Guidance Requested: Off-Grid, Peltier-Cooled, Cat House

[emua]

Apologies beforehand if I'm posting this in the wrong section, or if this has been covered before, but I'm new, and equally confused from the litany of references that I've found online.

I'm conceptualizing a new project, and the ultimate goal is to keep my outdoor cat cool during the hot summers (up to 110° F), and I'd love to begin learning about PV panel applications. I've decided to go with a Peltier-cooled 70+ quart ice chest (with a small pet door), and have begun ordering the parts:

• (1) TEC1-12710 peltier
• (2) 90mm² heatsinks
• (2) 12v Titan IP55 92mm case fans
• various fittings, thermal compound, insulating foam, caulking, etc.

Okay. So, now my foot's in the door with the basic cooling mechanism. But to provide power, I'd like a recommendation on the size of PV panel that I'd need, and whether the system will need any other components (e.g. controller, inverter, voltage regulator, etc.). Here are the power requirements of the fans and the Peltier (i.e. thermoelectric cooler):

	Hot Fan	Cold Fan	Peltier
Rated Voltage	12	12	12
Rated Current	0.3	0.3	10
Power Consumption (watts)	3.6	3.6	154

I've read that Peltiers generally perform more efficiently when the current is scaled back from its max-rated wattage, so I'm wondering if a 100w/10a PV panel will do the trick.

If a 100w/10a PV panel isn't sufficient, please let me know why. And, if the 100w/10a PV panel is sufficient, and there are other components that are needed, please let me know which.

I know that thermoelectic coolers are inefficient, but the alternatives - such as a 5,000 BTU window AC unit - are overkill (in terms of dBA) and just as costly.

Please help!! I'm just a newbie - not altogether an idiot - but I'd like to do this right the first time. And please, no haters on why my cat is outside. There's a long story behind that, and it's justified. Also, she's quite happy to be outdoors amongst the plants in our garden.

Thanks to all in advance!

[Logan5]

This thing will eat batteries and not save any money. I can give you some advice, make the area to be kept cool, top entry, with a roof and walls designed to prevent the cool air from getting stirred out of the cool box inside the house. You should get this working with an AC outlet and power supply before you switch over to battery solar. Iknowing this is going to cost 10 to 20 times more than powering it from PoCo. with annual battery replacement. Peltiers make very little cool and it must be kept in an insulated container it pools like water so any bottom leak you lose all of your cool air. and maybe still not actually cool the space. Plug it in and save yourself hundreds of dollars.

[sensij]

Before you get too far into the solar aspect of this, you might consider proving that it works using grid power first. How experienced are you with TEC's? A 90 mm2 heatsink for a 160 mm2 TEC already looks like a bad combination, and a single TEC will do very little to cool a 70 quart chest.

[emua]

Thanks for the response Logan. As I get further into the world of PV systems, I'd love to learn more about batteries, but I didn't envision this contraption to use batteries. I imagined that it'd be a simple on, as the sun comes up, and then off as the sun goes down. On the other hand, your feedback of dealing with condensation and heat intrusion is well taken.

I imagined leaving the cooling chest's drain port open to release any condensation. These chests nowadays are all graded so that fluids flow to and out of a point in the floor. Other than that, the pet door that I ordered has a lightly insulated flap to prevent too much leakage. Also, the argument that I made to my wife is such that this Peltier can supposedly lower ambient temperatures by 60° F. BUT! IF this contraption fails to satisfy...hey...I got a 100 watt monocrystalline panel out of this little experiment, which is then a compelling basis to purchase more panels to mount on top of a plastic storage shed. By my spatial calculations, I could add 700 watts more - for a total of 800 watts - and then add in the requisite charge controller, deep cycle batteries, and inverter(s) for an off-grid system.

No idea what I'd use such a system [on my shed, in my backyard] for...but the possibilities are broad.

[emua]

Before you get too far into the solar aspect of this, you might consider proving that it works using grid power first. How experienced are you with TEC's? A 90 mm2 heatsink for a 160 mm2 TEC already looks like a bad combination, and a single TEC will do very little to cool a 70 quart chest.

Howdy Sensij,

I've also been wondering how to prove it using grid power first. As far as I can tell, I might be able to use a transformer for low-voltage lighting to "bench test" the concept. A 200 watt transformer for a 161.2 watt system will run between $50-$90, which should be fine. My only concern is, once I see that it can run on grid power, I'd be more inclined to stay with grid power, and not explore the potential of PV power.

I've only installed one computer system using a TEC during my college days (two decades ago), so not very experienced at all, but most commercially available TECs - including this 12710 model - are 40 mm², which will be fine for 90 mm² heatsinks sandwiched by 92 mm² case fans. I'm hoping that the size of the case fans will produce enough airflow within the confines of the ice chest to drop the temperature by hopefully 30° F from ambient.

So going back to the low-voltage lighting transformer, is it acceptable to splice 0.3 amp and 10 amp wires together? Or should they be on separate circuits?

Thanks!

[sensij]

I've only installed one computer system using a TEC during my college days (two decades ago), so not very experienced at all, but most commercially available TECs - including this 12710 model - are 40 mm², which will be fine for 90 mm² heatsinks sandwiched by 92 mm² case fans. I'm hoping that the size of the case fans will produce enough airflow within the confines of the ice chest to drop the temperature by hopefully 30° F from ambient.

The TEC you selected is 40 mm per side, or 160 mm2 (40 x 40).

http://www.thermonamic.com/TEC1-12710-English.pdf

Although they can achieve 60 deg delta T surface temperatures, they need to be coupled extremely well to the surfaces from which (and into which) you want to conduct heat. A thermoelectric wine fridge will have several mounted with the cold side attached to the wall of the fridge with a thermally conductive paste, and the hot side with a set of fins attached the same way and plenty of airflow.

Edit... looks like your fan and heatsink measurements might also be of a single side, 90 mm, not 90 mm2. Best bet with the oversized heat sink is to spring load the mounting so you don't deflect it and jeopardize the thermal contact between the hot side of the TEC and the fins. The cold side of the TEC still needs to be held in intimate contact with the surface you want to cool.

2nd edit: If the sandwich is made so that the hot and cold side each have a heat sink with fans, and they are otherwise insulated from each other, that could work. Then the question is whether or not ~100 W of flux is sufficient to cool the chest. The cat would probably be responsible for about 10 W of heat load, but the seal on the door could be a problem.

Last edit: You might want to consider separate PWM type controller to regulate the voltage to the TEC. Hitting it with a steady 12 V can cook it, if the heat sinks + fans aren't moving enough heat.

[sensij]

So going back to the low-voltage lighting transformer, is it acceptable to splice 0.3 amp and 10 amp wires together? Or should they be on separate circuits?

You can supply all the loads from a single transformer. Fusing each load is a good idea in anticipation of the battery.

[emua]

The TEC you selected is 40 mm per side, or 160 mm2 (40 x 40).

http://www.thermonamic.com/TEC1-12710-English.pdf

Although they can achieve 60 deg delta T surface temperatures, they need to be coupled extremely well to the surfaces from which (and into which) you want to conduct heat. A thermoelectric wine fridge will have several mounted with the cold side attached to the wall of the fridge with a thermally conductive paste, and the hot side with a set of fins attached the same way and plenty of airflow.

Edit... looks like your fan and heatsink measurements might also be of a single side, 90 mm, not 90 mm2. Best bet with the oversized heat sink is to spring load the mounting so you don't deflect it and jeopardize the thermal contact between the hot side of the TEC and the fins. The cold side of the TEC still needs to be held in intimate contact with the surface you want to cool.

2nd edit: If the sandwich is made so that the hot and cold side each have a heat sink with fans, and they are otherwise insulated from each other, that could work. Then the question is whether or not ~100 W of flux is sufficient to cool the chest. The cat would probably be responsible for about 10 W of heat load, but the seal on the door could be a problem.

Ahhh! I see what you mean. Apologies: my math-speak is very rusty since it's been ages since I've taken any math courses.

Thank you very much for your responses - I'm glad to see that you're getting a sense of where I'm trying to go with this contraption, and my education on PV.

Correct: I would have 90 mm heat sinks on both the hot and the cold sides with Arctic Silver Céramique, and then 92 mm IP55 case fans on top of the heat sinks. I'd spray or adhere insulating foam in the 3.2 mm space between the heat sinks to mitigate a degree of temperature bleeding. I suppose that's the appropriate way to describe what I'm trying to do.

I'll try to find a spring loaded clamp, but I had originally thought that using long zinc-plated machine screws, washers, and bolts on the four corners of the fans might do the trick. And the reason why I thought this was because my wife and I came across a used thermoelectric cooler (albeit, a very small unit, probably a 5a Peltier), which we purchased and disassembled in order to get a sense of its general construction. And the manufacturer had used long machine screws to "clamp" the device together.

I agree that the pet door could be a problem, but I'll very likely test the concept using a low-voltage lighting transformer as you have suggested. If I can even achieve a 30 deg delta change within the chest, that would be a win. Sitting in 80° F is sure better than sitting in 110° F!

I did speak to a buddy who knows a little bit about electrical wiring, and he didn't think that I'd blow any of the components if I spliced all of the red wires together and all of the black wires together (from components using different amperages), and plugging them into the transformer.

[emua]

Last edit: You might want to consider separate PWM type controller to regulate the voltage to the TEC. Hitting it with a steady 12 V can cook it, if the heat sinks + fans aren't moving enough heat.

That is a GREAT suggestion. I was wondering what such a device was called - I knew that there had to be "voltage regulator" device out there. I'll look into it to see if it'll be useful to "throttle" the power a bit to prevent blowouts. Thanks Sensij!

[sensij]

That is a GREAT suggestion. I was wondering what such a device was called - I knew that there had to be "voltage regulator" device out there. I'll look into it to see if it'll be useful to "throttle" the power a bit to prevent blowouts. Thanks Sensij!

You could use a PID temp controller operating an SSR rated for the 10 A draw of the TEC. Something like this combination:

http://www.auberins.com/index.php?ma...&products_id=8
+
http://www.auberins.com/index.php?ma...&products_id=3

Might take some trial and error to tune the PID to keep the TEC from overheating, but I think this would be on the right track. Can probably find one that runs off 12 V if what I linked does not.

[sdold]

Before you get too far into the solar aspect of this, you might consider proving that it works using grid power first.

And to make sure the cat actually uses it

[DanKegel]

You could simulate the cat with a six-pack of beer if need be

[Sunking]

So let's see who would buy it.

It cost more than conventional cooling to build and operate.
It is less efficient. and way less effective than conventional cooling.

So who would buy something that cost more initially, cost more to operate (more energy) and does a poor job?

Not me.

[Mike90250]

Peltier coolers are VERY inefficient. But you know that already, So, my suggestion is:
1 PV panel for the 3 fans, 2 large, 12VDC computer cooler fans for the hot side, 1 small fan for the cold side, or even just a large aluminum plate for the cold side.
a 12V PV array to power the TEM Wattage = 200 - 300W of PV at about 18V
TEM = 16V 10A = 160 w of waste heat to fan away PLUS the heat you are pumping from the cold side (another 100w)
So you need heat sink and fan, capable of disposing 260 watts of heat, with only a few degrees of thermal rise.

The other problem with TEM's is attaching FLAT thermal paths to them, (hot & cold side) Few things you buy off the shelf are going to be FLAT, Most of the machined computer heat sinks are FLAT. Flatness = good thermal path, If there is a slight curve, or some bumps, they prevent the heat sink from being FLAT against the TEM, and it will overheat and not pump from the cold side, or it will overheat and melt. If you have a heatsink with curvature or micro bumps, when you torque the heatsinks to the TEM, it (TEM) will crack and that voids the warranty. The mechanical work will be a large problem, Heatsink Grease is for flat surfaces, when you get a thick layer to fill bumps and curves, it's not very heat conductive, only a 2 mil layer is usable, thicker is useless. Sorry.
You can try the thermal rubber pads, but at the thermal density you are dealing with, their Delta T is too high.

[cabindave]

emua,
Interesting experiment. I have to make one point. This could be obvious to you, but if not it is important. The cooler you describe would need to have the hot side heatsink on the outside of the box you are looking to cool. This isn't like a computer where you cool the cpu, then use the fan to blow the heat away. Rather you have to get the hot side of the contraption through the wall or your net effect will be zero. That also makes a problem for weatherization of the outside fan area.
CabinDave