CHAPTER 48 / Sampling and command paths

Discrete poles: from the left half-plane to the unit circle

Sampling turns a continuous pole into a multiplier applied once per sample, and the stable region moves from the left half-plane to the inside of the unit circle.

THE SHORT ANSWER

A discrete pole is the factor by which a sampled mode is multiplied at each sample; sampling a continuous pole s every Ts seconds gives z = exp(sTs). The left half of the s-plane maps inside the unit circle, so a linear discrete system is stable when every pole satisfies |z| < 1. A smaller |z| does not mean faster motion unless the sampling periods match.

z=esTs,∣z∣=eσTs,xk=zkx0z=e^{sT_s},\qquad |z|=e^{\sigma T_s},\qquad x_k=z^k x_0
WHAT YOU WILL LEARN

Map one continuous pole to a discrete pole and compare both over the same physical time.

A digital controller has a sampling period. That period also sets how a continuous exponential mode appears in a sequence of samples.

Review the previous lesson: A controller has a clock
START WITH AN EXAMPLE

First, picture it

Suppose a value keeps half of its size at each tick. Repeated ticks make it shrink. If each tick multiplies it by more than one, it grows instead. A discrete pole is this per-sample multiplier, possibly with a rotation that represents oscillation.

THE IDEA

Build the idea step by step

A continuous pole generates an exponential mode. Sampling that mode every Ts seconds gives the multiplier z below. A magnitude below one means repeated samples decay; above one means they grow. Exactly one is a boundary and does not give decay. For full linear-state stability, every discrete eigenvalue must be strictly inside the unit circle. Compare responses over the same elapsed time, because a different sample period changes how many multiplications happen in a second.

Read the equations

z=esTs,xk=zkx0z=e^{sT_s},\qquad x_k=z^k x_0

For an exponential mode sampled every Ts seconds, z is the multiplier per sample. Its magnitude must be below one for that mode to decay.

What the symbols mean

ss
continuous pole, in inverse seconds
TsT_s
sampling period, seconds
zz
dimensionless discrete pole
x0,xkx_0,x_k
initial mode value and value at sample k

Because zᵏ = e^{s·kTₛ}, the value after k samples depends only on the elapsed time kTₛ, not on how that time was divided into samples.

WORKED EXAMPLE

The same decay on two clocks

  1. Take a continuous pole at minus 2 inverse seconds and a 0.1 second sample period.

    z=e−2(0.1)≈0.8187z=e^{-2(0.1)}\approx0.8187
  2. After ten samples, one second has passed. From initial value 1, the mode is about 0.1353.

    x10=(e−0.2)10=e−2≈0.1353x_{10}=(e^{-0.2})^{10}=e^{-2}\approx0.1353
  3. At a 0.2 second period, z is about 0.6703. Five samples still give the same one-second decay. A smaller discrete pole here does not mean a faster physical system.

    (e−0.4)5=e−2(e^{-0.4})^5=e^{-2}
REFERENCE

Reference tables

Where z = e^(sTs) sends each part of the s-plane
s-planez-planeSampled mode
Re⁡s<0\operatorname{Re}s<0∣z∣<1|z|<1decays
Re⁡s=0\operatorname{Re}s=0∣z∣=1|z|=1neither decays nor grows (boundary)
Re⁡s>0\operatorname{Re}s>0∣z∣>1|z|>1grows
s=0s=0z=1z=1constant
s=±jπ/Tss=\pm j\pi/T_sz=−1z=-1alternates sign every sample
One continuous pole s = −2 s⁻¹ sampled on three clocks
Ts (s)z=e−2Tsz=e^{-2T_s}Samples in 1 sMode after 1 s
0.10.8187100.1353
0.20.670350.1353
0.50.367920.1353

A smaller discrete pole on a longer clock is the same physical decay: after one second every row reaches e⁻² from an initial value of 1.

More detail and model assumptions
A CLOSER LOOK

Compare seconds, not only samples

Record a prediction that |z| decreases, then change T from 0.2 to 0.5 s. The radius changes from about 0.905 to 0.779. Yet the decay envelope after one second is exp(−0.5) in both recordings: the physical mode has not changed. Larger intervals contain more decay per step but fewer steps per second. Pin both traces and inspect the same physical time.

TRY IT

Connect the example to the lab

Keep the same continuous mode and change the sampling period. Does a smaller discrete pole mean faster motion in seconds?

Open this experiment
BE CAREFUL

This mapping samples an existing exponential exactly. It does not redesign a feedback controller or include computation delay. Frequencies that differ by a whole number of cycles per sample can produce the same samples, which leads to the next lesson.

Check your understanding+

Does z=0.6703 at a 0.2-second period decay faster in real time than z=0.8187 at 0.1 seconds in this example?

No. Both sample the same continuous pole −2. Five longer steps and ten shorter steps each give exp(−2), about 0.1353, after one second.
PRACTICE PROBLEMS

Solve these before opening the solutions

Work each problem on paper first. Open the answer to check yours, then the full solution if you need the steps, and mark how it went.

  1. A continuous mode has pole s = −5 s⁻¹ and is sampled every 0.05 s. Find the discrete pole z and the value after 10 samples from x₀ = 1. Check the result against the continuous mode at the same time.

    Show answer

    Answerz = e^(−0.25) ≈ 0.779. After 10 samples, which is 0.5 s, x₁₀ ≈ 0.0821, the same as e^(−5 × 0.5).

    z=e−0.25≈0.779,x10=e−2.5≈0.0821z=e^{-0.25}\approx0.779,\qquad x_{10}=e^{-2.5}\approx0.0821
    Show full solution

    Solution

    1. Map the pole.

      z=esTs=e−5(0.05)=e−0.25≈0.779z=e^{sT_s}=e^{-5(0.05)}=e^{-0.25}\approx0.779
    2. Ten multiplications.

      x10=z10=e−2.5≈0.0821x_{10}=z^{10}=e^{-2.5}\approx0.0821
    3. Ten samples span 0.5 s, and the continuous mode at 0.5 s is e^(−2.5). Sampling a mode exactly reproduces the continuous values at the sample instants.

  2. A lightly damped continuous mode has poles s = −1 ± j4 and is sampled every 0.1 s. Find the discrete poles in polar and rectangular form. How many samples cover one oscillation period?

    Show answer

    Answer|z| = e^(−0.1) ≈ 0.905 and the angle is ±0.4 rad (±22.9°), so z ≈ 0.833 ± j0.352. One period of 2π/4 ≈ 1.57 s covers about 15.7 samples.

    z=e−0.1e±j0.4≈0.833±j0.352z=e^{-0.1}e^{\pm j0.4}\approx0.833\pm j0.352
    Show full solution

    Solution

    1. The real part sets the radius; the imaginary part sets the angle.

      z=e(−1±j4)(0.1)=e−0.1e±j0.4z=e^{(-1\pm j4)(0.1)}=e^{-0.1}e^{\pm j0.4}
    2. Polar form.

      ∣z∣=e−0.1≈0.905,∠z=±0.4 rad≈±22.9∘|z|=e^{-0.1}\approx0.905,\qquad \angle z=\pm0.4\ \mathrm{rad}\approx\pm22.9^\circ
    3. Rectangular form.

      z≈0.905(cos⁡0.4±jsin⁡0.4)≈0.833±j0.352z\approx0.905(\cos0.4\pm j\sin0.4)\approx0.833\pm j0.352
    4. The mode advances 0.4 rad per sample, so a full turn of 2π takes about 15.7 samples.

      2π0.4≈15.7\frac{2\pi}{0.4}\approx15.7
  3. A sampled loop running at Ts = 20 ms has a real discrete pole z = 0.6. Find the equivalent continuous pole and time constant. What does a pole at z = −0.6 mean instead?

    Show answer

    Answers = ln(0.6)/0.02 ≈ −25.5 s⁻¹, a time constant of about 39 ms. z = −0.6 decays at the same rate per sample but flips sign every sample: it behaves like a mode oscillating at the Nyquist frequency π/Ts ≈ 157 rad/s.

    s=ln⁡0.60.02≈−25.5 s−1,z=−0.6: s≈−25.5+j157s=\frac{\ln0.6}{0.02}\approx-25.5\ \mathrm{s^{-1}},\qquad z=-0.6:\ s\approx-25.5+j157
    Show full solution

    Solution

    1. Invert the mapping.

      s=ln⁡zTs=ln⁡0.60.02≈−25.5 s−1,τ=−1s≈39 mss=\frac{\ln z}{T_s}=\frac{\ln0.6}{0.02}\approx-25.5\ \mathrm{s^{-1}},\qquad \tau=-\frac1s\approx39\ \mathrm{ms}
    2. A negative real z has angle π, so its samples alternate in sign.

      z=−0.6=0.6ejπ,xk=(−0.6)kx0: 1, −0.6, 0.36, −0.216, …z=-0.6=0.6e^{j\pi},\qquad x_k=(-0.6)^kx_0:\ 1,\ -0.6,\ 0.36,\ -0.216,\ \ldots
    3. Equivalent continuous pole.

      s=ln⁡0.6+jπ0.02≈−25.5+j157s=\frac{\ln0.6+j\pi}{0.02}\approx-25.5+j157
    4. No real first-order continuous mode produces alternating samples. A negative real discrete pole usually comes from the controller or the sampling itself, for example a gain that overcorrects every period.

  4. Exam level. A first-order plant sampled through a zero-order hold at Ts = 0.1 s gives x_{k+1} = 0.9x_k + 0.1u_k. A digital P controller uses u_k = −Kx_k. (a) Find the closed-loop pole and the range of K for which the loop is stable. (b) Which K returns the state to zero in one sample? (c) For K = 4, find the equivalent continuous pole and the time to fall below 1% of the initial value. (d) Compare K = 4 with K = 14.

    Show answer

    Answer(a) z = 0.9 − 0.1K; the loop is stable for −1 < K < 19. (b) K = 9 gives z = 0, a one-step (deadbeat) response. (c) z = 0.5 and s ≈ −6.93 s⁻¹; 7 samples, or 0.7 s, are needed. (d) K = 14 gives z = −0.5: the same decay per sample, but the sign alternates every period.

    z=0.9−0.1K,−1<K<19,K=9: z=0,K=4: s≈−6.93, 0.7 sz=0.9-0.1K,\quad -1<K<19,\quad K=9:\ z=0,\quad K=4:\ s\approx-6.93,\ 0.7\ \mathrm s
    Show full solution

    Solution

    1. Closed-loop multiplier and stability range.

      xk+1=(0.9−0.1K)xk,∣0.9−0.1K∣<1 ⇒ −1<K<19x_{k+1}=(0.9-0.1K)x_k,\qquad |0.9-0.1K|<1\ \Rightarrow\ -1<K<19
    2. One-step response.

      0.9−0.1K=0 ⇒ K=90.9-0.1K=0\ \Rightarrow\ K=9
    3. Equivalent continuous pole for K = 4.

      K=4: z=0.5,s=ln⁡0.50.1≈−6.93 s−1K=4:\ z=0.5,\qquad s=\frac{\ln0.5}{0.1}\approx-6.93\ \mathrm{s^{-1}}
    4. Count the samples needed to fall below 1%.

      0.5k<0.01 ⇒ k>ln⁡0.01ln⁡0.5≈6.64 ⇒ k=7,t=0.7 s0.5^k<0.01\ \Rightarrow\ k>\frac{\ln0.01}{\ln0.5}\approx6.64\ \Rightarrow\ k=7,\quad t=0.7\ \mathrm s
    5. K = 14 gives z = −0.5, the same magnitude as 0.5, so the envelope decays equally fast, but each sample overcorrects past zero. Stability depends on |z|; the sign and angle decide what the samples look like.

Mark each exercise after you check your answer. Marks stay in this browser.

KEY TERMS

Terms this lecture teaches

All terms in the glossary →
QUESTIONS STUDENTS ASK

Frequently asked questions

Why is the unit circle the stability boundary in the z-plane?

A sampled mode is multiplied by z at every sample, so after k samples its size is scaled by |z| to the power k. It shrinks when |z| < 1, grows when |z| > 1 and does not decay at |z| = 1; the circle is the image of the imaginary axis under z = exp(sTs).

How do I compare discrete poles taken at different sampling periods?

Convert each back to a continuous rate with s = ln(z)/Ts, or compare responses over the same elapsed time. A mode with σ = −0.5 s⁻¹ has |z| ≈ 0.905 at Ts = 0.2 s and ≈ 0.779 at Ts = 0.5 s, yet both envelopes reach exp(−0.5) ≈ 0.607 after one second.

What does a pole on the negative real axis of the z-plane mean?

Its mode changes sign at every sample, so the sampled response alternates between positive and negative values. It decays if the pole lies between −1 and 0 and grows in magnitude if it lies beyond −1.

Does mapping poles with z = exp(sTs) design a digital controller?

No. It samples an existing exponential mode exactly, but it does not include computation delay or redesign the feedback. Continuous frequencies that differ by whole cycles per sample also map to the same z, which is aliasing.

Why the next lesson follows

Pole locations describe one mode at a time. The z-transform turns whole difference equations into algebra, the way Laplace did for differential equations.

YOUR PROGRESS

Finished the reading and the exercises? Mark the lecture as done to tick it on the syllabus. Progress stays in this browser.

EIGENROOM · https://eigenroom.com/theory/discrete-poles · Updated 2026-11-15