Hello everyone
I've tried myself to build a small, portable Peltier cooler for my car. Used a TEC1-12703 Peltier, 30x30mm, rated for 15.4V / 3.2A / 30 watt of heat.
After some unsuccessful attempts, found out the Peltier will not work with any sort of heatsink that fits its own size. Tried from ice cubes and ice gel packs to old fashioned chipset coolers.
So I've tried to use a small wooden cabinet as a cooler box. Heatsinks were a cooler removed from a Gigabyte GT440 graphics card, with 80mm fan, on the hot side, and an old CPU cooler (think it's from a Pentium II), about 55x50x50mm, on the cold side.
Measured the amps draw of the Peltier at various voltages and the temps it can achieve.
At 14.1-14.4 volts, it drew about 3.0-3.2 amps and cooled the cold heatsink to 11°C while heated the hot side to 40-44°C in open air at 20°C ambient.
At 12.4-12.6 volts, it drew 2.75 amps and achieved roughly the same temps. Decided to feed it ~12 volts to be able to use a solar panel for an energy source in the future. The cooling power must have decreased, from the original 30W at 15.4 volts, to maybe 20W. Total power draw in the 34W range, fan included.
Insulated the wooden cabinet (~20mm wooden walls) with multi-layered Styrofoam (~28mm, 4 layers of 7mm each, glued with polystyrene glue) and sealed each crack with RTV silicone (white). The inner space shrunk to about 0.9 cu ft, enough to fit about 12 one-pint bottles of beer or 20-30 cans of soda. Styrofoam R-value is about 5 and average R-value for wood about 1, so if both layers are 1 inch or so, it should have a total R of 6. There was no way to fit a rubber band seal on the door, due to rough surfaces of the Styrofoam edges, so used a "labyrinth seal", let the inner layer of the door protrude slightly inside and poured multiple parallel "ropes" of RTV silicone around edges.
Fed electrical power from a laptop charger, rated to a maximum of 12 volt / 5 amps, which made 12.4 volts in practice.
In 20°C ambient air, the cold heatsink inside cooled to 1-3°C and ice began to form on top of the fins. But the bottle placed inside stood warm even after 2 hours. Still air is insulative, even when cold.
Fit a 50mm fan rated to 11 cfm to the cold heatsink, to draw air through the center and exhaust to the sides. The plan was to use inner air as a cooling fluid: draws heat from the bottles, releases heat to the cold heatsink, draws again heat and so on. As long as there are no leaks, permanently churned air inside should cool a bit each pass through the heatsink, down to freezing.
Repeated the test with same bottle. Heatsink temp did not decrease as low and quickly and ice did no longer form, but after 15 minutes the bottle was noticeably cooler.
So the roughly 20 watts of heat drawn away by the Peltier should allow cooling to fridge temps after a few hours, assuming a load of 12 bottles. Ambient conditions should be better compared to a normal upright fridge: it stays in the most shaded corner of the house, there is no hot kitchen around to draw heat from, door should be open infrequently.
There will not be any thermostat fitted, 34 watts of electrical power is affordable and if a 35W-60W solar panel can be used, power is nearly free.