at 17'C and if you live in the tropics, 30'C @ 80% relative humidity will spell trouble for your PC at a mere 26'C.