Convergence
The Z-transform is only defined where its infinite sum converges. The Region of Convergence encodes the signal type, makes the inverse transform unique, and determines whether a system is stable.
Two completely different sequences can share the same X(z) expression. Without the ROC, the inverse Z-transform is ambiguous.
The ROC is the set of z values where the sum is finite. Since |z⁻ⁿ| = r⁻ⁿ with r = |z|, convergence depends only on the magnitude of z — making the ROC an annulus.
For causal sequences, convergence requires |z| to be large enough. ROC is the exterior of a circle.
- Pole at z = a sits on boundary, never inside ROC
- Causal FIR: ROC includes z = ∞
- ROC never contains any pole of X(z)
For anti-causal sequences, convergence requires |z| to be small enough. The same rational X(z), a completely different sequence.
Two-sided sequences must converge in both directions simultaneously — requiring |z| to be neither too large nor too small.
A system is BIBO stable if and only if the unit circle lies inside the ROC of H(z). For causal systems, this means all poles inside |z| = 1.
- ROC = set of z where X(z) sum converges — always an annulus
- ROC never contains poles of X(z)
- Right-sided (causal): ROC = |z| > r₁
- Left-sided (anti-causal): ROC = |z| < r₂
- Two-sided: ROC = r₁ < |z| < r₂ (annular)
- Same X(z) expression → different signals depending on ROC
- BIBO stable ⟺ unit circle inside ROC ⟺ all causal poles inside unit circle