[Laser] a free space optical network

[C. Turner]

With our availability of some *really* tall towers (a mile or so high 
and made out of rock) here in "7"-land, we've tossed about the idea of 
some "optical repeaters" - but there are some pretty daunting problems:

- The "Daylight problem".  At the very least, such a system would be 
nighttime-only, for obvious reasons.

A less-obvious problem is if any optical gain is utilized (mirror/lens) 
then a mechanical shutter would be required lest the electronics fall 
victim to accidental focusing of the sun onto them.  This problem is 
awkward - usually involving some sort of motorized device - but isn't 
insurmountable.

- The point<>multipoint problem.  This is probably the most challenging 
aspect.  A traditional "repeater" implies a system by which multiple 
stations spread across disparate locations can be afforded equal access 
- but this contradicts the essential requirement of "antenna" gain at 
the repeater site.

I've seen other proposals that involve conical mirrors (or similar 
odd-shaped reflectors) but the practical problem with this is that of an 
isotropic:  On the receiver, a lot of energy is intercepted from all 
around, diluting the desired signal - not to mention the practical 
difficulties of obtaining/fabricating a custom piece of precision optics 
in the first place!

If we take as an example for transmitting, the energy is diluted as the 
number of solid degrees over which energy is spread is increased.  Since 
there are about 41253 square degrees in a sphere, compared to a 
hypothetical emitter with 1 square degree beamwidth, an omnidirectional 
signal would be reduced by over 90dB at the output of the receiver's 
detector!  Even a hypothetical source that was radiated 
omnidirectionally in only 1-degree tall ring (e.g. narrow vertical 
beamwidth) would take a 50 dB hit compared to that 1 degree emitter!

***

If you consider that even a small (2.5"/6.35cm diameter) lens with a 
typical photodiode or Lambertian LED offers over 50dB of effective gain 
(w/respect to the resulting electrical signal) compared to the "bare" 
photodiode (which, itself, has somewhat less than hemispheric coverage) 
then the difficulties are easier to appreciate.

A more-practical system would be one with "sectored" 
detectors/lenses/emitters:  Linear photodiodes are available - or a 
number of them could be placed in a line - and a compromise could be 
reached, but still, over a very limited horizontal beamwidth with a 
still-narrow vertical beamwidth.  With our tall, mineral-based towers 
out west here, whatever we'd end up using would have to be able to 
accommodate the fact that it would have to accept a range of downtilt 
angles in order to work!

Several years ago, we experimented over a fairly short distance - only 
about 15 miles (24km) or so in which we did a "before" and "after" test, 
removing the LED assembly from its Fresnel lens housing and reverting to 
the secondary lens with a half-angle beamwidth of 20-25 degrees.  This 
results are documented at this page:

http://modulatedlight.org/optical_comms/Second_optical_test.html#no_fresnel

As you can see, the signal levels dropped by 55-60dB or so, going from 
"deafeningly-loud" and "blindingly bright" to "audible only on peaks" 
and "visible only with binoculars."

I'm not saying that it's impossible - but it's worth realizing that 
there are some really big numbers involved with respect to gain!

***

If one dropped the idea of an omnidirectional system to one using a 
point-to-point topology, one could make a "repeater" with several sets 
of steerable optics - but the number of users at any given instant is 
likely to be the same as the number of "steerable modules" that are 
available unless more than one user just *happened* to be in the 
beamwidth of a particular module.  This is doable, but it very rapidly 
gets complicated!

***

As for the San Diego-area optical repeater, it's worth reading the 
article that describes it in detail:

http://www.ham-radio.com/sbms/sd/kd0ifrepeater20070801.pdf

As one can see, it's possible to do this - but don't expect cheap, 
portable, "eye-safe", high-speed communications when doing so!

73,

Clint
KA7OEI