And what about that temperature sensor? This is what I hear you asking. If you somehow take apart the fan assembly you will find that it is pressed against a sheet of metal that looks like part of the heater. I think this is a significant point as it is not sticking out into the hot air plenum like I thought it would be. You can expect it to read a fair bit hotter than the feed air and probably even hotter yet than the bean temperature.
Once I get a thermocouple set up to measure bean temperature, I expect to determine a temperature offset for different weights of "bean loads". I expect the more beans you try to roast, the bigger that offset will be, but this is just an prediction that I hope to someday measure. If this offset is predictable, it might be a way to get a target bean temperature for a given size load.
The logic board has a pair of calibration trimpots. It actually has three. One is for the temperature sensor (RTD) and has already been discussed. The other two are almost certainly for the heater and fan. They simply provide a voltage to the PIC as some way of calibrating each unit. Why not use EEPROM or something like that? At any event, this calibration is something to be aware of. It may well mean that the timing offsets I measured with an oscilloscope on my unit for triac triggering may not be quite right for your unit. I am surprised that they took the trouble to do this. I can only imagine it would be a minor adjustment, but I could be very wrong, so procede with care.
Tom's coffee pages / tom@mmto.org