In particular I am having trouble with 6b).

From what I understand, we can split a difference LTI equation into two sums, the sum of the previous responses, and the sum of the previous inputs. Something like this:

Now to get a zero-state response, I have to zero out the first part of the formula. This I understand.

So to get the zero-state step response, I must use x[n] = u[n] and zero out all past outputs of the system.

The problem I'm having is deriving y[n] in the first place.

I have had a couple ideas, however am not confident in their correctness:

  • Since H(z) = Y(z)/X(z) I could compare the given H(z) to Y(z)(1 - 1/z) because that's the z-transform of 1/u[n]. Correct me if I'm wrong. Then I could make Y subject of the formula and figure out the inverse transform from there.

  • I thought about using convolution with the inverse transform of
    H(z) and input signal u[n].

However I'm not convinced these are correct because nowhere am I "enforcing" the definition of zero-ing out the previous outputs.

Any guidelines would be appreciated!


1 Answer 1



The (zero-state) step response is just the cumulative sum of the impulse response $h[n]$:


This follows in a straightforward manner from the convolution of a unit step $u[n]$ with the impulse response:

$$y_{ZS}[n]=(h\star u)[n]$$

  • $\begingroup$ Would you mind elaborating further on how to obtain the first statement? Particularly I don't understand how this relates to being the "zero state". I understand the use of the convolution sum, but not how it relates to the zero state. Thank you :) $\endgroup$ Commented Nov 5, 2019 at 18:46
  • $\begingroup$ @Novicegrammer: By definition, the impulse response $h[n]$ is the "zero-state" response to an impulse. So any "zero-state" response is obtained by convolving the respective input signal with the impulse response. $\endgroup$
    – Matt L.
    Commented Nov 5, 2019 at 21:47

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.