[Sphyrna-devel] 'Doppler' a close match with computed doppler.

[Ian Stirling]

http://www.mauve.plus.com/gps.gif
http://www.mauve.plus.com/2.dat
http://www.mauve.plus.com/8.dat
http://www.mauve.plus.com/26.dat
http://www.mauve.plus.com/gps

Using the gnuplot plot line below on the above files gives the GIF at
the top.

plot  "2.dat" with dots, "8.dat" with dots, "26.dat" with dots, "gps"
using (30*$0+27800):($14*-328), "gps" using
(30*$0+27800):($110*-328),"gps" using (30*$0+27800):($66*-328)

This indicates quite a good match between the measured from strace value
of doppler, and 'doppler'/-328 being how fast in m/s the satellite is
approaching you.

It is not perfect, and more data would be very nice.
There are indications that some of the inaccuracy may be in the internal
clock.


It also indicates that the initial doppler search phase goes up to
+-1280m/s or so. Which is sane, given that the satellite doppler
approaches 700m/s, and the GPS maximum of 1000Kt is 514m/s, or a total
of 1200m/s.
This means that it can track and lock to a worst-case (at least for a
ITAR compliant reciever) - when you are flying at 1000Kt directly
towards or away from a satellite (which would usually imply a fighter
aircraft to ascend that steeply).

If it performs as well at these frequency extremes is another question.


Looking at the raw data from the 'doppler' output.
It looks like it tracks these to typically better than 8m/s, and jumps
from one side of the error to another.