[Laser] more strobelight stuff

[TWOSIG at aol.com]

Here are some rather random thoughts on a strobe light communication  system. 
 They seem to be disjointed to me, so I hope there will be some  value within.
 
The overall system seems to be similar to a low duty cycle, high peak power  
pulse system.  Radar techniques may have been developed which could be  
applied, if you can find them and translate them into current terminology and  
available hardware.
 
If you are sending out a broadband light pulse (such as from an un-filtered  
xenon flash tube) you might look into using a photo-multiplier tube as your  
detection device.  The PMT is more sensitive to the higher light  frequencies, 
and is a very high gain device.  PMT do not seem to be a good  match for red 
lasers, but they might do well for you.
 
Someone correct me if I am wrong about this, but I don't think you can send  
a 31.25 baud PSK data stream on a carrier  under the 120 Hz buzz.  I  think 
there are more fundamental reasons,  but I think your first problem  is that the 
PSK31 software is not designed to send out a signal below 300 Hz  lower 
passband frequency of SSB transceivers.  
 
If you are thinking of using BPSK, you might try to find someone to modify  
the PSK-31 software on both receive and transmit.  All the software  versions 
that I am aware of are intended to put out a pure narrow signal for use  by a 
linear transceiver and to decode such a signal from within a receiver audio  
passband.   Your light transmission equipment is non-linear and  broadbanded.   
Think of it this way:  For a 31.25 Hz alternating  string of ones and zeros 
being sent on a 400 Hz carrier wave, the sound card on  a typical PSK31 signal, 
puts out a 400 Hz sine wave that is within a modulation  envelope that is a 
31.25 Hz sine wave, and the 400 Hz wave will reverse phase at  each zero 
crossing of the modulation envelope.  If you were trying to send  out BPSK with a 
laser diode, you would want a 400 Hz square wave that did not  change in peak 
magnitude ( no sine wave modulation envelope) that has 180 degree  phase 
reversals ( at the times when the linear system would have had zero  crossings).  The 
spectrum of the linear system would be two narrow spikes,  one at 368.75 Hz 
and the other at431.25 Hz.  The spectrum from the  laser diode would look like 
the sum and difference products of two square waves  (which it is)  ---- (the 
sum of 400 Hz plus one third of 1200 Hz plus one  fifth of 2000 Hz plus one 
seventh of 2800 Hz and so on...) multiplied by ( the  sum of 31.25 Hz plus one 
third of 93.75 Hz plus one fifth of 156.25 Hz and so  on...)  For the laser 
diode transmitter the real problem is getting the  laser to fire at the right 
time, including the phase reversals, which the linear  software will perform at 
zero volume.  Any timing error on transmit will  show up as carrier frequency 
shifts.  Receive software will probably work  on at least the "plus" half of the 
multiple images that are above the  noise.  I don't understand enough to say 
if the "minus" half of the  images will be readable or not.  The bottom line 
is the software was  optimized for narrow linear RF signals, not for the 
typical light signals,  that use very non-linear output devices.  The good thing is 
that light  signals are not an interference problem.  The bad thing is that 
your strobe  system is even more non-linear and more broadbanded than a laser 
diode  system.
 
The receive software you need to process the received signal needs to be  
adapted to accept the light pulse at the beginning (or end or some other  
reference point) of each time hack that would be the laser diode square  wave.  The 
time between pulses is the same as the period of the square  wave.  A phase 
reversal event would be marked as the time between pulses as  either one half or 
one and one half times the square wave time period.  If  the software cannot 
be adapted (perhaps because the sound can cannot process the  short pulses) 
then some form of hardware reconstruction would be needed.   That will require 
detecting the desired pulse and rejecting noise.
 
Strobe lights are usually very low duty cycle devices.  You might want  to 
consider a pulse width modulation scheme.  With laser diodes we think in  terms 
of 10 to 90 percent duty cycle, but you may need to think in terms of 1 to  10 
percent duty cycle or a tenth of that.  You might be able to adapt a  camera 
strobe technique that quenches the flash when the light meter integrates  
enough reflected light from the target.  You could use a long pulse (  normal) for 
a one and a short pulse ( quenched part way through the normal  cycle) for a 
zero.  Another approach would be to send a pulse train of  two or more 
flashes.  The number of pulses in the train, or the spacing of  the pulses would 
convey the information.  In the end, the system needs  hardware and software that 
can exploit the capabilities of you transmitter  source and detector.
 
If you get a strobe technique that will work, it could be used as a driver  
for a photo excited laser.  That may not be your objective, but what you do  
could be a stepping stone for something someone else might try.
 
As far as I know, the receiving software will not correlate the harmonics  of 
the pulses and square waves you will be receiving.  If it did, that  
additional information could be used to reject noise on the data channel.  
 
 
Good luck with you experiments.
 
James
N5GUI