my diy cooling system

[spazz]

i made this cooling grid in my basement crawl space that has a sand floor. it is made from 11- 1 1/2 pvc pipes 10 ft long. i could not see buying a chiller that would cost me about $700.00 plus the 4-5 amps of power to run it every day. the only power i use is a 800gph pump to circulate the water into the cooling grid. it runs about 80 to 100 watts of power. this type of cooling system wont work if you dont have some sort of crawl space under your house. unless you drilled 2 holes through your wall and dug up your yard to lay the pvc grid outside. i dont think your wife would like you digging up the yard just for the fish tank. the grid works well. i keep my tank temp at 79 deg. all the time. i have to wait for the summer to see if the cooling grid will handle the daytime heat or if i will needa bigger cooling grid. the idea is to run the water through the grid as slow as possable in order for the heat to transfer to the sand. i use a pvc ball valve on the return side of the grid to control the flow. the water cools down very rapidly comming out of the grid. i get about a 10 deg drop in temp comming out of the grid at (have not checked gpm)about 100 gph of flow. if you raise the flow rate i think you would shock the fish by such a drastic temp drop. i use a medusa temp controler to run the heating and cooling so the temp never goes above or below 1 deg of 79deg set point. i made this cooling grid for about $300 the medusa controler being the most expensive part of the grid.
i hope this system can help some of you guys that are worried about the cost of buying and running a chiller.

i will post a pic as soon as i can get it to work.

[spazz]

i hope this works!http://www.lunysworld.com/uploader/u...660639_cooling grid (Small).bmp

[spazz]

i think i got this right this time!

[Trackz]

Great idea!!! You could also do it with a coiled hose. Possibly take up less room? plus the walls of it would probably transfer the heat better than pvc. That's how most refrigerators with a water outlet do it.

You could also run it in a water trough or bucket of some sort. The liquid would transfer heat from the hose or pipe much better than the sand or air.

[spazz]

i think there is many different ways to make a cooling grid. i chose the one that was the easiest for me. it works great and didnt cost an arm and a leg to make. the water going in the grid is over 79 deg and the water comming out is right about 70 deg. i set my medusa controler to start cooling at 79.8 deg. so the water is droping 10deg in the short time its in the grid. if you cooled the water down much further it might shock the fish if they were to swim through the return water. you dont want to cool it down too fast, just at a steady rate. my chilling pump runs for about 5-10 min before it reaches 79 deg and shuts off. thats still fast. 1 deg in 10 min is almost as fast as a chiller would cool it down. i forgot to add that this grid is cooling down a 180g tank with a 55g fuge sump. thats alot of water!

[Soltaker]

Nice work spazz!!

If you have the room and means of doing it, geothermal cooling is awesome. Someday, I hope to use geo to help cool a decent size prop operation.

[T Sandman]

Excellent! I like economy. 20 bucks worth of pvc vs. 700$ chiller. Hmmm, I like your plan.

[xrunner1234]

Nice job.
And btw, if you increase the flow rate the difference in temperature between the inlet and exhaust will be LESS not more, so it will be less of a shock.

[Trackz]

Interesting...but how do you know that the sand isn't dissipating the heat faster than the water brings it? Maybe it could handle cooling MORE water.

You are probably right in your estimation xrunner, but it could technically go either way.

[AnnArborBuck]

From a previous thread I posted about heat transfer. Here is a link to the thread, or below is an excerpt.

----------------------------------------------------------------------
I have stated this several times (under previous username), but lets do this once more. The more flow through a cooler the better. There comes a point (as mentioned before) where increased frictional forces induce more heat then the added flow benefits, but generally speaking that only happens at really high flow rates.

When determining the size of a cooler/heater (same principal) the key component is the temperature difference between the hot and cold side.

Q=UA(Delta T)

Q = Amount of energy transfered (heat)
U = Overal heat transfer coefficient (constant for system)
A = The heat transfer area (constant)
Delta T = Temp difference between hot and cold (actually, this is the log mean temp which includes the inlet and outlet temps of both the hot and cold side)

Thot is Tank Water
Tcold is cooling water

Delta T = Thot - Tcold

The problem bececomes if your flow is too small, then Thot will approach Tcold therby your Q drops. You want a large Delta T to maximize cooling.

Ideally you want to maximize your chiller since these are expensive to run. There can be too much flow where Delta T does not increase and therefore you are not increasing your heat transfer. If you had the frictional loses for your piping, etc, this could also be calculated. But normally these things are trivial compared to the bulk energy transferred by the energy (heat is actually a flux of energy, it is a flow so to speak) exchanger.

If you want to figure out how much heat is transfered per time then add a flow rate to the above equation.

Think about it this way, do you think an ice cube will melt faster in a cup of hot coffee, or a cool Pepsi. Also, when you put ice into a warm pepsi it melts real fast to start with, but as the pepsi gets colder the ice melts slower. This is because the pepsi temp is reaching the ice temp (32F). The same principal applies to chillers for our tanks. In our case we want the ice to melt fast, therefore a large Delta T.

I tried to make this easy for everybody to understand, I hope i didn't confuse you too much.
--------------------------------------------------------------

I would also say that the reason why your water is comming out so cold is because it is probably sitting down there for a long time waiting for the pump to kick on again. You might be better off with a smaller loop and a pump that runs close to all the time (much smaller pump mind you). That way the exit temp is nearly as low.

You will not see a decreased temp with a higher flow, however you will remove more heat from the system. And since 10 degrees is a big change, you might want to think about tweaking the system. Maybe have a loop you could bypass for winter mode then open for summer mode for more cooling.

[xrunner1234]

AnnArborBuck-
I think you are talking apple and oranges. Your example about the ice cube and coffee is not at equilibrium and does not apply here.

Look at it this way- Your tank is at 80F and the ground is at 70F. You set you flow to a trikle and 1 drop of water enters in at 80F and leaves at near 70F. That is a 10deg difference- it's high because the ground removed enough energy from that 1 drop to reduce the temp by 10deg in a certain amount of time. The more time, the lower the temp.

Now the tank temperature changed by very little- say .1deg because the flow rate is very small and the amount of energy transferred is minimal.

Increase the flow and you will have less time that the 1 drop of water made contact with the ground. However, overall energy transfer DOES go up (since you can get more ‘drops’ of water to make contact per unit time) but delta T between the inlet temp and exhaust temp WILL go down. Now the tank temperature when at equilibrium will be say 76F and the exhaust at 74F (just picking numbers for sake).

And frictional loses in the pipe does not come into play until we reach flow rates magnitudes higher than what our tanks can handle.

Now the temperature of the ground does increase, since the energy does have to go somewhere. Maybe that is what you meant?

[cisco006]

Do you guys think this will work in miami fl, how deep would I have to dig. I am in the process of setting up a 375g reef tank and I am very interested in implementing this setup.

Thanks for the idea spazz
cisco

[pete_lockwood]

How about using that black Siloflex instead of (I assume) white PVC?

Just a thought.

[AnnArborBuck]

Look at it this way- Your tank is at 80F and the ground is at 70F. You set you flow to a trikle and 1 drop of water enters in at 80F and leaves at near 70F. That is a 10deg difference- it's high because the ground removed enough energy from that 1 drop to reduce the temp by 10deg in a certain amount of time. The more time, the lower the temp.

If you go and read the entire thread that I posted concerning cooling properties you will realize i was not talking apples to oranges. Thermodynamics is thermodynamics.

If the ground is at 70 degrees and your tank water is at 80, it will take a very long time to get to 70. The driving force of heat transfer is the temperature difference. as the hot and cold temperatures approach each other the amount of energy transfered from the hot to cold GREATLY decreases.

You may not believe what I am telling you, but trust me, I know what I am talking about. I have a masters degree in chemical engineering with way more heat transfer knowledge then I will ever need in my life.

When i was talking about increasing flow and decrease efficiency due to frictional losses what I mean is the heat caused by friction, not from lack of flow.

I promise you this, If I pump 1000 gallons of water through the coil and drop the temp .5 degrees that will remove heat then flowing 2 gallons that removes 10 degrees. Again, go back to the original thread and read about Q=U*A*Delta T. Q is the amount of heat removed per unit volumn. As Delta T approaches zero the amount of heat removed approaches 0.

If you have anymore questions feel free to let me know, but trust me on this on.

[AnnArborBuck]

How about using that black Siloflex instead of (I assume) white PVC?

It would depend on which one has the better heat transfer coefficient. Right off hand I have no idea which one is better, but if I can find my Perry's (chemical engineering handbook) I might be able to find out.

[T Sandman]

So, If you oversize the pump for this you could theoretically control the temp of the water coming back with a ball/gate valve. Is that correct? What about the sand being an insulator? could you heat soak it enough that it stops transferring heat?

[uscgbeachbum]

I'd listen to AnnArbor if I were you guys. I started off college going for chemical engineering. All the heat transfer, equations of equilibrium, blah blah blah, drove me nuts. The only profession that knows more about heat transfer is an actual thermodynamics professor.

[spazz]

i didnt meen to start an argument in the dynamix of geothermal cooling. ann arbor buck is right about one thing and that is the closer the tank water going through the grid gets to the temp of the sand, the longer it takes for the water to cool down. thanks to living in minnesota the groung is very cool. the sand temp at ground level where i put my grid stays at an even 60 deg or so. that makes the water cool down to 70 deg comming out of the grid. by putting 79.8 deg water through the grid i get a 10 deg drop in water at about 100 gph. now for someone in florida, you would have to take in to acount the size of the tank, the temp of the sand at the exact depth in the ground you would put the grid, and the amount of water going throught the grid. i dont have a PHD or any degree in science, but i just used a little common sence when i designed my grid. by using larger pipe for the grid (1 1/2") and smaller pipe(1/2") for the supply and return pipes, it causes the water to flow through the grid slower. and being in pipe thats got alot more sq. in. per foot makes the grid more efficient. i know my system is not the most practial system ever made, but it was just a way to show people that there is more than one way to skin a cat

these are just my thoughts, and i dont want any one to take my idea as the only way to do a geothermal cooling system. im sure there are alot better ways and maybe some that are cheeper too, but this is just my way and i thought i would share it so others could use it and maybe put big improvements to it for others to read. thanks, spazz

[vmichael]

Originally posted by cisco006
Do you guys think this will work in miami fl, how deep would I have to dig. I am in the process of setting up a 375g reef tank and I am very interested in implementing this setup.

Thanks for the idea spazz
cisco

If you dig 6 ft. your bound to hit water anyway.
As soon as you past the muck and the coral there is water.

It amazes me everyday that we cn actually live down here.

Sorry about the off topic!

[T Sandman]

I am not doubting his knowledge. I am asking a legitimate question. I think this grid is a super idea.

[AnnArborBuck]

So, If you oversize the pump for this you could theoretically control the temp of the water coming back with a ball/gate valve. Is that correct? What about the sand being an insulator? could you heat soak it enough that it stops transferring heat?

Ideally you want to have a pump and a grid system that just meets your needs. Sure you could oversize your pump and control it with a ball valve, but then you are paying for more pump then you really need. The pump is going to use so much juice no matter if it full open or you slow it down. So why use a pump that uses 100 watts if you can get by with a pump that uses 50 watts.

by using larger pipe for the grid (1 1/2") and smaller pipe(1/2") for the supply and return pipes, it causes the water to flow through the grid slower. and being in pipe thats got alot more sq. in. per foot makes the grid more efficient.

Actually, No. Slower flow through the pipe is not more efficient. Now in your case since your heat sink is free, efficiency doesn't matter as much. Your concern is having a large enough grid and making sure the temp difference between your cooled water and tank water is not too great. To keep the temp difference between the tank water and cooling water smaller, you actually want a higher constant flow. That way the exit temp of the water may only be 79 instead of 70. The thing to remember is that the amount of energy removed is the temp difference of the water in and out, and the amount of water being cooled (mass of water * Delta T * Specific Heat of water).

Like I said before, I bet one reason why your exit temp is so low is because when the pump turns off the water sits in the pipes getting real cold. When the pump turns back on this really cold water enters the tank. This cold water cools the tank really fast and the pump shuts off, only to let the water sit again. You might be better off running a smaller pump on a smaller grid, that way you don't have the extreme temp drop. Because it seems as if you have a great heat sink being in the north your system may actually be too big if you are getting alot of cycling of your pump.

[newbyreef]

Guys,

I've done a couple of successful cooling systems somewhat like this...However I used a stainless car radiator with an electic cooling fan for a car that was already mounted to the radiator. This worked, great..the water passes slowly through the radiators many very thin "fins" and as the fan blows through the fins, it cools quite effectively.

HTH
Dan

[cisco006]

vmichael,

Thanks for your reply, do you live in south florida? What is the temperature of the water down there?
Thanks
cisco

[Trackz]

Originally posted by newbyreef
Guys,

I've done a couple of successful cooling systems somewhat like this...However I used a stainless car radiator with an electic cooling fan for a car that was already mounted to the radiator. This worked, great..the water passes slowly through the radiators many very thin "fins" and as the fan blows through the fins, it cools quite effectively.

HTH
Dan

I've been doing this same setup for a long time on my water-cooled processors. Before I ever got into reefing, I was building coolers out of sheet PVC with a car heater core (very much like a radiator, but smaller) and a decent fan. I used a powerhead to push water through it to a copper block on the processor and back.

I've wondered for a long time why people didn't do this for reefs. It cools great and much cheaper than a chiller. Hell, the copper block was the most expensive part, and you wouldn't even use that for a reef.

[vmichael]

Originally posted by cisco006
vmichael,

Thanks for your reply, do you live in south florida? What is the temperature of the water down there?
Thanks
cisco

Yes! I'm from SOFLA
right now outside the temp in the pool is about 68-70
water coming out of the tap is about 70

I would imagine that this would go up about 15-20 during the summer months.

Current ambient temperature 07:00 AM is 61 the humidity at 93

Since most Floridians run their AC units 24/7 during the summer months; I've often thought of running coiled tubing inside the air handler and then back to the sump. This would bring the temp down. Also, in South Florida, because we have a critter problem (mostly ants) we tend not to keep to many food items in the cupboards instead we place a lot of stuff in the refrigerator and because of this we keep a 2nd refrigerator in the garage.
Since my sump is in the garage right behind the display wall. I thought about drilling a hole in the freezer and running tubing through that (coiled or wrapped around the walls of the freezer) and then back to the sump.

I would imagine that burying pipe or tubbing would also have some benefeit, however it would definitely have to be in the shade inside and not outside. It get so hot down here in the summer you can easily fry eggs.