[Laser] PSK and laser communication

[TWOSIG at aol.com]

Under the topic "Re: [Laser] Troposcatter and Information Theory-2" Tom 
Upton, AD6N, made the following comments:

>PSK31, also known as "warbler", is very narrow band, and is used by 
>the QRP (Reduced Power) Hams to talk amazing distances (1500 miles) 
>on much less than a tenth of a watt of radio frequency power.  
>
>It is quite able to be used on telephone freqs.  It is easily detected below 
>the noise threshold of receivers.  
>
>Should we not be looking at it in addition to long duration dits and dahs?
>
>Tom Upton  AD6N

Tom, 

I am glad you brought up the topic of PSK31 as a coding method for laser 
communications.  I think that it has some potential, but it also illustrates my 
point about how things are different when you apply HF methods to light 
communications.

Our radio club had a PSK31 station for Field Day.  Using readily available 
software and a hardware interface we used a laptop comuter's sound card to 
encode and decode PSK signals, which the hardware interface directed to and from a 
SSB transceiver for transmission/reception on 40M, 20M, and 15M.

I could connect the same audio output to the auxillary input to my Ramsey 
transmitter and take the audio from the Ramsey receiver and send it to the sound 
card input.  It would work as a PSK station, much like our Field Day setup.  
With the laser systems other configurations are available, and the system 
trade-offs and limitations are different.

Other Configuration:

When I operate my Ramsey stations, the are full duplex - both transmitters 
are running all the time, and I can hear the other station while I am talking.  
Our Field Day PSK station was simplex - we listened, then when the other 
station stopped transmitting, we turned on our transmitter, not being able to 
receive until we turned it off.  The software on our laptop might have a full 
duplex mode that we had no opportunity to try during Field Day, but I would 
certainly want that capability for a laser communications system.

Different system trade-off and limitations

The PSK31 communication mode was specifically designed for efficient use of 
the available spectrum.  The techniques used at HF to preserve its spectrum 
efficiency, do not translate well to laser communications, and if they did, it 
would not improve the effectiveness of a communications channel.

Using an example of BPSK - Bi-Phase Shift Keying, assume the I have a data 
stream of ones and zeros that I am going to encode on an 800 Hz square wave.  If 
I have a long string of ones, I will get a long sequence of high for 625 
microseconds followed by low for 625 microseconds and then alternating back and 
forth.  If I have a string of zeros, I will get a similar sequence.  If I 
compare the two sequences, when the ones are high the zeros will be low and when the 
ones are low the zeros are high.  In other words, the two sequences are 180 
degrees out of phase.

Now assume for the moment that at the end of a one the output is high and I 
am going to transition to the next digit which is a zero.  The last 625 
microseconds of the one will be followed by the first 625 microseconds of the zero 
which will also be high.  The previous alternating sequence of alternating high 
and low, will now have a blip that is a double length high that marks the 
phase reversal.  Every transition from a one to a zero, or a zero to a one, will 
have that phase reversal blip.

Now imagine what the signals look like on a spectrum display (which is 
usually available for "tuning" PSK31 stations).  If you have a string of ones or 
zeros you will see a spike at 800 Hz.  Since I said we were working with a square 
wave, you will also see a spike (theoretically one third as high) at 2400 Hz. 
 And another spike ( One fifth the first) at 4000 Hz.  Similar and 
progressively smaller ones at 5600, 7200, 8800, 10400, and so on.  You should notice 
that this is not efficient use of spectrum if you transmit all those harmonics.

The first thing that you do for this example of PSK31 is use an 800 Hz Sine 
wave.  Now your spectrum display shows one spike at 800 Hz.  But that is only 
if you are sending continuous ones or zeros.  Now imagine that you are sending 
alternating ones and zeros at 31.25 Hz.  The spike at 800 Hz disappears and is 
replaced by two spikes at 768.75 and 8031.25.  Ooooooops!  There are also two 
more spikes, a little smaller at 737.50 and 8062.50.  And two more even 
smaller at....

OK, I have talked too much about this. Simply switching from ones to zeros 
gives you a 31.25 Hz square wave times the 800 Hz tone.  To get the efficient 
use of spectrum out of PSK31 in the example of alternating ones and zeros you 
need to multiply an 800 Hz Sine wave times a 31.25 Hz Sine wave.  In physical 
terms it puts a 31.25 Hz envelope arround the 800 Hz tone and forces any phase 
reversals to take place when the envelope is at or near zero.  Encoding and 
Decoding this kind of digital data is easier to do with square waves rather than 
Sine waves inside Sine wave envelopes.

Well, the software has already been written, so why not just use it as is, 
rather than change it for laser communications.  In simple terms, lasers don't 
like to be analog modulated.  The PSK31 signal goes to the transceiver 
microphone input because it needs to be processed as a linear signal.  Any distortion 
of the audio from the sound card will increase the bandwidth of the 
transmitted signal.

I could send the typical PSK31 audio into my Ramsey transmitter and it would 
be received on the other end just fine, assuming that I have the signal 
strength to reproduce the audio.  If you look at the spectrum at the input to the 
audio filter of the receiver, it would show a typical PSK31 spectrum, arround  
800 Hz in the above example.  There would also be lots of energy showing up at 
18 KHz, which is necessary for the modulation method, but does not improve my 
communications.

What I would rather see done is the laser transmiter be driven by a square 
wave that represents the digital data stream controling the audio tone.  Perhaps 
the best way to do this is to send the serial data out to a one bit port that 
goes to one input of an exclusive OR gate.  The other input would be from a 
square wave tone generator.  The output of the exclusive OR gate would simply 
key the laser on and off.

The received spectrum would have harmonics all over the place.  Some of those 
harmonics might be useful in decoding the information being sent.  They are 
after all, redundancy in the communication channel.  If they are thrown way, 
that is OK too.  I have simplified my transmit equipment without degrading the 
signal.  I believe that I have improved the effectiveness of the signal, but 
that has yet to be proven.

I hope all these words have given some suggestions and encouragement for 
experimenting with PSK31 on light beams.

James
N5GUI