Planning for Fermentation Temp Control - Need some veteran advice

I'm looking at building a simple temp controlled fermentation cabinet. This is just for the next few months until I'm out of college, I don't want to buy freezers and lug them around with me when I get a real job.

Instead of just sticking an AC unit into the box and controlling it that way I think I might be able to save a lot of space in my tiny apartment by gutting a mini fridge and putting the cold heat exchanger in the cabinet while obviously keeping the rest of the components outside.

Does anyone have any experience with hacking mini fridges like this? I'm assuming its not feasible to remove the coolant from the system, re-plumb it and then repressurize so I might have to work with the shape and routing of the current system, which might not be a big issue anyways.

In the interest of saving some more dough I'm looking at using a microcontroller with temp probe as a simple PID controller. This could also give me some extra things like temp data logging, etc.

Any advice?

Also, anyone know anything about peltier cooling? Perhaps if I could find a big enough peltier cooler it could give me the ~5 degree temp difference to keep an ale at 65-68deg in a 68-72 degree room (plus and heat released from fermentation).

1. You could extend the fridge by cutting a side wall:

https://www.homebrewtalk.com/f51/4-6-cu-ft-fridge-10-1-cu-ft-fermentation-chamber-conversion-88554/

2. I'm looking into the peltier option, but only because I have a slightly broken peltier cooler. (though I'm having a hell of a time finding the replacement part that broke)

I'd be interested to see which way you went as I'm looking to do something similar. Come May our Craigslist piles up with dorm fridges.

I did this one in my garage, but no reason you couldn't do something similar in a cabinet indoors.

https://www.homebrewtalk.com/f11/new-fermentation-chamber-build-102846/

PID algorithms aren't necessarily the best choice for controlling a compressor, they typically switch too often. You're usually better off with a more simple differential hysteresis.

Yuri did some experiments a while back with peltier cooling, a good read:
https://www.homebrewtalk.com/f51/diy-thermoelectric-temperature-control-27524/