Passively cooled fish tank computer

[FakeGamerGuy]

Steal them from that one noisy neighbor.

 

I am the noisy neighbor.

 

I didn't really care about the case rusting (looked kinda mad max style and was a bit of a "how far will it go?" experiment), I was worried about the lead solder they probably use to put the fins together failing and then having a huge contamination problem with my desktop loop.  So I switched to stacked plate heat exchangers that are stainless and copper brazed with 4000W heat capacity at 1C: 

HOHOHOHOHO

 

Where does your cold water come from?

[AnonymousGuy]

I am the noisy neighbor.

 

 

Where does your cold water come from?

Chillers and a 22 gallon reservoir.  Chillers do what radiators would do.  Cooling capacity there is about 1200W but it's not a direct translation because I have heat (cold) loss throughout the system.  22 gallons is a pretty large thermal buffer though.

[FakeGamerGuy]

Chillers and a 22 gallon reservoir.  Chillers do what radiators would do.  Cooling capacity there is about 1200W but it's not a direct translation because I have heat (cold) loss throughout the system.  22 gallons is a pretty large thermal buffer though.

 

I saw chillers while researching fish tank cooling but assumed that they were only good for keeping cold water cold like a refrigerator or freezer.

 

Also, one of my friends suggested I use glass rather than acrylic due to it's higher thermal efficiency. With the exact same tank size and water/ambient temperature delta the radiation through the walls alone jumped from 475W to something like 1.5KW. Looking into making a smaller, sealed system that would radiate ~500W at ~80 degrees.

[Stefan1024]

I saw chillers while researching fish tank cooling but assumed that they were only good for keeping cold water cold like a refrigerator or freezer.

 

Also, one of my friends suggested I use glass rather than acrylic due to it's higher thermal efficiency. With the exact same tank size and water/ambient temperature delta the radiation through the walls alone jumped from 475W to something like 1.5KW. Looking into making a smaller, sealed system that would radiate ~500W at ~80 degrees.

You still calculating only the thermal resistance of the material, do you? But on both sides it's not perfectly coupeled to the surroundings.

The water makes good contact, but the air can't chatch up und will lower the dissipated heat a lot.

 

Seriously I can't imagine you can get rid of 1,5 kW with only 0.6 m^2 surface area and no additional airfolw.

[FakeGamerGuy]

You still calculating only the thermal resistance of the material, do you? But on both sides it's not perfectly coupeled to the surroundings.

The water makes good contact, but the air can't chatch up und will lower the dissipated heat a lot.

 

Seriously I can't imagine you can get rid of 1,5 kW with only 0.6 m^2 surface area and no additional airfolw.

 

I haven't changed my calculations because I haven't found any documentation that backs up your claim. It's not that I don't believe you, but every resource I've found calculates heat transfer this way and specifically mentions that it's to calculate heat passing through a window to outside.

[Stefan1024]

I haven't changed my calculations because I haven't found any documentation that backs up your claim. It's not that I don't believe you, but every resource I've found calculates heat transfer this way and specifically mentions that it's to calculate heat passing through a window to outside.

 

I just looked it up by myself. You have to search for convection instead of heat transfer, otherwise you find next to nothing. Be aware that the thermal resistance caluculated is in addition to the one of the acrylic / glas. The total system will look like this:

from http://ocw.mit.edu/courses/aeronautics-and-astronautics/16-050-thermal-energy-fall-2002/lecture-notes/10_part3.pdf, page 26. It is a bit complicated.

 

This is the best page I found on my short search: http://www.engineeringtoolbox.com/convective-heat-transfer-d_430.html

They have an online calculator as well as a nice and short description.

Calculatie convection is a tough task, and is dependend on a lot of parameters. That's why it is often only simulated or measured. And that's why the coefficient of the natural convection is so wide spread:

• Free Convection - air, gases and dry vapors : 0.5 - 1000 (W/(m²K))
• Free Convection - water and liquids: 50 - 3000 (W/(m²K))

You are better off with a tall geometry, so a chimney effect can build up. The convection is mainly based on the air speed, so it helps a lot. Thats why my heat sinks are 400mm tall.

[Stefan1024]

@FakeGamerGuy

 

Any updates?

[Zyphron]

all this work for a Tank aquarium. system with a Limited heat load?  I dont understand. you have this Huge mess for Failure and it can be solved by Adding in a Simple 240 rad. and some fans running at next to nothing. 

where do you live? you could use a Aspen mat Swamp cooler (evap cooler https://en.wikipedia.org/wiki/Evaporative_cooler) if your outside humidity is Low. Hell you don't even need to cool the house. the Unit could sit outside and cool the out doors for all it matters and  it will bring in STUPID cold water.... easily below Ambient... (slowing the flow will stop Condensation as the Water can only flow so fast.)
your only issue here is Mineral buildup, you would need a Zinc Anode for sure and mineral Inhibitors. 
if you live in a Dry Cold Climate installing a swamp cooler in a 5 gallon bucket will give you cool water while adding Hummidity to the room meaning your heater can warm the Air more efficiently and the air will stay Hotter longer. you can Also keep the house cooler this way since Humid air feels warmer. 

I work on mine every summer. If its been running I have to let it sit for an hour to work on it.  Because its hot? no no..... because the water in the Unit is ice cold and the pads are wet and will get me soaked in freezing water. (im not kidding) so I wait for the 90F summer temps to warm it up. before I go work on it.

I mean you are doing a Satellite Cooling Tank and Plumbing.. might as well make a Unit that will do more then Just act as a heat tank.
another idea is you could also use that heat for other things. if this water is clean, more copper and more piping is a non issue. you could set a float up to dump in fresh water and use the warm stuff to heat the bath or Sinks?.... I mean... talk about efficient. Cooled PC, and Hot water all in one...... 

[FakeGamerGuy]

No updates; I've basically been on vacation for the past 2 weeks and now that I'm back at work I'm pulling 10-13 hour days.

 

Zyph, I live in Seattle. The idea was that I wanted to built a completely passively cooled system because it sounded like a fun project. Using a fish tank I could achieve practical, silent cooling while adding a visually pleasing item to my desk rather than cramming fans and radiators into or around my computer. I also live in a rented house and my room is on the second story so large active coolers that require additional plumbing aren't really an option, nor is recycling the hot water.

[Stefan1024]

No problem, take your time....

[FakeGamerGuy]

So I've received a lot of criticism with this build idea, primarily focusing around how much heat is actually going to be transferred from the water to the air. Another potential issue is the ability of a custom-made tank to safely hold water. In the interest of doing good science I've decided to buy a cheap tank off Craigslist, attempt to model the heat loss through all sources as best as I can, and see how that compares to filling up a real tank with hot water and monitoring the temperature over time. If I can translate heat loss per hour into watts and it agrees with my thermal model to a reasonable degree (hah, degree...) then I can move onto building a water cooling loop to test the stability and real-world heat-up times under actual use cases. If THAT works I would feel comfortable moving onto a more permanent custom tank solution.

 

I found a perfect test tank, I'm just waiting to see if it's still available.

[Stefan1024]

This sounds reasonabe and I'm looking foreward for results to compare with my system.

I hope you don't fool offended by our criticism.

[FakeGamerGuy]

Some criticism is backed up with concerns based on formulas and previous research, which I appreciate. Other criticism is not so well thought out and easy to dismiss.

[Andrew Sh]

Sounds like a brilliant project to be honest, will just be a huge resi though, can't put fish in a 40*c + tank

Try it!

[FakeGamerGuy]

Yeah, "fish tank" is more of a visual description than a practical one. Once it's done it will be disguised as a functioning fish tank sans fish. Besides, I doubt the fish would appreciate the biocide.

[FakeGamerGuy]

Picked up my test tank. Work is nuts lately but hopefully I don't have to work both days this weekend and I can do some testing. Pretty excited!

[ThomasD]

to avoid contamination you could submerse a 360mm rad in the water. have a water-water cooling system like a liquid cooled boat engine.

That is what I would recommend - a closed loop with the rad submerged in the tank, and some sort of impeller circulating water through the rad.

 

Either way, in order to get decent heat exhausted into the room you will need to have the tank open to the air, and that will lead to serious evaporation.  Unless you plan on buying a reverse osmosis system to provide you with a steady supply of de-ionized water it would be best to keep the cooling loop sealed.

 

None of which will be all that quiet...

 

But then you could use actual fish in the tank.  Tropical tanks usually need a heater anyway.

 

Edit: as an FYI:  Looking at tropical fish tank heaters on Amazon a 300 watt heater is rated for up to 75 gallons...

[FakeGamerGuy]

I was originally going to rely on surface radiation and evaporation to provide the bulk of my cooling but later realized that, using a glass tank rather than an acrylic one, the glass will radiate enough heat and allow me to run the system sealed.

[FakeGamerGuy]

Finally have some time to test my tank. The 20" x 10.25" tank is fulled with 10" of water. The top of the tank is "sealed" with 2 layers of air-gapped cardboard with bottles of mystery liquid holding it down. I can do the calculations on gallons of water and surface area later. As of 9:15 pm PST the water is 105F. Ambient temperature is about 78F. I'll be taking pictures of the thermometer every 30-60 minutes to monitor heatloss so I can determine how many watts this thing can cool while "sealed".

[Stefan1024]

You do proper scientific work.

Keep going....

[FakeGamerGuy]

From 104F to 95F took 2.5 hours for a cool rate of 3.6F/hr. Ambient temperature was 75-78F with minimal airflow. Quick calculation shows this tank radiated 70W.

From 95F to 81F took 8.25 hours for a cool rate of 1.7F/hr. Note: This was overnight but also in the garage. Air temperature likely dropped but shouldn't have been by much. Quick calculation shows this tank radiated about 48W.

 

I haven't measured the thickness of the glass and I haven't measured temperature drops with the tank open but I'm not particularly excited right now.

[Stefan1024]

From 104F to 95F took 2.5 hours for a cool rate of 3.6F/hr. Ambient temperature was 75-78F with minimal airflow. Quick calculation shows this tank radiated 70W.

From 95F to 81F took 8.25 hours for a cool rate of 1.7F/hr. Note: This was overnight but also in the garage. Air temperature likely dropped but shouldn't have been by much. Quick calculation shows this tank radiated about 48W.

 

I haven't measured the thickness of the glass and I haven't measured temperature drops with the tank open but I'm not particularly excited right now.

I redid your math and it is correct.

 

And I also expected a result in this range. But you can heat up to water up to ~60°C and your power dissipianted double to 150 watts.

Also you can add a passive radiator to the loop to get rid of an other 100 watts. Also you still have the big tank to "store" the energy during heavy usage. While browing and office stuf your PC only uses around 60-100 watts. So the tank can actually cool a bit dispite of using the PC.

 

Or you could also add a standart active rad to the loop. You only have start the fans after 3-5 hours of gaming the tank is hot.

[AntVenom]

I have just recently finished a project that uses a big bucket and a thick quad radiator w/ 4 fans blowing air in my basement, and I can tell you that there's a decent chance that, if the radiator is big enough, this will likely work fine, even passive. Not 100% sure though. The ONLY thing having a big reservoir such as a fish tank will do is increase the amount of time the system will take to reach an equilibrium, but that equilibrium will likely not be TOO much different than if you were using a small reservoir. But hey, go for it if that's what you want!

[FakeGamerGuy]

I have just recently finished a project that uses a big bucket and a thick quad radiator w/ 4 fans blowing air in my basement, and I can tell you that there's a decent chance that, if the radiator is big enough, this will likely work fine, even passive. Not 100% sure though. The ONLY thing having a big reservoir such as a fish tank will do is increase the amount of time the system will take to reach an equilibrium, but that equilibrium will likely not be TOO much different than if you were using a small reservoir. But hey, go for it if that's what you want!

The idea wasn't to use a large amount of water to cool but rather use a large surface area to radiate heat. I still need to do more testing but at this point it's not looking promising. I could add a passive radiator but then I have to start considering if it's worth the build or if I should go back to quiet air cooling. At this point, even if I can radiate an extra 100w from the radiator I'm still only able to cool the CPU and not my dual video cards.

[Stefan1024]

Once finished, I will cool 500 watts passive. It's possible, but I thougth your idea was to use the big tank, so you deal with more heat than you dissipiate at a given time....