Key Concepts of Transfer Functions

Why This Matters

Transfer functions transform messy differential equations into algebraic expressions that reveal how a system behaves. Whether you're analyzing drone stabilization, thermostat response, or amplifier distortion at certain frequencies, you're working with transfer functions. The concepts here, including poles, zeros, damping, frequency response, and stability, form the analytical backbone of control theory.

You'll be tested on your ability to connect mathematical representations to physical behavior. An exam won't just ask you to define a pole; it'll ask you to predict what happens to system response when a pole moves. Don't just memorize formulas. Know what each parameter controls and how changes in the s-domain translate to real-world performance.

---

The Mathematical Foundation

Transfer functions exist because the Laplace transform converts calculus problems into algebra problems. By transforming differential equations into polynomial ratios, you can analyze system behavior using algebraic tools instead of solving integrals.

Definition of Transfer Function

A transfer function is the ratio of output to input in the s-domain, expressed as:

$H(s) = \frac{Y(s)}{X(s)}$

This applies only to linear time-invariant (LTI) systems and assumes zero initial conditions. That assumption is what lets the transfer function characterize the system itself, independent of any particular starting state. From this single expression, you can extract stability, transient response, and frequency behavior.

Laplace Transform and Its Role

The Laplace transform converts time-domain signals to the s-domain by replacing derivatives with powers of $s$ and integrals with division by $s$. This turns differential equations into algebraic equations you can manipulate with polynomial arithmetic.

The complex variable $s = \sigma + j\omega$ carries two pieces of information:

• $\sigma$ (the real part) represents exponential decay or growth
• $\omega$ (the imaginary part) represents oscillation frequency

---

Poles, Zeros, and System Character

The roots of the numerator and denominator polynomials determine everything about how a system responds. Poles dictate stability and natural modes; zeros shape the frequency response.

Poles and Zeros

• Poles are roots of the denominator: values of $s$ where $H(s) \to \infty$. Each pole contributes a natural mode (an exponential or oscillatory term) to the system's response.
• Zeros are roots of the numerator: values of $s$ where $H(s) = 0$. Zeros block or attenuate certain frequency components from reaching the output.
• Location in the complex plane matters most. Left half-plane poles mean stability. Right half-plane poles mean the response grows without bound. Poles on the imaginary axis produce sustained oscillations.

First-Order Systems

The standard transfer function form is:

$H(s) = \frac{K}{\tau s + 1}$

Here $K$ is the DC gain (the steady-state output for a unit step input) and $\tau$ is the time constant. The system has a single real pole at $s = -1/\tau$, which is always in the left half-plane when $\tau > 0$, guaranteeing stability.

The time constant $\tau$ sets the response speed. After one time constant, the step response reaches about 63% of its final value. After $5\tau$ seconds, it reaches approximately 99%.

Second-Order Systems

The standard form is:

$H(s) = \frac{K\omega_n^2}{s^2 + 2\zeta\omega_n s + \omega_n^2}$

Two parameters control the behavior:

• $\omega_n$ (natural frequency): how fast the system tends to oscillate
• $\zeta$ (damping ratio): how quickly oscillations die out

The damping ratio determines the response character:

• Underdamped ($\zeta < 1$): complex conjugate poles, oscillatory response
• Critically damped ($\zeta = 1$): repeated real poles, fastest response with no overshoot
• Overdamped ($\zeta > 1$): two distinct real poles, sluggish with no oscillation

For underdamped systems, the actual oscillation frequency is the damped natural frequency: $\omega_d = \omega_n\sqrt{1 - \zeta^2}$.

---

Time Domain Response

How a system behaves over time after receiving an input is characterized by specific performance metrics. These metrics translate directly to design specifications in real applications.

Time Domain Response Characteristics

• Rise time ($t_r$) measures speed: the time to go from 10% to 90% of the final value. It's inversely related to bandwidth, so a faster system has a wider bandwidth.
• Overshoot ($M_p$) measures how far the response exceeds the final value. For a second-order system, it depends only on $\zeta$: $M_p = e^{-\pi\zeta / \sqrt{1 - \zeta^2}} \times 100\%$.
• Settling time ($t_s$) is the time for the response to stay within a specified band (typically 2%) of the final value. For a second-order system, $t_s \approx \frac{4}{\zeta \omega_n}$ using the 2% criterion.
• Steady-state error reveals how well the system tracks its input. The number of integrators in the open-loop transfer function (poles at $s = 0$) determines the system type, which dictates tracking ability for step, ramp, and parabolic inputs.

---

Frequency Domain Analysis

Examining how systems respond to sinusoids at different frequencies reveals bandwidth, resonance, and stability margins. The frequency response is the transfer function evaluated along the imaginary axis: $H(j\omega)$.

Frequency Domain Analysis Fundamentals

For any sinusoidal input at frequency $\omega$, an LTI system produces a sinusoidal output at the same frequency but with a different magnitude and phase. These are captured by:

• Magnitude: $|H(j\omega)|$, which tells you how much the system amplifies or attenuates at that frequency
• Phase: $\angle H(j\omega)$, which tells you the time shift introduced

Bandwidth defines the useful frequency range, typically where gain drops to $-3$ dB (a factor of $1/\sqrt{2}$) below its maximum or DC value. Systems with low damping ($\zeta < 0.707$) exhibit a resonance peak, meaning they amplify inputs near $\omega_n$.

Bode Plots

Bode plots consist of two graphs: magnitude (in dB) and phase (in degrees), both plotted against log frequency. The logarithmic scaling lets you see behavior across many decades of frequency at once.

Sketching Bode plots by hand relies on asymptotic approximations:

1. Each real pole at frequency $\omega_p$ contributes a $-20$ dB/decade slope starting at $\omega_p$
2. Each real zero at frequency $\omega_z$ contributes a $+20$ dB/decade slope starting at $\omega_z$
3. Complex conjugate poles or zeros contribute $\pm 40$ dB/decade slopes
4. Sum the individual contributions to get the total magnitude and phase curves

Gain margin and phase margin are read directly from Bode plots and are critical for assessing closed-loop stability before building hardware. Phase margin is measured at the gain crossover frequency (where $|H(j\omega)| = 0$ dB), and gain margin is measured at the phase crossover frequency (where $\angle H(j\omega) = -180°$). Positive margins mean the closed-loop system is stable.

---

Stability and System Representation

A system is BIBO stable (bounded input, bounded output) if every bounded input produces a bounded output. Mathematically, this requires all closed-loop poles to lie in the left half-plane.

Stability Analysis Using Transfer Functions

• Left half-plane poles guarantee stability: the system returns to equilibrium after disturbances.
• Right half-plane poles cause instability: responses grow without bound.
• Imaginary axis poles produce marginally stable behavior: sustained oscillations that neither grow nor decay.

The Routh-Hurwitz criterion lets you check whether any poles are in the right half-plane without actually solving for the roots. You construct a table from the coefficients of the characteristic polynomial and check whether all entries in the first column are positive. Any sign change in that column indicates a right half-plane pole.

Block Diagram Representation

Block diagrams decompose complex systems into interconnected transfer functions. Each block represents a transfer function, and arrows show signal flow.

Three fundamental connection rules:

• Series (cascade): $H_{total}(s) = H_1(s) \cdot H_2(s)$
• Parallel: $H_{total}(s) = H_1(s) + H_2(s)$
• Negative feedback loop: $H_{total}(s) = \frac{G(s)}{1 + G(s)H(s)}$, where $G(s)$ is the forward path and $H(s)$ is the feedback path

---

Quick Reference Table

---

Self-Check Questions

1. A system has poles at $s = -2 \pm j3$. Is it stable? What type of time response will it exhibit, and why?

2. Compare how adding a zero near a dominant pole affects system response versus adding another pole. Which speeds up the system?

3. If a Bode plot shows a phase margin of $-10°$, what does this tell you about closed-loop stability? How would you fix it?

4. A first-order system has $\tau = 0.5$ seconds. A second-order system has $\zeta = 0.5$ and $\omega_n = 4$ rad/s. Which responds faster to a step input, and what characteristics would you compare?

5. You're given $H(s) = \frac{10(s+2)}{(s+1)(s+5)}$. Identify the poles, zeros, and DC gain. How would you sketch the Bode magnitude plot asymptotes?