0
$\begingroup$

Currently I'm having problem wrapping my head around the following.

Suppose you have a dynamical system described by the transfer function $$ G(s)=\frac{as}{(s+b)(s+c)} $$ depending on the variables $a$, $b$ and $c$. In order to calculate the frequency response of the system $s=i\omega$. With that one is now able to draw the Bode plot wherein the magnitude specified by $$ M(\omega)=20\log_{10}\lvert G(i\omega)\rvert $$ is plotted over $\omega$.

The bandwidth of this frequency response should now be defined (as far as I understood it) as the differences of frequencies for which the maximal magnitude is lowered by 3 dB.

Is this the correct definition? Because people mainly show example in which the maximal magnitude is 0 dB and the bandwidth then is calculated as the difference of frequencies for which the magnitude crosses 3 dB.

Would it then be correct to solve $$ 20\log_{10}\lvert G(i\omega)\rvert = 20\log_{10}G_\text{max}-3 \text{ dB} $$ for $\omega$ with $$ \begin{align} 20\log_{10}G_\text{max}&=\frac{d}{d\omega}\left(20\log_{10}\lvert G(i\omega)\rvert\right)\\ G_\text{max}&=\frac{d}{d\omega}\lvert G(i\omega)\rvert \end{align} $$ If so, could someone help me with that? I should be able to do that but seem to be slow on the uptake.

Thanks

$\endgroup$
1
$\begingroup$

The bandwidth of this frequency response should now be defined (as far as I understood it) as the differences of frequencies for which the maximal magnitude is lowered by 3 dB.

Both definitions are used. But I believe this also depends a bit on the field of study. For control engineering we use the cross-over frequency, the bandwidth is the frequency at which the magnitude is equal to 1, i.e., 0 [dB]. This is done because it is related to the stability at the system. Because if the magnitude response $|G(jw)|$ crosses at 0 [dB] and the phase $\angle G(jw)$ is larger then -180 degrees, you know your system is stable, assuming the magnitude does not cross 0 [dB] afterwards anymore. This has to do with Nyquist stability criterion.

However in electrical engineering, specific to filter design, The frequency defined at -3 [dB] is being used as the bandwidth. Because this is the magnitude at which the filter is not active anymore. From wikipedia

https://en.wikipedia.org/wiki/Bandwidth_(signal_processing)#X-dB_bandwidth

In basic electric circuit theory, when studying band-pass and band-reject filters, the bandwidth represents the distance between the two points in the frequency domain where the signal is $\frac{1}{\sqrt{2}}$ of the maximum signal amplitude (half power).

$\endgroup$
  • $\begingroup$ Thanks for clearing that up. So, regarding my example, if $a=b=c=1$ the magnitude is always below -5 dB. Hence, Matlab's bandwidth command results in Inf. $\endgroup$ – Pascal May 27 '16 at 12:30
  • $\begingroup$ From what I understands from Matlabs bandwidth command, nl.mathworks.com/help/control/ref/bandwidth.html, yes that is the reason it results in Inf. $\endgroup$ – WG- May 29 '16 at 20:40
  • $\begingroup$ ps. @Pascal if the answer, answered your question, could you mark it as such that people know your question has been answered. $\endgroup$ – WG- Jun 2 '16 at 15:34

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.