I am currently studying two Butterworth and Chebyshev low-pass filters of order $n =3$ and $n=2$ respectively, whcih are in fact two prototypes to make a bandpass filter. The transfer function that I have is the following one: $$H(j\omega) = 2\sqrt{\frac{R_G}{R_L}}\frac{1}{R_G(C+\frac{D}{R_L}) + A + \frac{B}{R_L}}$$ where $R_G$ is the impedance in series with the generator and $R_L$ is the load impedance

For the Butterworth filter the data is the following: $$\begin{align}R_G&=R_L=50\;\Omega\\ A &=1 -0.708\omega^2\\ B &=j59\omega \\ C &=j(0.024\omega -0.008496\omega^3)\\ D &= 1-0.708\omega^2\end{align}$$

and for the Chebyshev filter the data is the following: $$\begin{align}R_G&=50\Omega\quad R_L=18.8\;\Omega\\ A &=1\\ B &=j34.25\omega \\ C &=j0.036\omega\\ D &=1-1.233\omega^2\end{align}$$ and this is the poles diagram computed with the data above of such circuits: enter image description here

where: $$\begin{align} \omega_{pB_1} &= j1.66\\ \omega_{pB_2} &= -1.45+j0.83\\ \omega_{pB_3} &= 1.45+j0.83\\ \omega_{pC_1} &= -0.9 + j0.55\\ \omega_{pC_2} &= 0.9 + j0.55 \end{align}$$

It is clear that both systems are going to be unstable since they have poles in the right-half plane, but how will these poles look in the time domain? I have read articles like this one or this one and some questions on this website, and I know that if they were complex conjugated poles the response in time domain would be a sinusoid type signal.

I am asking this because the majority of the examples have two conjugated poles, which in this case they are not, and hence the time domain expressions won't be sinusoids, or I am wrong? Could someone answer with the inverse Laplace transform in Matlab of my transfer function?

Note that I am not asking for my job to be done, I have already made it, but some more information about those poles will be appreciated.

Thank you.


I think you calculated the poles incorrectly, namely for $\omega$ instead of the Laplace variable $s=j\,\omega$. You can find the appropriate poles in terms of $s$ by multiplying your poles by $j$. This will give you conjugate poles, with negative real parts.

  • $\begingroup$ I have computed them with Matlab and checked a thousand times, but maybe I am misleading some theory, in fact! So you are telling me that I have plotted them with the imaginary axis as $\omega$ and not $j\omega$? $\endgroup$
    – Martín
    May 2 '18 at 23:12

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.