Forums

Sampling Bandpass signals

Started by Darien Martinez September 23, 2000
Hi everybody,
I have some question about sampling bandpass signals.
There is sampling theorem for this kind of signals,
the sampling rate must be two times the frecuency of
the max information component instead of the max
frecuency afther modulation. If you play with the
spectrum of the discrete signal you will see that but
there is a distortion when the sampling is not
sychronized whith the center of the bandpass signal.
Anybody know something about this?
This is interesting because the adquisition system
migth be less expencive. Any comment is wellcome.
Bye
Darien _________________________________________________________



Hi Darien,

The Bandpass Sampling Theorem says that you can have 2X BW of the
signal provided fc1 (or fc2) is an even integer(odd integer) multiple
of BW. If this doesn't hold, then you get the distorted spectrum
after sampling.

Hope you are refering to this problem.

So, you can extend fc1 and fc2 to fc1' and fc2' respectively(since
the signal is bandpass) to get new BW' (= fc2'-fc1') > BW so that,
fc1' = k1 * BW' and fc2' = fc1'+ BW'.

regards
arun

-- In , Darien Martinez <darienmt@y...> wrote:
> Hi everybody,
> I have some question about sampling bandpass signals.
> There is sampling theorem for this kind of signals,
> the sampling rate must be two times the frecuency of
> the max information component instead of the max
> frecuency afther modulation. If you play with the
> spectrum of the discrete signal you will see that but
> there is a distortion when the sampling is not
> sychronized whith the center of the bandpass signal.
> Anybody know something about this?
> This is interesting because the adquisition system
> migth be less expencive. Any comment is wellcome.
> Bye
> Darien > _________________________________________________________
>





Darien,
Hi! Oversampling is a good rule of thumb. For example,
if your highest frequency component is 10Khz. Instead
of using just 20Khz as sampling rate, you could use
44.1Khz or 48Khz as sampling rate.
--emer

On Sat, 23 Sep 2000, =?iso-8859-1?q?Darien Martinez?= wrote:

> Hi everybody,
> I have some question about sampling bandpass signals.
> There is sampling theorem for this kind of signals,
> the sampling rate must be two times the frecuency of
> the max information component instead of the max
> frecuency afther modulation. If you play with the
> spectrum of the discrete signal you will see that but
> there is a distortion when the sampling is not
> sychronized whith the center of the bandpass signal.
> Anybody know something about this?
> This is interesting because the adquisition system
> migth be less expencive. Any comment is wellcome.
> Bye
> Darien