Given the input $$e(kT_s) = {0.4, 0.8,1.2, - 0.9}, \quad k = 0,1,2,3$$

and transfer function $$T\left(z^{-1}\right)=\frac {z^{-1}-0.8z^{-2}}{1-1.1z^{-1}+0.3}$$ Find the output $Y(kT_s),\quad k=0,1,2,3$

(I don't know why $0.3$ and $1$ haven't been added).

Starting with the inverse transform I found $$T(k)=\frac{1}{1.3}\left(\frac{1.1}{1.3}\right)^{k-1}u(k-1)-\frac{0.8}{1.3}\left(\frac{1.1}{1.3}\right)^{k-2}u(k-2)$$

I'm not sure if that is correct as I'm not so familiar with $\mathcal Z$-transform. Anyway, I don't have a sampling period in $T(k)$. Do I just put a $T_s$ next to $k$? Do I ignore $T_s$ and do the following? $$Y(0)=T(0)E(0)$$

$T(0)=0$ so $Y(0)=0$, but why am I given $0.4$ for $E(0)$? This has to be wrong. $$Y(1)=T(1)E(1)=\frac{1}{1.3}0.8=10.4 $$ etc etc.


1 Answer 1


There should probably be a $z^{-2}$ term with the $0.3$. It looks like a typo.

Once you've transformed to the $z$ domain (or discrete-time), and are not interested in transforming back to the continuous-time domain, there is no real need to carry the $T_s$. Just refer to $e[k], y[k]$, etc.

  • $\begingroup$ In my solution I inverse z transform though so I'm no longer in the z-domain . Except if y[k] , e[k] etc are still in the z domain and I haven't understood something. $\endgroup$ Jun 28, 2017 at 13:56
  • $\begingroup$ @JohnKatsantas : Right: if you're in the $z$ domain OR in discrete-time (the transform domain). Once you've sampled the signal and got the scale right, the sampling frequency doesn't enter into $z$-transform / discrete-time manipulations (except for some scale factors). $\endgroup$
    – Peter K.
    Jun 28, 2017 at 14:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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