optimal BPSK receiver with Raised Cosine

Hi,
I have a practical problem, which I want to discuss with you.
I have a baseband signal, which is BPSK modulated with a pulse shape as
: alternative polariy pulse for 1s and zeros for 0s. My transmitter
employs a wave shaping filter (which is a RC filter). The receiver
starts from a first an FIR low pass filter, then a matched filter
(which is an FIR filter derived from MATLAB's multi-band filter with
transition bands, to get the transmitted signal freq response). Then
runs an early late gate timing recovery scheme, and then the top levels
schemes like framing and FEC etc... The symbol length is 10 samples.
All works fine, even in noisy conditions too. But, when I want to go
more at more lower power levels (i.e. when recieved signal strength is
weakened, and the signal gets more noisy), my system fails. actually, I
have figured out that in noisy conditions, my timing recovery and bit
recovery ultimately fails. Thus, i loose the frame boundary, and then
looses the FEC sync and goes on.
I thougth that I need more good front end receiver signal processing
(e.g. better matched filter or timing recovery schemes). Can some body
hints at what could be done for the system to work at more noise
levels.? THe system should fail when the FEC is stretched completely,
and the front end should be the last one to fail.
OR what are the normal modem transmission (for simplex) protocols?. I
want to know about the overall system synchornization schemes.
Hope to hear you soon, many thanx
Best Regards
Khubaib

If you're confident that the timing loop is failing first you may want
to look at a better phase detector or a better optimization of the
loop filter parameters.

Just for a benchmark, though, good, continuous signal BPSK
demodulators can maintain synchronization around 0dB SNR (and below).


On 8 Jun 2005 11:13:12 -0700, "khubaib" <khubaib92@gmail.com> wrote:

>Hi,
>I have a practical problem, which I want to discuss with you.
>I have a baseband signal, which is BPSK modulated with a pulse shape as
>: alternative polariy pulse for 1s and zeros for 0s. My transmitter
>employs a wave shaping filter (which is a RC filter). The receiver
>starts from a first an FIR low pass filter, then a matched filter
>(which is an FIR filter derived from MATLAB's multi-band filter with
>transition bands, to get the transmitted signal freq response). Then
>runs an early late gate timing recovery scheme, and then the top levels
>schemes like framing and FEC etc... The symbol length is 10 samples.
>All works fine, even in noisy conditions too. But, when I want to go
>more at more lower power levels (i.e. when recieved signal strength is
>weakened, and the signal gets more noisy), my system fails. actually, I
>have figured out that in noisy conditions, my timing recovery and bit
>recovery ultimately fails. Thus, i loose the frame boundary, and then
>looses the FEC sync and goes on.
>I thougth that I need more good front end receiver signal processing
>(e.g. better matched filter or timing recovery schemes). Can some body
>hints at what could be done for the system to work at more noise
>levels.? THe system should fail when the FEC is stretched completely,
>and the front end should be the last one to fail.
>OR what are the normal modem transmission (for simplex) protocols?. I
>want to know about the overall system synchornization schemes.
>Hope to hear you soon, many thanx
>Best Regards
>Khubaib

Eric Jacobsen
Minister of Algorithms, Intel Corp.
My opinions may not be Intel's opinions.
http://www.ericjacobsen.org