Sequence	Transform	ROC
delta[n]	1	all z
u[n]	z/(z-1)	\|z\| > 1
a^n u[n]	z/(z-a)	\|z\| > \|a\|
-a^n u[-n-1]	z/(z-a)	\|z\| < \|a\|
n a^n u[n]	az/(z-a)^2	\|z\| > \|a\|
cos(w n)u[n]	z(z-cos w)/(z^2-2z cos w+1)	\|z\| > 1

Z-Transforms for Discrete-Time Systems

The Z-transform is the discrete-time counterpart to the Laplace transform. It converts sequences indexed by n into functions of the complex variable z. Digital filters, sampled-data control systems, difference equations, and DSP algorithms are often analyzed in the z-domain because delays become powers of z inverse and convolution becomes multiplication. This makes discrete-time systems easier to design, reason about, and implement.

A Z-transform table lists common sequence-transform pairs. The unit impulse delta[n] transforms to 1. The unit step u[n] transforms to z/(z-1) with region of convergence outside the unit circle. The right-sided exponential a^n u[n] transforms to z/(z-a) with region |z| greater than |a|. These pairs appear constantly when solving linear difference equations or deriving digital filter responses.

Manual Transform Use

To use the table, identify the discrete sequence and whether it is right-sided, left-sided, or finite length. Then match the sequence to a transform pair and region of convergence. For example, a causal first-order IIR filter has an impulse response proportional to a^n u[n]. Its transform contains a pole at z = a. If |a| is less than one, the pole is inside the unit circle and the causal system is stable. If |a| is greater than or equal to one, the response does not decay sufficiently for bounded-input bounded-output stability.

Delay is especially simple in the z-domain. A one-sample delay corresponds to multiplication by z^-1. Difference equations such as y[n] = x[n] + a y[n-1] become algebraic equations involving z^-1. Solving for Y(z)/X(z) gives the transfer function. That transfer function can be evaluated on the unit circle z = e^(jw) to obtain frequency response.

Region of Convergence

The region of convergence, or ROC, is not optional. The same algebraic expression can correspond to different time-domain sequences depending on the ROC. z/(z-a) with ROC outside |a| corresponds to a right-sided causal exponential. With ROC inside |a|, it corresponds to a left-sided sequence. Stability and causality depend on the ROC as well as the pole locations. A table that omits ROC hides essential information.

For causal rational systems, the ROC lies outside the outermost pole. For stable systems, the ROC includes the unit circle. Therefore, a causal stable IIR filter must have all poles inside the unit circle. FIR filters have finite impulse responses and are inherently stable under ordinary coefficient assumptions, but their zeros still shape frequency response. Z-transform reasoning connects time-domain recursion to pole-zero geometry.

Engineering Applications

Z-transforms are used in digital audio, communications, motor control, sensor filtering, control loops, embedded DSP, and sampled measurement systems. A moving average filter, biquad filter, PID controller, decimator, or predictor can be described with a difference equation and analyzed in the z-domain. Engineers use pole-zero plots to understand resonance, damping, notch behavior, and stability margin.

In firmware, the z-domain connects directly to implementation. A factor of z^-1 is a stored previous sample. A denominator term creates feedback. A numerator term creates feedforward. Coefficients become multiply-accumulate operations. Quantization, overflow, and sample-rate choices then determine how closely the implementation matches the math. The table helps bridge the symbolic design and the code that runs on hardware.

Use this reference to identify common sequence pairs and remember the role of ROC. For serious filter design, combine table lookup with frequency-response plots, coefficient quantization analysis, and time-domain simulation. The transform is the map; testing shows how the actual implementation behaves.

Manual inverse Z-transforms often use algebraic manipulation, long division, or partial fraction expansion. A transfer function can be expanded into powers of z inverse to reveal impulse-response samples directly. For IIR systems, partial fractions expose exponential modes tied to pole locations. For FIR systems, the numerator coefficients often correspond directly to feedforward taps. Moving between these views helps engineers connect equations, code, and measured impulse responses.

Sampling period matters. A pole location in the z-plane corresponds to a discrete-time behavior at a particular sample rate. Changing sample rate without redesigning coefficients changes the relationship between digital poles and physical time constants. This is why filter coefficients should be documented with sample rate and why firmware that supports multiple rates often needs separate coefficient sets or a runtime design step.

Fixed-point implementation adds another layer. Coefficients may be quantized, internal states may overflow, and feedback can create limit cycles. A mathematically stable floating-point filter may behave poorly on a small microcontroller if scaling is wrong. Z-transform analysis identifies the intended system; implementation review ensures the deployed system follows it closely enough.

Study Notes

Z-Transform Table works best when the article is read as a chain of ideas: Manual Transform Use, Region of Convergence, Engineering Applications. In Z-Transform, that chain explains the assumptions behind equations, domains, variables, units, vectors, matrices, or data sets represented by sequence, causality, transform pair, and region of convergence. The Z-Transform inputs are sequence, causality, transform pair, and region of convergence, and they should be connected to the specific problem before the output is treated as meaningful.

For Z-Transform, build one small example with numbers simple enough to check by hand, then change one input and explain why the output moved. Next, change one Z-Transform input from this list: sequence, causality, transform pair, and region of convergence. Predict the direction of the change before recalculating, especially because Z-Transform mistakes often come from applying the formula before checking domain, sign, units, order of operations, or the meaning of each variable.

A strong homework or lab note for Z-Transform should record the governing relationship, domain or unit assumptions, one intermediate step, and the way sequence, causality, transform pair, and region of convergence enter the final result. If Z-Transform Table disagrees with a later hand calculation or lab observation, those Z-Transform notes make it easier to locate whether the mismatch came from arithmetic, convention, measurement setup, or an input entered in the wrong form.