So I tried out the peltier scheme this summer, with every intention of running chilled water through a stainless coil in my fermenter as it sounds like you're thinking. It seemed like a great plan in that it really wouldn't take much room at all outside of the fermenter, and would be portable and collapsible for storage.
The peltiers themselves are often rated to something like 67 degrees temp difference in Celsius - I thought that with proper heat management on the peltier I should easily be able to get to 0 f. In practice I was able to get roughly 25 degrees f cooler than ambient in a small test loop.
I tried with 2 different types of peltiers - the smaller 80 watt ones, and a big honkin 245 watt. I cannibalized a few different computer power supplies to provide plenty of 12v juice. My initial plan was to attach computer cpu heatsinks and fans to the hot side and computer liquid cooling water blocks on the cold side. The heatsinks would do their best to dissipate 80+ (or 245+) watts of heat, and bring that side as close to ambient as possible. The cold side would then presumably be 67 degrees C colder than the hot side, and the water block on that side would allow me to pump glycol through a copper heat exchanger designed specifically to tranfser heat to/from a liquid. I suggest browsing the liquid cooling section at
www.frozencpu.com .
I got copper blocks to put between the various parts, and lapped those pretty damn smooth. I used heat sink compound sparingly, then in quantity. I tried without the copper blocks. I verified that the cold side would generate frost in my garage (Houston) in the summer. The water temp never got as low as I would have expected. The heat sinks were definitely blowing hot air, but my IR temp guage (cheap, likely very inaccurate, or at least not able to get into position to properly measure the temp through the fan and fins of the heatsink) never read terribly high. Based on that temp - even accounting for what I thought of as gross miscalculations on my part - I expected lower water temps.
SO I went to the next level and bought a computer liquid cooling radiator, presuming I wasn't keeping the hot side close enough to ambient. This at least tripled the surface area for cooling the hot side. I sandwiched the peltier between two water blocks - one circulating cold glycol (to the fermenter obstensibly), the other side hot (to the radiator). The radiator and fans (3 now) worked like a charm and kept the temps reasonable - it got warm but not hot by any stretch (and I know the hot side of the peltiers can cook). Cold side still generated frost, but the water temp stayed moderate.
I suppose getting into the 57-58 degree range isn't too bad, but I was hoping for a lot more. As well, this was a low volume test loop - I presumed 5 gallons would be a much tougher job to keep that cool. In fairness, I had no insulation, but my hoses were silicone. I put a temp probe in the cold side reservoir and let it run for 30+ minutes - temp was more or less stabilized at that point.
My plan was to have N units in series (peltiers sandwiched by water blocks) to increase cooling power. Given my lack of success with such a small volume (< 2 liters) I couldn't imagine much improvement when stepping up to 5 gallons with 2 or even 3 units, so I halted my experiment.
Nifty gizmos, definitely. Not feasible for (south Texas at least) garage fermenting. Not to rain on your parade, and it's entirely possible that my experiment was flawed. I feel like the number of configurations I tried was pretty comprehensive though. I say buy a mini-fridge or 3 instead for cost efficiency.