I know this is feasible, and they gave us enough EE in the mechanical engineering program that I can figure it out, but picking the specific components a good EE could probably do off the top of his head... and probably somebody's done it before (or maybe there's something off the shelf?) Anyway... I have a gas heating stove in my living room. Does a great job, overkill even, of heating the living room and kitchen. I want to move heat from there to a bedroom that's on the other side of the wall from the stove. Mechanically that's simple, an air inlet high on the wall above the stove, and a fan blowing into the bedroom near the floor, the space between two adjacent studs forming the duct. A thermostat in the bedroom controls the fan. So far, simple. However, even if the bedroom is cold and calling for heat, I naturally don't want the fan to come on unless the air at the inlet of the duct is warmer than the bedroom. My thought is this: A J-type thermocouple near the duct inlet in the living room, connected to a transistor in the bedroom. The output of the transistor switches a small relay, which is wired in series with the t-stat in the bedroom. Then whenever the the thermocouple in the LR is warmer than the cold junction in the BR, the voltage is positive and switches the transistor on, which in turn turns the relay on, and the fan runs. Simple in theory, but the choice of the specific transistor (and any related components?) is less clear to me. I'm also not clear on what happens when the temperature difference (and thus the B-E voltage) is below the saturation level, i.e. would the C-E current flow, with the voltage drop, be too much power for the transistor to dissapate?. Or is the thermocouple output voltage (say 150μV for a 5°F teperature difference) even enough to switch the transistor?