Weird ideas...

N

nephilim

Limp Gawd
Joined
Jun 10, 2004
Messages
178
Just some gonzo concepts I've been tossing around:
Why not run a metric buttload of tubing underground hooked to a strong pump? Constant sub-ambient watercooling and no need for a compressor. Granted it's not phase change, but that'd wick a lot of heat.
I'm positive this has been done, but I've not seen it personally: Hang an aircon from your window, slap on a duct to the fan hole on the side of your case, and duct back outside from the exhaust. At last glance a cheap window AC was about $90, and would probably cool to around 40 degrees or lower depending on the compressor. We know air is a lousy conductor of heat, but pre-cooled forced air works very well. It'd also only put dry air through, since the moisture would be frozen or at least condensed in the AC, not the computer box. Since pressure would be greater inside your case, there'd be no external moisture introduced either. The drawback here would be that the higher pressure environment would probably burn out the AC fan PDQ.
Or, if you're truly brave, shove a motherboard directly into a frost-free mini fridge. Same general concept as above, but perhaps a bit less efficient, slower to take effect, and more susceptible to precipitation.
 
I have been kicking around the idea of building a complete PC into one of the small frost free fridge's. It works in theory but havent quite figured out what to do with all the drives and such.
 
I dont think the systems are strong enuph in those mini fridges / freezers. I think it would work for a period of time, but after a few hours the fridge would be hotter than ambient.
 
I'm actually wanting to try the air conditioner one, since that would deal with the issue of condensation. Refrigerators are built to cool things that aren't actively creating heat, and without air flow that is accomplished by a single cold plate powered by a compressor.
A/C has an advantage there because it pulls air in from the room and then cools it. Moving air is far more efficient for carrying heat than still. If you exhaust the (comparably) hot air outside of the room after it leaves the PC case then your house's insulation does half the work for you, rather than in watercooling systems where a lot of energy is spent cooling the same water over and over again. The hot stuff is outside the system completely.
There WOULD still be an issue of condensation on the case surface, but higher pressure inside the case coupled with moving air would very probably work to keep moisture out of the case itself, though rivulets might form outside the case.
Obviously this would take a LOT more power than your average watercooler, or even a peltier or special-purpose (IE, computer) phase change system. Still, power's cheap where I live and a small A/C would cost about $90.
It'd put a huge strain on whatever air conditioner you were using though. You'd want to keep the compressor on almost full time to maintain a stable temperature, or I suppose you could wire the sensor from one of the thermostat models inside the case itself. I'd definitely wire a kill switch to shut down the PC if the A/C ever failed, and set the system shutdown temperatures very aggressively.
Still, it might be a fun project. Grab one of those cheapo Pentium D's and a solid mobo and see how far ghetto tech can take it.
 
I have been kicking around an idea like that (quite a few different ideas for my next build.)
what I am considering doing, is to remove the fan and ducting from the a/c unit and building a platform (to keep the computer parts dry.) then I would put a 12v electric radiator fan from a car blowing air in through the evaporator core.
this would have 2 effects. It would supply a cool dry air supply to the computer, and it would keep the condenser core cooler than normal.

so basicly, it would suck air in from the room, push it accross the computer components, then as it exhausts it would cool the high pressure side of the a/c, and exit the unit outside.
 
nhusby said:
I dont think the systems are strong enuph in those mini fridges / freezers. I think it would work for a period of time, but after a few hours the fridge would be hotter than ambient.
Click to expand...

This is true...especially with mini fridges.

I actually took my older system (3.4E and X800XT-PE watercooled) and put it in my parents 6'x3'x3' chest freezer. At first I had load temps on my X800 of about 18C and the NB was below 10C...but after awhile you could put your hand on the outside of the freezer close to the top and feel the heat.

Using mini/chest freezers for cooling is kinda like LN2...it works pretty well for short OC/benchmark runs, but is not something that can typically be used for a 24/7 system.
 
I would be interested in the ground cooling though. I have not heard of anyone doing it yet. But I amagine you should get something in the range of 20*C because the ground is a constant at about 11*C year round (below the frost level anyways) my only question is, how well does the ground conduct heat? if coolant temps get above 15*c (like 20 or better) then I would add a water bong before the ground loop.
 
I made a post in another subject. I'm thinking of trying it out with just a water block running off the faucet.
 
nhusby said:
I would be interested in the ground cooling though. I have not heard of anyone doing it yet. But I amagine you should get something in the range of 20*C because the ground is a constant at about 11*C year round (below the frost level anyways) my only question is, how well does the ground conduct heat? if coolant temps get above 15*c (like 20 or better) then I would add a water bong before the ground loop.
Click to expand...

People control homes this way, so it's certainly enough too cool a lowly PC.
The only issue is you have to get into the water table to get good heat transfer. Basically you want your cooling pipes laying in wet mud all the time. Around here you have to go through tens of feet of red clay to accomplish that.
 
I have seen people using the below ground cooling before and it has worked quite well. I haven't looked into it (in terms of setup/prep specifics) but I'm sure it could be made to work, and would/could be a fairly realistic option if done right.
 
minimum 6ft deep to 10ft deep is optimal... Now digging that deep of tranch isn't easy nor cheap. What happens is that you certainly get more consistent temperature that is cooler in general all year long :) I think it is one of the most extreme watercooling method i've seen so far :)
 
I was thinking of using an ice auger (I'm in wisconsin) to dig about 4 or 5 feet, then get another couple feet with a post hole digger. Then just coil up some thin copper tubing around the outer edge of the hole in a spiral and bary it.
 
jinu117 said:
minimum 6ft deep to 10ft deep is optimal... Now digging that deep of tranch isn't easy nor cheap. What happens is that you certainly get more consistent temperature that is cooler in general all year long :) I think it is one of the most extreme watercooling method i've seen so far :)
Click to expand...

Good advise man...granted that kind of digging is no small task...but it'd definitely be something worth considering doing beforehand if someone was building a house. Try explaining to the construction guys why you need this 10 foot deep hole dug out from your house. :p
 
Just tell him it's so the magic computer fairies can make your games go faster.
If I lived in the basement underground cooling might be practical, but piping to the second floor would be hard to justify.
"You're going to run a pipe WHERE?!?"
Ground cooling would be by far the most energy efficient way to cool below ambient. You'd need a pretty strong pump though, and whatever material you use for tubing it'd have to be very corrosion resistant. Copper would probably work so long as you're careful about scale and particulates. (Which is probably a given, but I'm used to addressing the unwashed masses, so forgive me.)
It'd also be the most sustainable. I mean, air conditioners and mini fridges would probably die in a few months due to strain, but I doubt the ground is going to start being hotter than the outside air anytime soon.
 
nephilim said:
Just tell him it's so the magic computer fairies can make your games go faster.
If I lived in the basement underground cooling might be practical, but piping to the second floor would be hard to justify.
"You're going to run a pipe WHERE?!?"
Ground cooling would be by far the most energy efficient way to cool below ambient. You'd need a pretty strong pump though, and whatever material you use for tubing it'd have to be very corrosion resistant. Copper would probably work so long as you're careful about scale and particulates. (Which is probably a given, but I'm used to addressing the unwashed masses, so forgive me.)
It'd also be the most sustainable. I mean, air conditioners and mini fridges would probably die in a few months due to strain, but I doubt the ground is going to start being hotter than the outside air anytime soon.
Click to expand...

Exactly...it's just a shame there isn't a less invasive way to go about the whole underground cooling thing...sounds like it'd be real fun. :D
 
live in the country? got a well? its technically renewable. run the well water through a water block, use the wastewater to water your grass. Its not a biohazzard, and contaminants are insignificant.

I live in the city. So such an option is not renewable on the pocket book, as I pay for my water by the gallon. Its renewable over time logistically. Wisconsin water comes from the ground, and I would just return it. at about 5 cents a gallon... How many gallons would I use per hour?
 
nephilim said:
Just some gonzo concepts I've been tossing around:
Why not run a metric buttload of tubing underground hooked to a strong pump? Constant sub-ambient watercooling and no need for a compressor. Granted it's not phase change, but that'd wick a lot of heat.
Click to expand...
See here for ground coupled / closed loops systems used for the home. Everything you need to know. No reason you can't adapt one of them.
 
how many gallons per hour would you estimate it would take to adequately cool a computer? I am thinking something along 30? assuming the water was at or around 50-60*F and I use peltiers. (more efficient heat transfer, and better capacitance per gallon because I am flushing the hot water)
 
I am an IT professional, and I actually had to do the AC one in server emergency once!!!!

We had a remote site with two servers, and one of them went down all of a sudden. Up-time was vital on the server so I rushed over there to check it out. When I got there and tried to reboot it told me the CPU fan wasn't working (compaq, so it had lots of built in reporting). The fans are hotswappable so I tried moving a less vital one in to cover the CPU, but then two of them had red lights indicating they were bad even though the one I moved was showing good. So I realized it wasn't the fans, but the fan harness system.

The people in the office were going ballistic by this point, and they said they would lose a million or so dollars of current and future work if they couldn't get back on the server for at least a couple hours to complete client deliverables. So I called Compaq to come in, but it was 3pm already, and they have a 4 hour turnaround time on this kind of thing...

So I was trying to figure out how to cool the thing when I realized suddenly that I had a portable AC in the room, some packing tape, and a 3 foot piece of 5 inch diameter plastic bendable and strechable ducting (like on the back of a dryer) that would reach out to about 15 feet if I needed it to...

Needless to say within about 10 minutes the server was back up and running... And it was probably the coolest it had ever been, what with all that AC powered air blowing straight onto the CPU!
 
Good stuff, nice story man.

Ducting AC cooled air straight into your case is another way to do things...as you're then limited moreso by the temperature of the AC'ed air when you're working with your CPU temps.

Kinda like when I had my comp on a chair beside the front door with the screen open when it was -40C outside...with a huge floor fan blowing the outside air right at the comp...does wonders. :p
 
hmmm.... I'm thinking phase change unit coupled to a ground cooling loop.
 
nhusby said:
hmmm.... I'm thinking phase change unit coupled to a ground cooling loop.
Click to expand...

Ground cooled loop with Plate HX acting as condenser will enable use of r410a which is cooler obviously vs most ss units :)
 
You must log in or register to reply here.
Back
Top