Modulating a constant current driver

[xanatos]

One thing I'm finding in my research is that LDs don't like to be switched fully on/off when modulated, and so it's better to switch the drive current between max and a lower value that brings the diode below the lasing threshold, but not actually off. Is this true to what you know?

[taggalucci]

Yes, keeping the diode powered, but just under threshold improves response time, and therefore linearity, when going from dark to powered.

[xanatos]

Thanks for the confirmation... I'll add that into my design, unless Daedal comes through and I get a much better schematic that's just as simple! :-)

[xanatos]

I have modified the original schematic I posted after breadboarding the circuit and adding another resistor to provide a minimum current so the diode is never fully off. I have tested it with a 1.0 ohm resistor in place of the LD to measure current, and the calculations hold for various values of R. I will add a reverse bias diode for LD protection, but this circuit SEEMS pretty good so far.

It appears the calculation for Max I ( I = 0.7/R ) is dependent of the transistors Vbe. The calculations seem to suggest that for the 2N2222A, that 0.7 value should actually be 0.66 - but then again it could be due to the resistor values being off slightly... not sure which yet.

I'll post my results again when I have a fully functioning LD with modulation applied at the desired current (or after I cook my LD, whichever comes first... :-) )

Dave

[xanatos]

I changed the location of the minimum current setting resistor and added the reverse spike protection diode, and I'm PWM'ing my diode with a propeller (a microcontroller if you aren't familiar with what a propeller is). I've run it to 10,000 Hz and have a nice square waveform going in, and a fairly square response as read on a photodiode.

Am I missing anything here - does anyone see any demons lurking in my schematic waiting to bite my faithful diode? :-)

Thanks for your thoughts.

Dave