passive water cooling

How big of a container would i need to passively cool the water from a waterblock attached to a intel 2.4 running at 3.0? I have two empty 50 gallon and one 180 gallon fish tank that i can use as containers. I am also thinking about creating more surface area for the water to cool on by inserting tubes into the water (sort of like straws that run up and down and allow air in and out by normal convection of hot air up cold air down).

good lord, using one 50 gallon tank with the top off would be fine. a fan blowing across the top in some way would possibly keep it at or below ambient... but not necessary. 50 gallons is a BIG heatsink

The 50 gallon one would be fine, if not excessive.

Good luck getting 50 gallons of distilled water though

(Naah, i don't recommend using distilled water in there. Might help to use soft water though)

The thing is, I am worried about if the heat will disapate fast enough so the water will never increase in temp. 50 gallons is alot and beable to absorb alot of heat but will it be enough surface space for the contacting air to act as a radiator. If it will only work for like 2 days before the water is warmer than ambient I don't want to do it.

Only one way to find out

make sure you have some way of removing humidity from the air around the container. It will peel wall paper, discolor surfaces, cause mildew or mold to form in places.


Also, make sure input and output tubes are at the same level in the water. Having them near the top will not be as effective as having them at the bottom. and of course, having tubes from the top to bottom will seriously increase the length of tubing use and decrease flow rate. so hope you have tube connections at the bottom of the tank. I don't see how this is going to be efficient or even practical. Better idea would be to get a huge heatercore with internal water routes larger than the tubes bringing water in and out to minimize flow reduction while still giving plenty of surface area for water to be against.

I'd consider a more practical setup, i.e. an inline loop with a proper airflow design. Really, a largish radiator could accomodate two 120mm fans easily and these have to run at very mild speeds only. I'd prefer a soft breezing sound over a bucket of water standing in my room, not to mention the mold/humidity problems it brings along...

I think if you where to put the tubes in and connect them to a sealable top, you'd be killing 2 birds with one stone. Think of a nuclear reactor and its control rods for the top for the setup...better yet a swiftek pin-fin HSF. With a 50 gallon tank I think the sides of the glass could cool it fairly well, so this convection idea would probably help put you over the top. You might not need a 100% seal, but enough to keep moisture in and let air pass through.

I don't know where people get this, but glass is NOT a good conductor of heat. If it were made out of any metal he's see a tremendous change/improvement.

people would see a tremendous change and improvement if they kept this in mind all the time whenever thinking of a novel way to cool their computer.

The novel solution shouldn't require more work than conventional solutions.

here we define work as any expenditure required by you, the users of the computer/device to use your solution. That includes, money, time required for maintainence, change in regular activity, alteration of environment.

If it fails at this then why are you wasting your time on a idea that sucks more than what's already out there? I'm not saying that's what this idea is. But if it was, it's not something that really belongs here. I'm sure the majority of regular readers here are interested in the ideas that dont fail the above test, and find the ones that do fail to be annoying because they make it harder to find the ones that aren't just toy ideas. toy ideas really belong in the extreme cooling section since extreme implies impractical in this case.

real world usage:

i got 30 liter brewing barrel(what's that, 3-5 gallons?) that i use for cooling, top off. with a fan blowing on the water surface the water temp stays around ambient/1c under it. without fan the water temp jumps to 30c. (these numbers come with 1.92v tbred-a @ 1900mhz).

i have to add water about once in 5 days, i think it would survive 7 because as the water level drops the fan is further away from the surface and the evaporation slows down and the temps get slightly (1c) warmer.

normail airflow needed to keep the air breathable is enought to get rid of the humidity, of course, if you live in a tropic swamp this might be diffrent(the humidity transferred to air is around the same as if you had an warm water aquarium or just another person breathing in the room).

"glass is NOT a good conductor of heat" - umm, duh? If it was a glass radiator, it would really suck. But in this case the water stays in the res (50 gal fish tank) for about 40 minutes to cool (50gal/~70gph average = 40minutes for complete cycle). Even at half a gallon (if that), a normal heater core holds the water for less than one half of a minute to cool. A 50 gallon tank would easily have as much/more surface area than most of the smaller rads people use here. While glass isn't a great conductor of heat, it still IS a conductor. Everything strives for an equalibrium - a passive glass tank just takes longer than a rad (just needs to be less than 80 times longer than a rad, as it holds the water 80 times longer). A 50 gallon tank has a lot of surface area; most are rectangular, which have a relatively high surface-to-volume ratio. If the water can spend [a lot] more time radiating heat, the transfer doesn't have to be as efficient.


"If it fails at this [saving time/money over conventional means] then why are you wasting your time on a idea that sucks more than what's already out there?" Even at 5 gallons, the water sits in the res for 4+ minutes passively cooling (I'm using a 70 gph working average most people would see). I know for a fact that you can easily go to a flee market/garage sale and find a cheap fish tank [as in much less than most rads]. Many people have them but don't have a use for them atm. Also, one of the main ideas of water cooling is to reduce noise. A pump in a 50 gallon tank is never going to make more noise than an external pump, nor will you really need a fan. If this system [50-100 gallons] was used w/o a top, the evaporation alone could cool the cpu, albiet not as well. Water usally takes energy from the water to become vapor, not from the air.

No offense safemode, but, and I'll even use your quote, I think your post is "not something that really belongs here [in this thread]". I'm not flaming you, as I can't tell you what you consider a toy idea and what isn't; also, I can't tell if you are just lecturing or saying this idea is impractical. The fact is that this is EASILY the cheapest way to cool; no leaky barbs/joints or a rad that traps air bubbles, just a smallish fan and a bunch of water. Likewise some people spend many hours modifying thier rads. If you get a used tank, you need only rinse it a few times with cleaner, but you might have to soak a used heater core overnight, and then mod it a bit. They may be the same price used, but the rad needs much more equipment to modify.

All things considered, this rivals a normal rad setup. Cu Rad = compact, fish tank rad = simple.

I use to run 14 gallons, with a 25cfm fan in the corner blowing across the water, It kept it +/- 1c from ambient, worked great! just get some humidifier stuff so icky stuff doesnt grow in it. It works great! I was gonna switch back to it, but summer is comming and humidity is going up

You don't really need that much water to run passively though, I'm running my T-bird 1.4 passively with one small internal rad (black ice pro, which is not doing much anymore, I just left it in cuz I used to have a fan on that) and one larger rad on the outside of my case, about 10x5 inch maybe. Of course I'm not runing ambient, but it's fanless and works fine. Pics at http://bernhard.spalteholz.ca

FIZZ3 said:

I'd consider a more practical setup, i.e. an inline loop with a proper airflow design. Really, a largish radiator could accomodate two 120mm fans easily and these have to run at very mild speeds only.

Click to expand...

I'll drink to that.
After all, if it's noise you want to control, the psu fan will be louder than two panaflows on a rheostat with shrouds.