0
$\begingroup$

enter image description here

I found something very interersting when I was trying to simulate a discrete time system. So first I discretized the continuous system $G(s)$ using "c2d" function and "dbode". The resulting bode plot for the discretized system is displaying some sort of periodic behaviour after reaching about 3 hz. So is the phase plot.

Can someone explain why this maybe happening?

$\endgroup$
  • $\begingroup$ looks like the sample rate is 1 Hz (which is $2\pi$ rad/s). sampled spectra is inherently periodic with period equal to the sample rate. $\endgroup$ – robert bristow-johnson Apr 6 '15 at 18:53
2
$\begingroup$

When you are discretizing you make transformation $s \to e^{s T}$. This transforms the left half plane into unit disc. But actually it only transforms the $(-j\omega_s/2, j\omega_s/2)$ band and repeats itself by circling for outside of the band, where $\omega_s = 2 \pi / T$. This is because of the periodicity of exponential function.

When you plot the frequency response, you plot $|G(j\omega)|=|\hat{G}(e^{j\omega T})|$ for all $\omega \in [0, \infty)$ where $\hat{G}$ is the discrete transfer function. Therefore the frequency response of any sampled system is periodic.

$\endgroup$
0
$\begingroup$

My guess is that you're looking at $\pi$ radians/second rather than 3 Hz.

Look at the $x$ axis. It is labeled as rad/s.

And $\pi$ radians/second corresponds to $\frac{f_s}{2}$, the maximum (baseband) frequency the system can deal with. Above that frequency, the spectrum / frequency response will start to repeat.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.