Forums

demodulation problem

Started by kalden July 14, 2004
Hello.

I am working on a software demodulator and have a few questions. 
First off let me say that I am just getting into this DSP stuff.

Anyways... I have a file of samples representing a bpsk modulated
wave.  The popular method of demodulating seems to be through I/Q
channels so that the phase and magnitude can be detected.  Is there a
way to convert my non complex samples to an I/Q format?  If not, is
there another reasonable demodulation method?

Once i get a grasp on this I plan on adding support for qpsk and
different forms of qam.

Thanks in advance.
kalden <aldenk@rl.af.mil> wrote:
kalden,

I am not an expert in this area, but just the same, if I'm not mistaken,
Rick Lyons 2nd ed. book ("Understaning Digital Signal Processing") may
have some information on this topic. I don't have my copy handy at the
moment but I think Rick updated the secion discussing I/Q. A very "user
friendly" book.

FWIW,

Dave

> Hello.
> I am working on a software demodulator and have a few questions. > First off let me say that I am just getting into this DSP stuff.
> Anyways... I have a file of samples representing a bpsk modulated > wave. The popular method of demodulating seems to be through I/Q > channels so that the phase and magnitude can be detected. Is there a > way to convert my non complex samples to an I/Q format? If not, is > there another reasonable demodulation method?
> Once i get a grasp on this I plan on adding support for qpsk and > different forms of qam.
> Thanks in advance.
kalden wrote:

> Hello. > > I am working on a software demodulator and have a few questions. > First off let me say that I am just getting into this DSP stuff. > > Anyways... I have a file of samples representing a bpsk modulated > wave. The popular method of demodulating seems to be through I/Q > channels so that the phase and magnitude can be detected. Is there a > way to convert my non complex samples to an I/Q format? If not, is > there another reasonable demodulation method? > > Once i get a grasp on this I plan on adding support for qpsk and > different forms of qam. > > Thanks in advance.
To convert to I/Q format modulate your sample set by cos(w*t) for the inphase and sin(w*t) for quadrature, and lowpass. You'll get two real-valued numbers which you can then play with. -- Tim Wescott Wescott Design Services http://www.wescottdesign.com
Hi Kalden,

I just went through the exercise of post process bpsk demodulation from a
file of samples just like yourself. It was a fun exercise and turned out
really good in the end. I started down the I/Q path just like you are
thinking. I tried a Costas Loop which worked really well but I was unhappy
with the processing time due to the math on both  I/Q channels. So what I
did was to go to a digital PLL with narrowing loop filters. Works excellent
and super quick.

However the PLL approach might not be the best solution with the
consideration of qam or qpsk looming in the future. I would go Costas for
that.

Are your samples RF? How big is the file?  What type of S/N ratio are we
looking at? Is there preamble to lock and snap down on?  If yes how long is
the preamble. I have done bpsk with and without preamble.  Without is a
little more tricky.

I might be able to help if you give me a little more info. Like I said I
just finished the exercise myself.

Thomas


"kalden" <aldenk@rl.af.mil> wrote in message
news:57834864.0407141121.58c1fa26@posting.google.com...
> Hello. > > I am working on a software demodulator and have a few questions. > First off let me say that I am just getting into this DSP stuff. > > Anyways... I have a file of samples representing a bpsk modulated > wave. The popular method of demodulating seems to be through I/Q > channels so that the phase and magnitude can be detected. Is there a > way to convert my non complex samples to an I/Q format? If not, is > there another reasonable demodulation method? > > Once i get a grasp on this I plan on adding support for qpsk and > different forms of qam. > > Thanks in advance.