An attempt at building a SAH (sample-and-hold) circuit switched by an NPN BJT (analog switches are usually done with FETs, I haven't come across a BJT-based implementation).
To my great surprise, it is actually seems feasible, even if that means the transistor is working in a rather unusual manner:
- signal is injected in the collector
- when storage cap must discharge to "read" a lower voltage than the previous one, transistor operates in "reverse-active" mode (collector becomes emitter, emitter becomes collector).
The literature I found says reverse active mode is useless because beta will be very low. However, for this circuit, beta is irrelevant: we only use off mode and reverse saturation mode, and reverse active actually comes in handy.
The cap on the base is a "speed-up" cap, here to compensate for the fact that turn-on and turn-off characteristics of BJT's aren't symmetrical, which leads to slower turn-off. Not sure that it helps in any way in the sim.
A question / challenge for you electronics afficionados: when clock goes low, BJT switches off, hold mode starts, the storage cap voltage drops by a smidge from the correct hold voltage... I'd like to eliminate that little bump. Any ideas ?
Enjoy.
