LED as IR Photodiode

Thread Starter

tom66

Joined May 9, 2009
2,595
Okay, this is weird. The PWM waveform (a modified triangle) itself is 88 kHz. But the PWM output is at 255 kHz ???

It must be picking up some noise from somewhere or oscillating.

I was looking at this phototransistor: http://www.rapidonline.com/pdf/58-0115.pdf

Rated 50ns rise/fall time, which is very fast.

I'll have to use these with the peak detector to detect the minimum amplitude, so I can compare against that, because a phototransistor in a normal open collector configuration would invert the signal. Then again, is there anything stopping me using these in an open emitter configuration, given the base is not referenced to ground?
 

Hi-Z

Joined Jul 31, 2011
158
Do you have the triangle wave and audio at the correct levels? Both should be in the range 0V to +5V. If you have them referred to ground, you'll run into problems (the comparator can't handle negative voltage excursions)!

Your link in your last post is to a PIN photodiode, which may be fast but would need a fair degree of sophistication in use. I rather like the idea of the matched LED/phototransistor pair in your previous link, especially as part of a cascode arrangement.

If I were you, I'd get hold of a pair and get them on a bench and have a play: you could insert a single LED in place of the 3 you have (and make R1 39 ohms for an on-current of somewhat less than 100mA), you'd be able get a feel for the phototransistor response to ambient lighting, and to the PWM from the LED (including the risetimes for the cascode arrangement), and at various distances.

This would enable you to find out what sort of collector load you'd need for the cascode circuit. One possibility is to make this an inductor, rather than a resistor. The advantage here would be a total rejection of steady-state current due to leakage and response to ambient light, plus a partial rejection of low-frequency noise.

In fact, if you made the inductance small enough you would end up with a differentiation of the PWM pulse train, which would end up as a train of alternating positive and negative pulses. The original PWM might be recovered using a comparator with hysteresis. You could think of some sort of agc system, and the overall result would be an audio link with fantastic abilities, even in a noisy environment (which ought to demonstrate a high degree of superiority over an analogue link).
 

Thread Starter

tom66

Joined May 9, 2009
2,595
Sorry I haven't got back. I've been very busy...

Anyway I got my photodiodes, and my high speed TL072 op-amps. Will let you know how it goes.
 

Thread Starter

tom66

Joined May 9, 2009
2,595
Really high speed?
High enough for 130 kHz PWM... ;) Main advantage of them over the LM324's I use is they are FET input, so I use them in a transimpedance amplifier for the photodiode. Rise time 330ns, fall time 830ns.

Also faster than the LM324's I use elsewhere in the circuit, those are just used for general purpose audio stuff.

Peak detector part is next, then the comparator, then an actual audio output.

I *could* just use an RC filter on the PWM to get audio, but where's the fun in that? I'll try and devise an integrator and peak detector, or maybe even use a 555 timer, to measure the pulse width and produce a voltage. :D
 

Attachments

Top