[Laser] Polarization modulation (Was: High-power polarized light?)

[C. Turner]

Hi,

If I understand the post correctly, you propose "polarization 
modulation" of data at the transmitter, or simply put, one polarity (or 
"sense" if circular were used) would represent a "mark" and the other a 
"space."  IIRC, one interesting place where the use of polarization 
modulation showed up was in the novel "Contact" by Carl Sagan as one of 
the means by which the ET signal was modulated!

I suppose that the question would be:  What would one modulate in this 
manner?

Modulating a subcarrier onto the light (say, a tone) has the obvious 
advantage in that the tone rides atop an ambiguous DC level from the 
detector and by virtue of the fact that we are detecting the frequency 
of the modulation rather than the presence of the light itself, we have 
greatly simplified the receive system.  To be sure, a very 
slowly-modulated signal would quickly get lost in the 1/F (and similar) 
noise from the detector and modulating the light in that manner allows 
us to sidestep that problem - assuming that one managed to tame the 
problem of a DC offset from our radiometric detection scheme in the 
first place!

(There are some applications in which - in a radiometric detection 
scheme not unlike what we might use - that an intervening mirror is 
mechanically modulated.  After electronic detection and amplification, a 
synchronous detector removes that mechanical modulation and restores the 
original "DC" signal - the entire point of this exercise being to get 
that signal away from the "1/F" problem in the detector!  A scheme like 
this is employed in radio astronomy in applications where radiometric 
rather than heterodyne methods are used for detection.)

Now, if we "quickly" modulated the polarity at a reasonable rate (e.g. 
100's of Hz or higher, for example) then we have adequately removed the 
baseband signal from DC so that we once again avoid the "offset 
problem".  Doing this implies that our coding scheme is such that it 
maintains a reasonable "minimum" frequency component (e.g. no long 
strings of "1'" or "0's") as these are fundamentally incompatible with 
AC coupling.  One must be careful in doing this, however, as some 
schemes (such as "bit-stuffing" as used in AX.25 packet) can actually 
worsen the error rates.  If one wanted to get clever, FEC coding with 
appropriate interleaving and "soft decision" thresholds would be 
appropriate - but then you've left the realm of "simple"!

Having two "identical" receivers - one looking at each optical polarity 
- would then, in a sense, resemble the traditional detection of FSK 
where we can differentially detect the state of the bit being sent by 
observing the outputs of the mark and space filters:  If more signal was 
detected from detector "H" than from detector "V", then we can assume 
that it was likely on the "H" plane that the last bit was sent.

Now, is this better than OOK (On-Off Keying) - which would be the 
equivalent with a signal that was not sent on both polarities at the 
same time?  Owing to the presence of a positive "1" or "0" indication 
based on the presence of two detectors, OOK suffers a slight 
disadvantage to differential detection.

An interesting article on differential keying may be found here:

http://homepage.mac.com/chen/Technical/FSK/ATC/

While this discusses AFSK in particular, one can draw parallels to the 
schemes when it comes to the use of two detectors.

Another possible method would be to differentially-modulate the light of 
the two planes in amplitude and then differentially detect them.  Since 
noise will be present in both channels, anyway, the improvement won't be 
huge compared to the other losses noted below.

Considering that running randomly polarized light will result in only 
about 25-40% of the light remaining (using commonly-available 
polarization materials) - and those same losses again at the receiver - 
that would mean that you'd have weigh those additional system losses 
against any benefit that such added complexity would entail!  (There are 
other, more-exotic ways to produce/detect polarization of light, but 
we're talking - for the moment - about using things that an experimenter 
would likely have available and be able to afford!)

Regardless of the above, such a system would be fun to mess with and 
cool to look at!

73,

Clint
KA7OEI


>> Message: 3
>> Date: Tue, 8 Mar 2011 18:39:29 -0500
>> From:<n5gui at cox.net>
>> Subject: Re: [Laser] ~OT: High-power polarized light?
>> To: Free Space LASER Communications<laser at mailman.qth.net>
>> Cc: Tom Becker<GTBecker at RighTime.com>
>> Message-ID:<20110308183929.MN1YX.1314871.imail at eastrmwml47>
>> Content-Type: text/plain; charset=utf-8
>>
>> I do not know if it has been done.  Perhaps there are others on the list that have heard of such.
>>
>> Years ago, I put up a post suggesting that it could be done.  I do not know of any benefit, just that it is a way to impress information on a light beam.
>>
>> A very simple system would be to shift polarization by 90 degrees at the transmitter.  On receive, split the beam into two paths, each going through a polarizing filter so that one peaks on each of the transmitted segments.  Assuming there are no path related polarization effects, the beam would appear a non-polarized sensor to have the same intensity regardless of the information transmitted.  The intensity might fluxuate due to other effects.  Perhaps this could be used to compensate for "twinkle", or at a minimum mark un-reliable signals.  Said another way, the two polarization paths should compliment each other.  Where they have the same state, then the data is corrupt.
>>
>> Another way to use two differing polarized paths is to sent two data streams:  different data, compressed data, or redundancy.
>>
>> Might be interesting to tinker with.
>>
>> James
>>   n5gui
>>      
>>