Hi,
I am trying to plot the BER verses SNR using SVN estimation method (I am
using dbpsk.py and dbpsk2.py), but it is not matched with theory. Here
is my
problems and questions:
1- when I use dbpsk.py with using feedforward_agc_cc I see three centre
constellation points, one with the real part around -1.5, one with the
real
part around 1, and the third one with the real part around 1.5. Is it
the
problem of using feedforward_agc_cc or I am doing it wrongly? (I tried
it
many times).
2- When I use dbpsk2.py the constellation points are much more natural,
but
still they are not appropriate for snr estimation, since the variation
of
constellation points along the imaginary axis is much less than their
variation along the real axis. Another problem here is that the
constellation points are not symmetric along the real axis. In fact, one
of
the centre constellation points has a real part around -0.9 and another
around 0.5. Also, the distribution of the points are not completely
matched
with the AWGN channel assumption (I am using cable).
To me it seems that the problems comes from the time_recov part, but I
am
not sure. Does anybody else have the same problem with this modulation
or
can anyone help me to solve this problem?
Many thanks,
Mostafa
View this message in context:
http://old.nabble.com/problem-with-dbpsk-and-dbpsk2-tp28206560p28206560.html
Sent from the GnuRadio mailing list archive at Nabble.com.