Group discussion on crystal diodes

To launch the discussion, the moderator raised the question of the nature of the life characteristic of crystal diodes—whether the number still usable drops off suddenly at some particular life, whether the number falls off steadily with time as an exponential curve, or whether an asymptote is approached, with the remaining diodes appearing to have indefinitely long life. No one quoted carefully-controlled experiments to distinguish among these possibilities, but some opinions favored the last—a relatively long life after an initial period of "weeding out" faulty units. The question was then asked whether tests to a given percentage of the initial back current of a diode would indicate more rapid aging for high-resistance diodes than for low ones. Measurements were quoted which indicate that initially-highback resistance units vary faster than those with lower initial back resistance. There was some discussion of the methods of testing diodes for end of life. Since the resistance characteristic is so nonlinear, one speaker made a plea for measuring and talking about voltages and currents rather than resistances. It also was suggested that ac observations should be made, with oscilloscopic observation, since this gives a rapid check of the shape of the characteristic and shows changes of this shape during the observation period. The need was also mentioned for observing intermittent shorts or opens, particularly under vibration or tapping.