Peltier-cooled hatchery

[RonWill]

Folks,
I'm frustrated by the lack-lustre hatching/fry survival results for the A. amoenum 'Sakbayeme' that Bill sent. He mentioned hatchwater was between 22-24ºC

A freezer-unit to chill the whole tank, like what Kwek Leong did, isn't viable in my multi-tier tank setup but I've been considering peltier cooling for some customized tanks (14"L x 8"W x 10"H). It's small enough for these Thermoelectric coolers (TECs) to be efficient.

Besides better hatch rate, it's a chance for me to maintain diaptrons and cool water killies!

Here's the crunch... I'm electronics-handicapped but dearly want to rig one up, together with no-frill temperature controller, and will need guidance in this area. What I do will be compiled into an article to be shared here.

Issues I'm trying to figure out...
What's the cooling unit/term used for peltiers? (air conditioners are in BTU).
What's an appropriate peltier size for a 1-foot tank?
I understand peltiers to be powered by 12v DC. Can I use a PC's PSU for that and to drive the cooling fan?
For heatsinks, are there circular-finned types? I've seen one before but forgotten where. [normal HS are either square or rectangular footprint]

Some of the links I've went through...
IceProbe Aquarium / Water Chiller by Coolworks
CoolWorks Microchiller
How Does the IceProbe Small Aquarium Chiller Work

Peltier chillers
http://www.inbthermoelectric.com/thermo.html
http://www.peltier-info.com/products.html
http://www.peltier-info.com/tims.html
http://www.kryothermusa.com/
"Peltier Radiator (Chiller) Formulas"

Peltier Thermoelectric Cooler (TEC) 72 Watt
"Thermoelectric coolers (TECs) are solid state heat pumps that utilize
the Peltier effect. During operation, DC current flows through the TEC
causing heat to be transferred from one side of the TEC to the other,
creating a cold and hot side..."

I'm mad enough to even consider radiators, especially the Black Ice Extreme Radiator discussed in a overclocking forum. Suggestions, proposal... help?

[whuntley]

Ron,

Let me add to your links:

http://cgi.ebay.com/ws/eBayISAPI.dll...837530031&rd=1

This is a 62mm-square, 270-350-Watt unit on ebay. $6 extra to ship outside the US, but distributorships available!

The pic shows the basic unit, a thin plate with red and black wires attached.

To this you will want to add an aquarium thermostat (usually available in a burned-out heater) and some heat spreaders (fins) for outside and inside the tank. Outside, you may want to add a fan to improve thermal transfer with the air. Inside may need coating to be chemically neutral -- a good black hard anodize should be OK. I'd make the fins considerable larger than 62mm square to reduce sharp thermal gradients that might crack glass (i.e, provide a transition region around the edges).

The ideal system would use the thermostat to drive a reversing relay so that the unit was either heating or cooling depending on what the tank needs. I would build a small circuit to prevent rapid switching back and forth -- a little hysteresis is often a blessing for saving contacts.

Caution! This much Watts may be a lot more chilling than a little freezer or bar fridge. IDK. Likewise, the thermostat or relay must handle 30+ Amps. That is in the neighborhood of the headlamp or starter relay from a car. It may be too much for the usual computer power supply, too. Check the 12V current ratings.

Wright

[stormhawk]

Ron, came across this Aussie site by a teacher I think, who was trying to answer a question from a student about the use of Peltier effect in refridgerators. Here's the discussion:

http://www.abc.net.au/science/k2/homework/s95611.htm

From the information regarding Peltier's discovery, I came up with this simplified diagram.



I was wondering which junction was the one that got hot. Anyone knows? It would be feasible to try this out on an experimental basis using two kinds of thermoelectric wire. List of thermoelectric materials in the link below:
http://www.xyroth-enterprises.co.uk/thermser.htm

Other links to Peltier Effect:

http://www.uni-konstanz.de/physik/Ja...wer/node2.html
http://www.digit-life.com/articles/peltiercoolers/
http://www.xyroth-enterprises.co.uk/peltier.htm

[stormhawk]

Found the Japanese site I was referring to in an earlier post on another thread. DIY article on making a Peltier cooler.

http://babelfish.altavista.com/babel...%2fkuuraa.html

[whuntley]

Hi all,

I have found a number of useful sites, many with good precautions and warnings.

One is that condensation must be studiously avoided, so painting, enclosing, running at the right levels must all be considered. Water will short out the semiconductor devices if allowed to condense on the cold junctions.

There's a really neat graph at:

http://www.tellurex.com/csweet.html

that describes a so-called "sweet spot" for thermoelectric cooling. From a cursory look, it may be that we can't avoid condensation in SG at the temperature differences where the device would like to work. A solution would have to be found.

Check out that whole site as well as:

http://www.melcor.com/faq.html ,

http://www.marlow.com/TechnicalInfo/...tions_faqs.htm and

http://www.ferrotec.com/usa/thermoel.../ref/index.htm

Efficiency seems to lie somewhere in the 25-40% area, which is a bit lower than liquid-based refrigerant systems. Maybe half?

Wright

[whuntley]

Some useful conversion factors:

Look at well-designed Peltier coolers as about 1/3 as efficient as most aquarium heaters (that are nearly 100% efficient).

A heater to give a certain temperature change may be selected (for us ice folks) as big enough to overcome room chilling but small enough to not boil the fish if the contacts stick in the thermostat. For 5G tanks, I prefer 25W heaters as meeting those criteria in human-comfortable surroundings.

One might then guess a similar ability to overcome room effects might be possible with a 75W Peltier chiller, or a bit over 20W/G on other similar shaped tanks. Insulating bottom, sides, and back could drop that number quite a bit, I suspect. We are talking killies, so covered tops are, of course, mandatory. Right?

Basically, 1 Watt is about .85 KCal. so could chill a liter about 0.85 degrees Celcius. The significant factor is to balance room heating against the cooling to get the desired end effect at a price we can afford.

I'd assume I could design a far more sensible heat-sink arrangement than the Japanese one, so would probably start with the 20W/G rule of thumb until I had one built and tested. YMMV.

Wright

[RonWill]

Folks,

I've a couple of DIY items on the drawing board and the following images are for a peltier-based cooler (prior to fan installation) that's been sitting on the shelf



When I followed Jian Yang's URL link to the Japanese hobbyist's page, I realized that my thoughts are not new. IMHO, the short-coming of that design is that the aluminum support plate is too wide and too far from water, allowing the alum sheet to act as a primary HS (heatsink) and absorbing the 'chill', dissipating before it hits the water.

The 2nd unit was an improvement with a larger and more efficient HS. However, with the additional weight, placement at the tank's corner doesn't allow the unit to be efficient. IMHO.

I loath the idea of electricity near water and the 3rd pic, an adaptation from Coolworks, allow connection terminals to be further away from water.

The online auction site has many peltiers available, in different sizes and wattage. Question is, which model/rating is better suited for a 14"L x 8"W x 10"H tank? Temperature-wise, how much can I expect the device to 'pull down' and should I over-engineer, ie. buying a bigger peltier than I need? Numbers in the 'formula link' doesn't ring a bell either.

Presently, electrical supply to the multi-tier is from one powerpoint that's running a Hi-Blow 40, Eheim 1060 pump and 2 Eheim hang-on filters. I wouldn't want to overload and trip something

With fan installation, the acrylic cover for the 1st 2 illustrations (Left & Centre) need modifications for fan clearance, but still not very killie-safe

One more thing... is there something I can use that's thermo-conductive and bonds permanently, at junctions between the peltier 'hot-side'/HS and peltier 'cold-side'/HS?

Pennies for your thoughts?

[RonWill]

Found one more link;
http://fins.actwin.com/nanf/month.9907/msg00207.html

[RonWill]

In response to a few PM's asking... "How to adjust height of lower HS after installation of fan", here's the solution;


"How do I hold the assembly together"
Hopefully, there's a epoxy-like material that's thermal-conductive, ie. to transfer heat or 'chill', and has a permanent bond like those tiny HS's found on graphic cards and motherboards.

For the 'through-panel' assembly, the 'cold-side' of the peltier is bonded to a plate of aluminum or brass, which is wide enough to overlap the size of opening in the tank. Silicon is being considered;


I think for this contraption to be efficient, there must be circulation passing through the HS's fins, provided either by airstone or output from powerhead.

Another questioned, "I thought you're making a chill hatchery? Looks like you're going for a planted tank!"
Ok... for a start, the 'through-panel' design is safer (I can be a real klutz with dangling wires) and tested on a medium acrylic tank or critter-keeper. It's still a hatchery, that I float the egg-trays in it rather than having a peltier cooler for each egg-tray. A bigger volume of water should be more temp-stable.

Another possiblity is just to shove the egg-trays into the fridge but with 3 crumb-snatchers, fridge space is a premium... until I get my own 'fish-fridge' :wink: