Skip to main content

Short note: Obtaining a pulse transfer function

Short answer

Draw the block diagram with all samplers and name signals. Write s‑domain equations, then take starred transforms noting that blocks not separated by a sampler must be combined as (G₁G₂)*. Obtain pulse transfer functions via partial fractions or the residue formula; with a ZOH use G(z) = (1−z⁻¹)Z[Gₚ/s]. Solve the algebraic equations for C(z)/R(z); if a continuous block precedes the first sampler, only C(z) can be found.

Write a short note on general procedure for obtaining pulse transfer function.

The pulse transfer function G(z)=Y(z)/X(z)G(z) = Y(z)/X(z) relates the z-transform of the sampled output to the z-transform of the sampled input, with zero initial conditions. The general procedure is:

  1. Draw the block diagram with all samplers. Name the input of each sampler (for example EE, UU). Name the outputs as starred signals E∗E^*, U∗U^*. Add a fictitious sampler at the output if the output is continuous.
  2. Write the s-domain equations. Express each sampler input and the output in terms of the starred signals and the input R(s)R(s), e.g. U(s)=G1(s)E∗(s)U(s) = G_1(s)E^*(s).
  3. Take the starred transform of each equation. Use [G(s)X∗(s)]∗=G∗(s)X∗(s)[G(s)X^*(s)]^* = G^*(s)X^*(s). Elements not separated by a sampler must be combined first: [G1G2X∗]∗=(G1G2)∗X∗[G_1G_2X^*]^* = (G_1G_2)^*X^*, and this is not equal to G1∗G2∗X∗G_1^*G_2^*X^*.
  4. Find each pulse transfer function using partial fractions or residues: G(z)=Z[G(s)]=∑Res[G(s)zz−eTs]G(z) = \mathcal{Z}\big[G(s)\big] = \sum \text{Res}\left[G(s)\frac{z}{z-e^{Ts}}\right] With a ZOH, G(z)=(1−z−1) Z[Gp(s)/s]G(z) = (1-z^{-1})\,\mathcal{Z}\left[G_p(s)/s\right].
  5. Solve the algebraic equations for C(z)C(z) and form C(z)/R(z)C(z)/R(z). If R(s)R(s) enters through a continuous block before the first sampler (as GR(z)GR(z)), only C(z)C(z) can be found. In that case no pulse transfer function exists.

Example: Unity feedback with a sampler on the error, a ZOH and plant Gp(s)G_p(s). Then E∗=R∗−(GhGp)∗E∗E^* = R^* - (G_hG_p)^*E^*, which gives C(z)R(z)=G(z)1+G(z)\dfrac{C(z)}{R(z)} = \dfrac{G(z)}{1+G(z)} with G(z)=(1−z−1)Z[Gp/s]G(z) = (1-z^{-1})\mathcal{Z}[G_p/s].

✓ Answer checked by Claude Opus 5.5 (high effort)

All Digital Control System chapter 3 questions →