Have you ever bought a wine cooler? They are huge, bulky and ri-donck-u-louse looking. First off, it does not make any sense to design a wine cooler as a cube. The main goal should be to minimize the surface area of the enclosure as that is where heat transfer occurs and thus is a large dictator of how much energy you must feed the system. Wine bottles naturally stack in a triangle and in this fashion they are packed the tightest. So, why not design a triangle shaped enclosure that minimizes how much heat we must pump out of the enclosure and thus how much energy we are using? Seems reasonable, right? Here is a first pass at doing so.

First off, some initial calculations to get a feel for how much heat we gotta pump outta this thing. We will shoot for a cooler roughly sized to hold three wine bottles or a sixer of Busch Light, yum. We will only consider conductive heat transfer for the time being. The amount of energy we need to pump out of the cooler through the TEC (Qpump) is then equal to the amount of heat weaseling its way through our insulation (Qin) and the amount of power we are feeding to the TEC (Qjoule). Qjoule will not actually be the amount of power we are feeding the TEC but we can use it as a worst case scenario.

Based on some proposed dimensions for the cooler, an estimate of the k value of our insulation (the k value just quantifies how good or bad insulation is), and the desired temperature of our wine (55*F) relative to normal room temp (75*F) we can calculate Qin. I doubled the Qin value as originally calculated it was the absolute minimum amount of heat needed to be pumped to retain desired temperature if all else were to remain constant. As there will also be some small amount of conductive and radiative heat transfer taking place we will need a little more heat pumping capacity. We also might run into different environments where it is necessary to pump more heat, aka: if room temp is greater than 75*F.

So as it stands, with some Busch League All-Star (definition: bad, crappy, swagged, horribly awesome) code I slapped together here is what it looks like and how it is functioning.

First off, some initial calculations to get a feel for how much heat we gotta pump outta this thing. We will shoot for a cooler roughly sized to hold three wine bottles or a sixer of Busch Light, yum. We will only consider conductive heat transfer for the time being. The amount of energy we need to pump out of the cooler through the TEC (Qpump) is then equal to the amount of heat weaseling its way through our insulation (Qin) and the amount of power we are feeding to the TEC (Qjoule). Qjoule will not actually be the amount of power we are feeding the TEC but we can use it as a worst case scenario.

**Image 1:**Initial thoughts on size and heat transfer involved

Based on some proposed dimensions for the cooler, an estimate of the k value of our insulation (the k value just quantifies how good or bad insulation is), and the desired temperature of our wine (55*F) relative to normal room temp (75*F) we can calculate Qin. I doubled the Qin value as originally calculated it was the absolute minimum amount of heat needed to be pumped to retain desired temperature if all else were to remain constant. As there will also be some small amount of conductive and radiative heat transfer taking place we will need a little more heat pumping capacity. We also might run into different environments where it is necessary to pump more heat, aka: if room temp is greater than 75*F.

**Image 2:**Solving for conductive heat transfer and power added by TEC

Here is an ugly looking diagram of how the components will all be connected. I will post later exactly what was used to create the cooler once I fine tune it along with a cleaner diagram.

**Image 3:**Component diagram

**Note**: a wine bottle at room temperature that is put in a wine cooler holds latent heat that also needs to be pumped out of the enclosure. This was not included in the initial sizing of the TEC due to the fact that if you can pump more heat out than is entering from the outside surroundings (Qin) this latent heat will eventually be pumped out as well... it just might take awhile.So as it stands, with some Busch League All-Star (definition: bad, crappy, swagged, horribly awesome) code I slapped together here is what it looks like and how it is functioning.

Something has gone horribly wrong with uploading this video so here is the link to how the user interface is currently functioning (can't embed vimeo :( ):

**Image 4:**Outside of enclosure, wiring yet to be subdued

**Image 5:**Looking in the side hatch, foam insulation visible

More to come on this project...