Sat Dec 8 11:30:05 CET 2007
synchronization to modulated carrier
i tried to following: use the sign of I to steer the direction in
which the feedback works. works ok for clean carrier, but in full
reversal this leads to problems.
looks like a conceptual problem.
maybe the synchronizer should be slowed down? in a sense that a symbol
transition, which moves through a zero feedback point (in which the
carrier is effectively not controlled), has no noticable effect on the
setting of the tuner, but when this transition is complete, full
feedback is in effect to pull the oscillator in sync again.
using just Q feedback, the PLL seems to stabilize around Q = -120,
with an amplitude of about 30.
say -128, that's -#x0080
#2000 -> #1F80
it's 1/64 th of the frequency, which is exactly the difference between
symbol rate and carrier: the PLL locks to another attractor..
a simple solution seems to be to limit the PLL frequency correction.
anyways, the sign stuff is necessary.