Water cooling with a ground loop (no radiator)!

[eaglescouter]

My outside air temperature gets quite high during the summer, (100 to 112 degrees is not unusual), my inside ambient may be 95-98 degrees at times.

So.... Knowing that my temperatures are already high, (on air cooling) and believing that even water cooling cannot go sub-ambient, and not wanting to mess with refrigeration, I am considering an alternate source of cooling.

My concept is to run a water cooled system replacing the radiator with the natural cooling power of soil. Three concepts are offered for peer review.



Here are the three concepts for comments:

1. Replace the radiator with a large coil of buried copper tubing, buried under the house.

or

2. Replace the radiator with a buried steel tank under the house. The tank provides for the heat exchange with the soil, and allows increased btu capacity due to more water mix being stored. (Is 5 gallons enough? )

or

3. Combine both options one and two. The addition of a copper coil allows for an increased heat exchange effeciency.

All three of these options allow the pump to be located under the house so I don't hear it. The tank options (one and three) allow the use of a submersible pump.



Assumptions for each concept:

a. Under the house, there is no direct sun, and the soil is always cooler than the summer daytime high ambient temperature. I'll insulate the lines that are exposed to the air (between the floor and the soil) so that the sysem can't freeze during the winter or gain heat during the summer. (Note: I'm in a region where the soil does not freeze during winter. Folks in colder regions would have to bury below the frost line and insulate down to the frost line.)

b. A larger pump will be needed to handle the extra head pressure and the line resistance for the longer lines and buried components. This will increase my electrical costs a bit, but no fans except for power supply!

c. I own the house, so I can drill the necessary holes in the floor.

d. I can powder coat the outside of the buried components to prevent corrosion. I can also powder coat the inside of the steel tank to prevent/minimize the battery effect.

e. I can include a pressure or flow switch to shut down the PC power if the water flow stops due to failure in the system (pump stops, lines freeze, system leaks.) I have not designed this part of the system yet, but I'm sure it can be done.

f. I have access to a milling machine so I can make my own CU water block, and I can make and powder coat the steel tank.


I need your opinions, what am I mising? Are there other problems that I have missed?

 

[toxik co2]

Your gonna need a serious pump to keep a good flow. I wonder how deep you have to deep where the soil/mud is noticable colder? Sounds like a good idea to me if you can pull it off!

 

[The Spyder]

3. Combine both options one and two. The addition of a copper coil allows for an increased heat exchange effeciency.

All three of these options allow the pump to be located under the house so I don't hear it. The tank options (one and three) allow the use of a submersible pump.
~did that, works great, but i dont have neopreme, and i get LOTS of condensation, still cant find the blopdy pics of when i did this....

 

[eaglescouter]

24 inches of buried depth should provide substantially cooler soil during the summer. While it is true that a fishtank pump is not designed for this much head pressure I can acquire a 110 V AC water pump designed to deal with this amount of head pressure.

 

[GTA]

Bladerunner has done the tank method, and finds that it works very well indeed, he gets 18 degree C water all year round. That is in the UK however, and I think the tank is about 8 feet underground. You might not get temps like that, I don't know if theres much difference in soil temp to this where you are, but I think that what you suggest would be very effective. Bladerunner uses a Eheim 1260 ( same as the old 1060 )

I'd suggest something like a swimming pool or central heating pump would be ideal, large head.

You may want to drop him a line, he obviously knows more about it than I do.

 

[eaglescouter]

Soil temperatures as measured for farmers (6 inches down) near my community:

Jan 7-11 C
Feb 7-11 C
Mar 9-14 C
Mar 9-14 C
Apr 15-17 C
May 16-22 C
Jun 17-23 C
July 22-23 C
Aug 21-23 C
Sep 18-21C
Oct 13-18 C
Nov 11-13 C
Dec 8-12 C

Looks like a great temperature potential since I'm currently at 28C on the motherboard and 55C on the CPU.

 

[TeRRaN`uSER]

be sure to post pics and results once you finish this project up


-Patiently Waiting-

 

[eaglescouter]

Right now this is only a concept. I'd like to give the forum at least a few days to work the bugs out of the concpt before I start building.

I promise to take pic's when building begins.

I was afraid that my concept would have been poked full of holes by now. Nice to see that the replys have all been positive.

 

[Literatii]

That sounds like a hellava project! Take lot of pics... I'm sure some mag somewhere would buy them!

 

[inspectorhammer]

Geez thats about enough to convince to me try that! Love the idea. Would the coils and or res. be removable at some point or would have to dig them out?

And are you in a basement or on the ground level? Because your talking more like 4 - 5 feet under than 24 inches if in a basement

keep us up to date on progress!

 

[Diggrr]

You may want to dig a test hole before you begin. BladeRunner already stated a couple of times that his tank needed to be in wet soil to transfer heat well. In a dry soil, the heat doesn't transfer from the system well enough, and temps rise in matter of hours.
If your system's only on a couple of hours per day, then no worries using a tank alone, but if it's on much more than 4-6 hours per day (totally guestimating-tank size) then your temps will start rising to where you are now.

And by wet soil, I mean when you dig a hole and water fills the bottom. Let it. Then bury your apparatus below that water level.

Now if you're only option is dry soil, I'd use tubing and lots of it, ie. rent a backhoe unless you got time to kill.

The ground is a nice cool 'sink, but it's totally crap for transfering temperature if it's dry. You'd just warm the soil surrounding your system, then your temp transfer would grind to a halt. Water's the secret (as we all know).

 

[eaglescouter]

I don't have a basement.

I assume that this is a bury it and leave it system, I do not plan to remove it at a later date.

I did not plan to dig down to the water table, too deep, and impractical in a four foot tall crawl space under the house.

I believe the natural soil moisture will be adequate if the coil is 24" down. If not, it would be easy to add a line off of the drip irrigation system to provide a bit of moisture for our project.

 

[Miralcos]

That is a great idea! Might want to read this site because it might help you. http://www.groundloop.com/

Geothermal air conditioning has been around for a while now.

I hope you do it because I believe you will see great results, and I would think the coils of copper tubing would work better than a tank. Just my $0.02.

GOOD LUCK!

 

[Molybdym]

If it is going to be a 'leave in' project, imagine the resale value of the house!

Water Table is a little far, but soil will stay moist (man I hate that word) probably a little farther than 2 feet, when my dad dug a foundation for his garage, he went about 4 feet and the soil was damp to the touch.

This is a great idea and I would love to see it accomplished, or if Blade's already done it, again

 

[inspectorhammer]

How many square feet are you looking to consume with these coils?

I can picture a nice grid of 10ft copper pipe about 5 wide and joined at the ends with elbows.

thatd be sweet

and youd have a couple gallons of water into it without any reseviors.

Man i wish i could go outside and dig some test holes but the grounds still frozen here!

 

[FreeRider]

I think this concept will work good. I would use a copper coil, because the surface is larger than a tank. If you can reach the groundwater it would be great. But, the pump has to be able to lift the water 1-3 meters, so you need a heavy one... Try to test the setup before you dig it in (with the same height-differences).

I cool my cpu with the outside temperature too, and you can get some nice results with it !

 

[TyRex]

I would bet the best place would be to dig a hole in the crawl space, that is if you dont mind, it's be easier to bring the tubing and connectors up through the floor that way. Or you could just put a big tank full of water in the crawl space instead of digging.......many many options with this idea

 

[BladeRunner]

Here is a recent bit I wrote about the underground tank:-


"On the underground tank cooling there are a few things I've learned after living with the set-up for a year.

For it to work efficiently the tank will need to be fully immersed in the ground water. How deep this is will depend on your specific location, but when the hole you've dug fills with water, that's the ground water level. It can change from winter to summer so it's best to find it when it's at the lowest point, generally in the summer after a longish dry spell. If enough coolant is used, then it will work without being immersed in the ground water but not as effectively, and will tend to slowly heat up. My current 100-litre propane tank 8-foot underground only has the bottom foot or so immersed and it does slowly heat up. I don't run my PC 24/7 so it's not a major issue. But it needs to be appreciated we are only talking maybe 3°c to 4°c rise in a 12 hour period of full load PC use, today for instance my PC has been on for almost 10 hours and the temp on start up was 12.8°C and has risen 2.5°C since this morning. When off for a period the underground temperature returns the coolant to whatever the underground temp is. This in use temp rise would not continue until the coolant boiled of course, as there would be a stabilisation point. I have never run it long enough to find this point as I don't like leaving the PC on when I'm sleeping, but my guess is it would never get to a temp as high as a "normal air-cooled rad water-cooling system" starts out at. What it possibly would not cope with is a high watt TEC especially run for long periods. I will be making a second underground cooler build sometime, as I'd like to see if it can be improved with copper coils immersed under the water table to cope with high watt TEC use.

One last point is the assumption a mega powerful pump is required for this type of set-up. The way I have it set-up with a header tank / reservoir above PC level is exactly the same principal as a UK domestic household central heating system. Just as in the central heating system there is no pressure pumping, just circulation, and central heating systems cope with a two or three story house using a normal low-pressure circulation pump. Circulation is something aquarium pumps do well. I'm using an Eheim 1060 that gives me good flow rate through all the blocks in my system, although it's a bigger pump than would normally be required in a "standard water-cooling system" this is really only due to the extra distance / tubing the coolant has to be circulated around going to & from the tank. The header tank also makes system bleeding a non-issue."

This basic schematic below should help explain the workings

Those that don't know what I'm on about read HERE (and go to the "bomb" page).

 

[eaglescouter]

BladeRunner:

Thank you for a great post. You are well ahead of me but on a slightly different track.

My goal is sub ambient water cooling without a mechanical chiller. However I want to use a coil of copper tubing for the majority of my heat transfer, and a tank for increased btu storage.

I believe most readers will be put off by the need for a back hoe and an eight foot hole is a bit much for hand digging.

Thus I am focused on work that I can do under my existing house ( so I cannot use my back hoe for this project).

I think a flat coil of 1/2 inch copper tubing buried 24 inches deep will offer adequate heat transfer. Tubing coil will be about 24 inches accross (has to fit through the access door) and tightly coiled upon itself. (Like the coils of a snail shell.) I still need to determine the length of copper tubing for the coil. If I don't get adequate cooling with the first unit, I can add an additional coil.

The return line from the CPU will run through the coil, then to the tank so that stored water is precooled then stored in the buried tank.

My tank is not for thermal transfer, rather it is for increased water storage capacity, any thermal transfer is an unexpected bonus.

Since I live on a hill, it is unlikely that I can reach ground water (with its excellent thermal transfer properties), however, the soil in my area is heavy and moist, so adequate thermal transfer should still be feasible.

 

[NO LIFE]

Keep in mind it wont be easily reached for maintanence(i cant spell). Do it right the first time or your toast.