"Rob Dekker" <rob@verific.com> writes:

Oh, another thing :

If all transmitters jitter over 'a few centimeters', how does that
affect the beam accuracy ? The jitter will create phase noise, which
will certainly affect the 'thick-ness' of the beam. Do you know how much ?

The effect on the beam power is easy to estimate.  Each location error
will produce a phase error p.  The E field for this transmitter, 
in the desired direction, is reduced by cos(p).  Then you integrate over
the expected error distribution.  This means the beam strength reduction 
will be very small unless the jitter is a significant fraction of a 
wavelength.  

In the simplest case of errors which are uncorrelated and uniformly
distributed, then if the peak-to-peak error is lambda/N, then the
reduction is sin(pi/N)/(pi/N).   Cell phones are about 2 GHz, or about
15 cm.  So a +-2 cm error would be about lambda/4 peak to peak.  Then the
reduction in beam strength is 0.91 db, or less than 10%.

If the errors are uncorrelated and 0 mean, there is no affect on the aim
of the beam.  If they are correlated it may, depending on the correlation,
affect the aim.  This would need to be looked at in detail for such
as scheme, as you would expect a GPS type scheme to generate correlated
errors.  (Otherwise differential GPS would not help).  Over a wide enough
area the correlations may average out, or they might not, depending on
the details.

   Lou Scheffer