[Laser] Optimizing sound card modes for optical communication

[F1AVYopto at aol.com]

Hi all
The optimal program  choice is not easy.
Most of the simple laser driver systems are not linear  and sending multi 
tunes simultaneously is difficult.
So all the modes  "one tune at a time" seem better
Because this problem all the experiments we  did, we only used fast chopper 
to give perfectly square waves.
The most  important point to check is to keep a "ON time / OFF time" optical 
flux constant  ratio independently the input level and the modulation 
frequencies.
To check  the symmetry from the real optical signal with an oscilloscope and 
a photodiode  loaded by a resistor is very important.  
The low pass filter  effect from our slow RXs rejects harmonics and the 
fundamental  frequency is mainly FFTDSP processed.
For our NLOS laser experiment to more  40 Km we have tried 3 programs: JT65, 
JASON and WOLF-GUI.
The JT65 mode gives  two problems:
The audio frequencies are rather high and so just compatible  with our ultra 
sensitive but slow RXs.
The useful bandwidth often contains  some spurious lines from the sky 
background ( 50 Hz x n or 60 Hz x  n).
JASON.
JASON works very well and allows - 35 dB under noise at slow  speed.
But two problems:
JASON is very  extra noise sensitive. A punctual flickering optical source 
(scintillation)  provides two noisy lateral bands both side the useful signal 
that often avoid  the JASON efficiency. 
JASON has no loop protocol to accumulate signal pattern via a statistic  
process.
That means that for slow signal level variations, characters could  be lost.
WOLF-GUI.
WOLF-GUI is the best
It uses the phase modulation and a very  sophisticated algorithm for signal 
recovery.  
But the sound cards must  be first match together to less than ONE Hz 
accuracy.
A 144 MHz 2500 Hz FM  link with the WOLF-GUI calibrating process allows a 
less than 0,1 Hz accuracy  for a further laser modulation to 175 Hz !
WOLF -GUI allows a -45 dB signal  recovery in 2500 Hz noise BW for 30 mn and 
one data frame !!!!
It is with  this program we succeed our 40 Km QSO from the clouds.
It is a very noise and  QSB resistant program !
Yves F1AVY 
http://pageperso.aol.fr/YvesF1AVY/