I have done some testing on this the past few days, I have a fan wired into my cooling and heating circuit so whenever cooling or heating is called on by the controller, the fan runs also, so here are my observations....
large unresized pic here:
Scenario 1
I hung the probe in the air and put a separate thermometer probe insulated against the carboy and set the controller for a 4 degree difference, with this setting the compressor was attempting to come on before my compressor delay(10 minutes). The carboys were staying within .2 degrees F with this setup.
Advantage: carboys temp most stable
Disadvantage: Compressor was running every 10 minutes. During fermentation you would have to manually adjust the controller to compensate for heat produced in fermenting wort.
Scenario 2
I hung the separate temperature probe in the air and placed the controller probe insulated against the carboy, I set the controller for a .3 degree C (that was the lowest it would go) difference, with this setting the compressor was cycling at a much longer interval (maybe 30-40 minutes). The air in the chamber was varying upwards of 10 degrees F during the cycles.
Advantage: you can set and forget and the controller will control the temp of the fermenting wort. Less cycling of the compressor, longer life of freezer and most likely more economical for power consumption.
Disadvantage: Large temp swings inside the compartment, although due to the large thermal mass of the wort this doesn't seem to be a problem
I brewed an Oktoberfest last night (with Kolsch WLP029 yeast) and set the controller to 63 degrees with a .3 difference and insulated the probe against the fermenter wall.....I will report back tonight once fermentation takes off and I see how things are performing.