DIY Chiller Project

[inzzane]

Anybody had experience creating a low-cost, yet professional appearance, DIY chiller?
I'm currently considering use of a peltier unit, software controlled to reduce power consumption. I'm thinking of a more professional appearance, like something we all would like to buy, and would not want to tuck it in the back corner like sooo many of our other items.
My end goal is to be able to produce a low to no-maintenance, compact, reliable digital temp controlled chiller for a 100 gal tank (15 degree reduction at room temp) with no more than a 10x10x10" footprint, unless of course there's already one being sold.:confused2:

Over the next couple months I'm hoping to finish one... but with the designs I keep running into one issue... metal contacting tank water.

My question is with transferring the cold temperature to a tank... (using metals great thermal properties.) If metal needs to be in contact with the tank water, what metal is preferred? I've read that copper (and alloys (brass, bronze, lead, zinc)should never be used, but I'm not sure of others, like gold plating... possibly something else?


Also, is there some sort of coating that could be placed over metal to prevent it from contacting tank water yet has high thermal transfer characteristics? I've heard of some epoxy's, but would worry about flaking off over time.
I have a few alternative ideas, but hoping to keep a small amount of metal in contact to improve performance. I have no ideas how the commercial ones operate internally. Peltier? Plastic radiators? Freon?

Any ideas or knowledge you have would be appreciated!

 

[skriz]

The best metal for sw use is titanium.

 

[lt_tweak]

peltier junctions?! following this oh and gold/titanium heatsinks dont sound cheap

 

[inzzane]

@Skriz
Thanks! I've been looking into Titanium as well, because of what i read here on it's tolerance with saltwater but just now (after your post) found it's only problem is that its thermal conductivity appears to be extremely low, rated around a 13, where gold is 183 and copper is 200-380... although this is conflicting because I have seen titanium heat exchangers. I may have to look into titanium some more tomorrow, as i don't see why anybody would make them if it was rated so low.

@lt_tweak
lol, if I have to go with one of these, I'm sure it will be a VERY THIN layer
Although it won't actually be a heat sink, just a transfer plate, so it would only need to be a small flat area about 3"-6" square

 

[myvwrock]

 

[inzzane]

@myVWrock
Thanks for the link. It is similar in some aspects of the cooling side to what i had planned, contains a lot of good information and brings a lot more design ideas.

The original poster had a similar idea as mine but didn't go with it due to his small tank size.
"At first I was going to make a titanium "heatsink" for increased contact area (it was going to be a plate with lots of little tabs cut out and folded up) but ....."
Considering this threads information though, I may just go with 4 in-line peltiers @ 120 watts using similar methods and see what the data shows.

 

[rbredding]

I'd recommend scrapping the temperature controller for a unit that is completely uncontrolled.. (plug it in, chiller turns on)

With the newly prevalent use of sophisticated controllers, there's really no need for you to add a controller to your setup.. people can buy a RKL for just over $100 that is programmable/expandable to handle other aquarium equipment/functions..

these types of controllers were not really available at a price point that made them affordable to the "typical hobbyist" 4 or 5 years ago.. (unlike today).



well.. THAT.. and it's just one less component that you need to worry about malfunctioning..

 

[lt_tweak]

i didnt mean to use as a traditional heatsink. I meant moreso as a heat exchanger, more surface area on a finned heatsink than a block of Gold or Titanium, not to mention less materials, so probably cheaper. I dont know i never bought a Titanium heatsink

 

[mojo]

Inzzane;663194 wrote: @Skriz
Thanks! I've been looking into Titanium as well, because of what i read here on it's tolerance with saltwater but just now (after your post) found it's only problem is that its thermal conductivity appears to be extremely low, rated around a 13, where gold is 183 and copper is 200-380... although this is conflicting because I have seen titanium heat exchangers. I may have to look into titanium some more tomorrow, as i don't see why anybody would make them if it was rated so low.

Because copper will destroy your reef tank, and gold is... expensive. Platinum would be better than Ti, but even more expensive than gold.

Some vendors use ceramic lined copper, but I'd suspect that ceramics thermal conductivity rivals that of titanium, and if it ever chips, you have reef water in contact with copper.

Either way, titanium or high grade stainless steel is about your only choice.

 

[lilrobb]

Thermal conductivity:

 

[enderg60]

you need the equivalent of a 1/4 - 1/2hp chiller to do what you want.

1hp = 1 ton = 12,000 btu = 3516w

Taking into account efficiency you basically need 1000-1500w of TEC cooling power...if not more.

ill be amazed if you can part that together for less then a regular used chiller.

 

[jef4y]

I replaced the cold-water-bladder on the fridge in my old house and instead of a bladder, they used 50' of plastic tubing coiled up. It chilled quite well.

Personally if I were to do a DIY on a smaller tank I'd run 100' of coiled tubing into a bucket of water that sat in a mini-fridge and was powered by a small pump (aqualifter maybe) and regulated by my apex.

 

[ichthyoid]

JeF4y;663361 wrote: I replaced the cold-water-bladder on the fridge in my old house and instead of a bladder, they used 50' of plastic tubing coiled up. It chilled quite well.

Personally if I were to do a DIY on a smaller tank I'd run 100' of coiled tubing into a bucket of water that sat in a mini-fridge and was powered by a small pump (aqualifter maybe) and regulated by my apex.

+1

There was a DIY chiller made from a dorm type fridge and plastic tubing, that I read about a few years back. Add a Ranco temp. controller and you are in business.

If all you need is a nano chiller, Costco has a Peltier cooler right now for $100. Of course, you can buy one for about the same price, or about $150 with controller.

 

[spiderman097]

most diy chillers involve a minifridge

 

[grouper therapy]

They may cool a 25 gallon or less nano with a light heat load. Dorm refrigerators only remove a very small amount of heat per hour. The compressors they use are not designed to run an extended duty cycle with warm refrigerant. Coils of tubing in buckets of water does not increase the capacity of a dorm refrigerator .

 

[morganatlanta]

This could be fun for an experiment, but you're not likely to come out ahead cost wise.

I'd think about doing an aluminum box of some sort and mount the cooling units on the outside, with the cold side to the box, and mount heatsinks on the hot side of the cooling units. Fill the box with water. Coil some plastic tubing in the water and run your tank water through that. The tank water only ever touches the plastic tubing, not the water in the box or the aluminum. You'll need a fan to move air across the heatsinks take the heat away. 160 Watt cooling units are around $20 each, so you probably need 10.

 

[inzzane]

First, Thanks for all the replies... it's a lot of good information, and I may be re-evaluating some steps. I'm sure glad i asked here before jumping into this. :thumbs:
Now I do still plan to continue, but I may be building one of each (peltier, and compressor) before I'm through.
I'm really trying to stay away from the dorm fridge styles, as I'll be using these on my tanks in my living room (36G+20G sump, 72G+30G sump, and soon a 55G with 20G sump). 3 mini-fridges would be pout of place, unless there were filled with beer :up:

@Rbredding
"I'd recommend scrapping the temperature controller for a unit that is completely uncontrolled.. (plug it in, chiller turns on)"
Thanks, I was contemplating this myself just for cost reasons. Now that I see the prices on an RKL I'll probably skip over it.

@mojo
Thanks. I think I'm going to go with creating a copper reservoir off the peltier. This will maximize thermal transfer to a low temp fluid section, and my run a stainless steel pipe coiled tightly inside. This will maximize exposure, prevent contamination, and no copper will come in contact with tank water.

@EnderG60
"you need the equivalent of a 1/4 - 1/2hp chiller to do what you want.
1hp = 1 ton = 12,000 btu = 3516w
Taking into account efficiency you basically need 1000-1500w of TEC cooling power...if not more.
ill be amazed if you can part that together for less then a regular used chiller."
I see that the power consumption is a definite issue, and one that will probably end up requiring me to use a compressor style, but I'm curious to know if multiple passes from a tube would reduce the requirement. (Worst case scenario I can use it on my smaller tank).
Meaning if you have a 2 inch square, 1 watt peltier, and you create a block of ice with it. I'm thinking if 1 watt (for example) will produce a block of ice, then If I pass through a tube through that same block 4 times, I would require less total watts in the long run. Really, I guess I'm left just questioning the efficiency possibilities.
I'm hoping I could get away with a max of 550watts, or at least determine how well it performs.


@MorganAtlanta
"I'd think about doing an aluminum box of some sort and mount the cooling units on the outside, with the cold side to the box, and mount heatsinks on the hot side of the cooling units. Fill the box with water. Coil some plastic tubing in the water and run your tank water through that. The tank water only ever touches the plastic tubing, not the water in the box or the aluminum. You'll need a fan to move air across the heatsinks take the heat away. 160 Watt cooling units are around $20 each, so you probably need 10."

Thats pretty much the plan so far :up: creating a liquid reservoir that will be below freezing that I can coil a tank hose (or possibly a stainless steel tubing) through.

Again, Thanks to all. I'm obviously still planning and this site helps. I should be picking up parts over the next couple weeks and will do a build thread on whatever the final decision is.

Edit:

MorganAtlanta;663455 wrote: This could be fun for an experiment, but you're not likely to come out ahead cost wise......
160 Watt cooling units are around $20 each, so you probably need 10.

If the first test works out, I'll be buying in bulk so I should be able to keep the costs per unit down. If I can get 1 to work for a good overall price, then I'll be building multiples of it. :thumbs:

 

[enderg60]

Inzzane;663524 wrote:
I see that the power consumption is a definite issue, and one that will probably end up requiring me to use a compressor style, but I'm curious to know if multiple passes from a tube would reduce the requirement. (Worst case scenario I can use it on my smaller tank).
Meaning if you have a 2 inch square, 1 watt peltier, and you create a block of ice with it. I'm thinking if 1 watt (for example) will produce a block of ice, then If I pass through a tube through that same block 4 times, I would require less total watts in the long run. Really, I guess I'm left just questioning the efficiency possibilities.
I'm hoping I could get away with a max of 550watts, or at least determine how well it performs.

No and no, a watt is a watt, power in is power out. if anything you will need more power due to inefficient heat transfer. multiple passes only works for exposure. you can get the same effect by slowing down the water.

at 550 watts you will be lucky to get a 3-5 degree drop, which granted is more then most people actually need but not what you said you wanted.

 

[jef4y]

grouper therapy;663424 wrote: They may cool a 25 gallon or less nano with a light heat load. Dorm refrigerators only remove a very small amount of heat per hour. The compressors they use are not designed to run an extended duty cycle with warm refrigerant. Coils of tubing in buckets of water does not increase the capacity of a dorm refrigerator .

Very true.. and at the end of the day, while it might be fun to play around with assuming you have the materials needed; if you had to buy all the stuff you'd be better off with a purpose built chiller.

 

[inzzane]

EnderG60;663542 wrote: No and no, a watt is a watt, power in is power out. if anything you will need more power due to inefficient heat transfer. multiple passes only works for exposure. you can get the same effect by slowing down the water.

Dang engineering logic strikes again! Always bringing reality back to the table :doh: I knew I was overlooking/thinking something simple.

EnderG60;663542 wrote: at 550 watts you will be lucky to get a 3-5 degree drop, which granted is more then most people actually need but not what you said you wanted.

Thanks, I spent the morning running the numbers based off of the ice probes results i could find, and agree that the results on a 100 gallon would be approximately 5 degrees with high efficiency. My end goal was a 15 degree drop in tank temp</em> (One of my tanks was 12 degrees hotter than normal when my HVAC went out a couple weeks ago) however it appears I may not be able to reach it with the peltier setup:shout:... That is without raising the electric bill quite a bit compared to other options.
Unfortunately I know this project is really borderlined on determining the max capabilities of a peltier system, and the affordability of a compressor system, primarily due to the 100 gallon goal.

Soooo... I have to go with both styles and create a new plan and goal. My priority is my 50 gallon tank, as it's the only one that is more than a few degrees out of whack. (peaks at 78 during hot days,(currently set at 74 but would prefer to set it around 69)

Under normal conditions, I am hoping to drop my 50 gallon test tank by around 8 degrees to get it where I want it.
I see that the ice probe claims it is capable of dropping 1-2 degrees in a 40 gallon system, using a 53 watt peltier, so there still appears to be some hope reaching this goal using the peltier for this tank if i use around 300 watts... if my calculations are right. (correct me if you see an issue).
I'll likely start ordering parts this weekend.

I think I'll also run through the plans this weekend for a compressor style system to determine the price i could build one at for a 100 gallon system...