I have a typical NPN phototransistor. I have it working in a common-collector configuration; see figure 2 of this app note.
Increasing Re will increase the sensitivity, but decrease the speed. I have been studying phototransistors for a few days now, and I think that a transimpedance amplifier can give me additional sensitivity without sacrificing speed, since I won't have Re loading the emitter anymore.
However, I can't seem to find straightforward implementations. The vast majority of app notes describe photodiodes. Unlike a photodiode, a phototransistor needs to be biased, and the few application notes that discuss using phototransistors assume the presence of a negative biasing voltage in their transimpedance amplifiers. I need a solution that works with a single-supply op-amp.
Would a virtual ground on the non-inverting input of the transimpedance amplifier correctly bias the phototransistor? Usually virtual ground is half-way between VCC and GND, but I don't think it has to be. My phototransistor saturation voltage is 0.15V; given VCC=3.3V, does that mean my virtual ground could be at ~3V?
Is there a better way to design this circuit? I would like the output to get as close to GND as possible, because there will probably be a second-stage amplifier.
EDIT:
More details on the application. I'm sensing light levels; low, very low, and off. There are no problems with ambient light, so I'd rather not focus on the phototransistor aspect of this question too much. The bandwidth of interest is around 1-10 kHz. The common-collector almost works; I've raised Re as high as it can go while maintaining the bandwidth I want, but I'd still like Re about 2x larger which results in a signal that's too slow.
No comments:
Post a Comment