OK, so I mounted all the components into the fridge, connected the battery and sat down with a couple cold beers to see what happened. I did not have a multimeter temperature probe inside the fridge, so the door sealed properly, but the insulation on the rear is still "imperfect" because I don't want to plug all the tiny gaps with spray foam until I'm sure that it doesn't have to all come apart again.
At an ambient temp of 21.5 C, it took approximately 45 minutes for the fridge to reach 8 C. I set the cutoff temp to 8 degrees for testing, so the fans shut down as expected and the TEC went into maintenance mode. I added code to direct the ATTiny to wait until the fridge internal temperature rose 4 degrees above the low setpoint before putting the TEC back into active cooling mode, to prevent too many unnecessary starts and stops.
When the fans shut down, I started a stopwatch to see how long it would take for the internal temp to rise high enough to require more active cooling. Disappointingly, I wasn't able to go longer than 2 minutes before the fans and TEC fired back up. This cycle repeated several times until I ran out of beer and patience.
Now, I know I have small air leaks and insufficient rear insulation, but I still hoped for something in the 10 minute neighborhood. I am not satisfied with where I have the microprocessor's temperature probe installed -- it's near the cooling mechanism, which might be skewing its data and causing it to sense the temperature of the cold air coming from the cold side heatsink fan rather than the actual air temperature inside the box. Also, I'm struggling to cut and shape hardboard insulation precisely enough to avoid gaps and ugly edges with missing "pebbles" of insulation.
So, I'm going to pull everything apart and remake the cooling system insertion point out of ABS sheeting. Then, I can use spray insulation to fill the back area to a greater R value and with fewer gaps.
Here are a couple pics of the current setup that will probably be gone by the time you read this --
The test rig.
The rear of the cooling insertion point. The gaps are obvious, but I can't fill them without using spray foam, which I don't want to do for initial testing.
The cooling insertion point from inside the fridge. I circled the ATTiny's temperature probe in red to illustrate why I think it's in a crappy spot. It gets a direct hit from the wash of cold air when the heatsink fan is running, and probably freaks out when that cold air disappears.
I mounted it here to avoid drilling unnecessary holes in the fridge's plastic shell, but I think it's got to move.