The problem with Diodes and the like is that it is possable to induce a pattern into the output of these things. Well, might we be able to look at how and set up a counter such that if the open end is held high on one diode it gets driven down on a second circuit which get added togethor before we go onto to 0/1 ballencing? Next I would set up a drifting clock rate on the device reading the diode, this messes up many efects of interference and might help look for induced patterens! Hey, imagen a random number generator that tells you when the bad guys are trying to influence your random numbers! ||ugh