Strange result of a filter + modulation

Started by mel January 17, 2006

        I am writing a filter ( for pulse shaping ) followed by a
modulation ( BPSK ) scheme for a satellite transmission.

        I have  a result, but I would like to hear your opnion about

        I am using a root raised cosine filter for the input stream,
and applying quadrature modulation after it.

        For instance, suppose I want to transmit this symbol stream :

        1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1

        After I apply the root raised cosine filter, I get this

        I thing this is right, 'cause it has the same input stream form
but shaped ... Does this look ok for you ? It looks nice for me =)

       Well, ok. I guess this is right. But things get a bit stranger
when I do the quadrature modulation.

       As far as I know, the nice thing about quadrature modulation is
that it shows when symbols change values with phase changes ... so, for
this simple input, with only one symbol change ( from 1 to -1 ), I get
a nice phase change in the carrier signal ( first graph, in red ).

       The second graph shows the carrier signal modulated by the
filtered sinal, and it also shows the phase change. My problem is :
isn't pulse shaping ( with root raised cosine filter ) done to
eliminate high frequencies ? In this second graph ( wich is the
filtered signal modulated ), seems to have higher frequencies than the
first one ( wich isn't filtered ). I was expecting a smoother wave for
the filtered signal. What is the problem ?

       Thanks in advance, samuel goto