Z-Transform Region of Convergence

Introduction

The Region of Convergence is not merely a mathematical technicality — it is the key that resolves the fundamental ambiguity of the Z-Transform. Without specifying the ROC, a given algebraic expression like $\frac{z}{z-a}$ could correspond to infinitely many different time-domain sequences. The ROC tells you which sequence you are actually dealing with, determines whether the system is causal, and reveals whether it is stable. In examination problems and engineering practice alike, neglecting the ROC leads to incorrect answers.

Think of it this way: the Z-Transform definition involves an infinite series. That series does not converge for all values of $z$ — it only converges in certain regions of the complex plane. The ROC specifies exactly where the series "makes sense" mathematically, and this geometric region carries direct physical information about the signal.

Mathematical Definition

The ROC is the set of all complex values $z$ for which the Z-Transform sum converges absolutely:

$$\text{ROC} = \{z \in \mathbb{C} : \sum_{n=-\infty}^{\infty} |x[n]| \cdot |z|^{-n} < \infty\}$$

Since the convergence depends only on $|z|$ (not on the angle of $z$), the ROC always takes the form of an annular (ring-shaped) region centered at the origin.

Derivation for a Simple Example

Consider $x[n] = a^n u[n]$ (causal exponential):

$$X(z) = \sum_{n=0}^{\infty} a^n z^{-n} = \sum_{n=0}^{\infty}\left(\frac{a}{z}\right)^n$$

This geometric series converges when $|a/z| < 1$, i.e., $|z| > |a|$.

Therefore: $X(z) = \frac{1}{1-az^{-1}} = \frac{z}{z-a}$ with ROC: $|z| > |a|$

The ROC is the exterior of a circle of radius $|a|$ in the z-plane.

Properties of the ROC

Property 1: Annular Shape

The ROC is always an annular region of the form $r_1 < |z| < r_2$. It can extend to include $|z| = 0$ or $|z| = \infty$ as special cases:

• Right-sided signals: $|z| > r_{\max}$ (extends to $\infty$)
• Left-sided signals: $|z| < r_{\min}$ (extends to $0$)
• Two-sided signals: $r_1 < |z| < r_2$ (bounded annulus)

Property 2: No Poles in the ROC

Poles are points where $X(z) \to \infty$, so the series obviously cannot converge there. Poles always lie on the boundary of the ROC, never inside it.

Property 3: Finite-Duration Signals

If $x[n]$ is non-zero only for $N_1 \leq n \leq N_2$ (finite duration), then the ROC is the entire z-plane, possibly excluding $z = 0$ (if $N_2 > 0$) or $z = \infty$ (if $N_1 < 0$).

Property 4: Right-Sided Signals

If $x[n] = 0$ for $n < N_1$ (right-sided), the ROC extends outward from the outermost pole: $|z| > r_{\max}$. This includes $z = \infty$.

Property 5: Left-Sided Signals

If $x[n] = 0$ for $n > N_2$ (left-sided), the ROC extends inward from the innermost pole: $|z| < r_{\min}$. This includes $z = 0$.

Property 6: Two-Sided Signals

For two-sided signals, the ROC is an annulus bounded by adjacent poles: $r_1 < |z| < r_2$. If no such annulus exists (poles at the same radius), the bilateral Z-Transform does not exist.

ROC and System Properties

The ROC carries direct information about fundamental system properties:

Why These Relationships Hold

Causality and $z = \infty$: A causal signal has $x[n] = 0$ for $n < 0$, so $X(z) = \sum_{n=0}^{\infty}x[n]z^{-n}$. As $|z| \to \infty$, every term $x[n]z^{-n} \to 0$ (for $n \geq 1$), guaranteeing convergence. Therefore, the ROC of any causal signal always includes $z = \infty$.

Stability and the unit circle: BIBO stability requires $\sum_{n=-\infty}^{\infty}|h[n]| < \infty$ (absolutely summable impulse response). The Z-Transform evaluated on the unit circle is $H(e^{j\Omega}) = \sum h[n]e^{-j\Omega n}$. This sum converges absolutely precisely when $\sum|h[n]| < \infty$. Therefore, stability is equivalent to the unit circle lying within the ROC.

Causal + Stable: If the system is causal, ROC is $|z| > r_{\max}$ where $r_{\max}$ is the largest pole magnitude. For the unit circle to be in this region, we need $r_{\max} < 1$, meaning all poles must be inside the unit circle.

The Unit Circle Test: Worked Examples

Example 1: $H(z) = \frac{1}{1-2z^{-1}}$, pole at $z = 2$.

• If causal (ROC: $|z| > 2$): unit circle ($|z|=1$) is NOT in ROC → unstable
• If anti-causal (ROC: $|z| < 2$): unit circle IS in ROC → stable (but non-causal)

Example 2: $H(z) = \frac{1}{1-0.5z^{-1}}$, pole at $z = 0.5$.

• If causal (ROC: $|z| > 0.5$): unit circle IS in ROC → stable and causal ✓

Example 3: $H(z) = \frac{1}{(1-0.3z^{-1})(1-1.5z^{-1})}$, poles at $z = 0.3$ and $z = 1.5$.

• If causal (ROC: $|z| > 1.5$): unit circle NOT in ROC → causal but unstable
• If two-sided (ROC: $0.3 < |z| < 1.5$): unit circle IS in ROC → stable but non-causal
• You cannot have both causal AND stable here because one pole is outside the unit circle.

Graphical Representation

On the z-plane, the ROC is visualized as a shaded region:

• Poles are marked with × symbols
• Zeros are marked with ○ symbols
• The unit circle is drawn as a dashed circle at $|z| = 1$
• The ROC is shaded, bounded by circles passing through the poles

For a causal stable system, the pole-zero plot shows all × marks inside the unit circle, with the ROC being the entire region outside the largest-radius pole (which is still inside the unit circle, so the unit circle is included).

Key Takeaways

• The ROC is an annular region in the z-plane, determined by the magnitudes of the poles
• Right-sided (causal) signals have ROC extending outward: $|z| > r_{\max}$
• Left-sided (anti-causal) signals have ROC extending inward: $|z| < r_{\min}$
• BIBO stability requires the unit circle to be within the ROC
• Causality requires $z = \infty$ to be in the ROC
• A causal system is stable if and only if ALL poles lie inside the unit circle
• The same algebraic $X(z)$ with different ROCs represents entirely different sequences
• Always specify the ROC when presenting a Z-Transform — it is not optional information