Hi there Ken,
It would be very hard to specify a run of the mill mini frige i think because they probably vary from manufacturer to manufacturer, and the hysteresis would vary too i would think, and not be very dependable for very tight control.
Will there be frequency cycling...the tighter the control band is, the more frequent the cycling. If this worries you then you can think about introducing some controlled loss, where you might even consider a small heater in conjunction with the cooling. I guess it depends partly on how much energy you can get away with using for this thing. Also, a big block of something helps to maintain temperature, such as a big block of steel
acting as a heat capacitance.
Circulating fan...well, if you want to minimize gradients what you could do is build up a very simple sub chamber, where you have a good heat conductor chamber inside a bad heat conductor chamber, with the sample held inside the good conductor chamber. The idea there is that any drafts have to first get through the insulator, then once they do get through whatever makes it has to distribute through the good conductor. The whole thing could be a thick wall metal can inside a styrofoam enclosure like a cheap cooler. A very thick wall would help reduce cycling also.
The thermo probe would of course be inside next to the sample, or you could use several probes along the height of the sample depending on how fussy you are.
Testing of course would be a must, to make sure that the temperature stays within the predetermined levels.
I would bet you could get away with thermistors with the right circuit as long as there was a calibration phase followed by periodic re cal.
I would ask how long the sample has to be kept like this...days, weeks, months, years, etc., and possibly make some mods depending on if it is going to be a very very long time (cryo for example).
I would also ask what are the consequences if something goes wrong...this would help to define what level of reliability we would have to build into the system.
I would also ask how long the sample can be subjected to indifferent heat levels, for example when the sample is transferred into the chamber.
BTW if you are good with electronics, then you are good with heat too...simply transpose variables where heat conduction is current, temperature is voltage, etc. We could actually draw up an electrical circuit describing this systems heat profile.