---
theme: default
title: 'Lecture 37: Barkhausens criteria for oscillation'
info: |
  ## DI02011011: Electronics Circuit and Application (ECA)
  Oscillators
---

# DI02011011: Electronics Circuit and Application (ECA)
## Lecture 37: Barkhausen's criteria for oscillation
### Oscillators

<!--
Welcome to Lecture 37. Today we examine Barkhausen's Criteria for oscillation. These two conditions—magnitude equal to unity and loop phase shift equal to zero or 360 degrees—form the exact foundation for designing and analyzing sine-wave oscillators.
-->

---

# Agenda: Barkhausen's Criteria for Oscillation

- Mathematical Statement of Barkhausen's Criteria
- Complex S-Plane Analysis & Pole Trajectories
- Startup Criterion ($|A\beta| > 1$) vs Steady-State Criterion ($|A\beta| = 1$)
- Phase Slope and Frequency Stability Factor ($S_F$)
- Application Case Study: RC Phase-Shift Oscillator
- Application Case Study: Wien-Bridge Oscillator
- Application Case Study: Colpitts LC Oscillator
- Non-linear Dynamic Limit Cycles and Transconductance Compression
- Limitations of Barkhausen Criterion & Nyquist Encirclement Test
- Phase Noise Analysis & Leeson's Model
- Summary & Engineering Design Principles

<!--
Here is our lecture outline. We will mathematically define Barkhausen's criteria, analyze pole movements on the s-plane, work through detailed circuit examples, and explore practical considerations like phase noise and stability.
-->

---

# Mathematical Statement of Barkhausen's Criteria

- Loop Gain Definition: $T(s) = A(s)\beta(s)$, where $A(s)$ is amplifier open-loop transfer function and $\beta(s)$ is feedback transfer factor.
- Magnitude Condition: $|T(j\omega_0)| = |A(j\omega_0) \beta(j\omega_0)| = 1$. Energy supplied by amplifier exactly replaces energy dissipated per cycle in feedback network.
- Phase Condition: $\angle T(j\omega_0) = \angle A(j\omega_0) + \angle \beta(j\omega_0) = 2\pi n$ ($n = 0, 1, 2, \dots$). Signals returning through feedback loop arrive perfectly in phase with original input.
- Combined Complex Representation: $A(j\omega_0)\beta(j\omega_0) = 1 + j0 = 1 \angle 0^\circ$ at resonant frequency $\omega_0$.

<!--
The Barkhausen stability criterion states that for a feedback circuit to sustain steady-state oscillations, the loop gain must equal 1 + j0 at frequency omega_0. This means the magnitude must be exactly 1 and the phase shift around the loop must be zero or a multiple of 360 degrees.
-->

---

# Complex S-Plane Analysis & Pole Trajectories

- Characteristic Equation: System closed-loop transfer function denominator $1 - A(s)\beta(s) = 0$ determines pole locations $s = \sigma \pm j\omega_0$.
- Left-Half Plane ($\sigma < 0$): Closed-loop poles at $s = -\alpha \pm j\omega_0$ produce exponentially damped oscillations $v(t) = V_0 e^{-\alpha t} \cos(\omega_0 t)$ (sub-critical feedback, $|A\beta| < 1$).
- Right-Half Plane ($\sigma > 0$): Closed-loop poles at $s = +\alpha \pm j\omega_0$ produce exponentially growing oscillations $v(t) = V_0 e^{+\alpha t} \cos(\omega_0 t)$ (super-critical feedback, $|A\beta| > 1$).
- $j\omega$-Axis Alignment ($\sigma = 0$): Steady-state Barkhausen condition places poles exactly on imaginary axis at $s = \pm j\omega_0$, maintaining constant amplitude sinusoidal output $v(t) = V_0 \cos(\omega_0 t)$.

<!--
We can analyze Barkhausen's criterion using s-plane poles. If poles are in the left half plane, oscillations decay. If they are in the right half plane, oscillations grow. Steady-state oscillation occurs when poles rest precisely on the j-omega axis.
-->

---

# Startup Condition vs Steady-State Criterion

- Small-Signal Startup Requirement: At power-up ($t = 0^+$), circuit requires $|A(j\omega_0)\beta(j\omega_0)| > 1$ (typically $1.05$ to $3.0$) to force poles into RHP and ensure reliable self-starting from thermal noise.
- Exponential Buildup Phase: Output voltage grows according to $v_{out}(t) = V_{noise} \cdot e^{\frac{|A\beta|-1}{2\tau} t} \cdot \sin(\omega_0 t)$.
- Transconductance Saturation: As AC voltage amplitude increases, active device operates across non-linear $I-V$ characteristic, causing large-signal transconductance $G_m(V_{amp})$ to decrease.
- Dynamic Pole Movement: Pole pair shifts leftward from RHP toward $j\omega$-axis, automatically settling at $|A(j\omega_0)\beta(j\omega_0)|_{large-signal} = 1.0$ at target amplitude $V_0$.

<!--
Notice the crucial practical distinction: to start up from zero, an oscillator must initially have $|A\beta| > 1$. As the oscillation grows, non-linear gain compression naturally pulls $|A\beta|$ down to exactly 1.0 in steady state.
-->

---

# Phase Slope and Frequency Stability Factor

- Frequency Selection via Phase Zero-Crossing: Operating frequency $f_0$ is fixed where feedback loop phase $\phi(\omega) = \angle A(j\omega) + \angle \beta(j\omega) = 0^\circ$.
- Frequency Stability Factor Definition: $S_F = \left. \omega_0 \frac{d\phi}{d\omega} \right|_{\omega = \omega_0} = \left. \frac{d\phi}{d(\Delta f / f_0)} \right|_{f = f_0}$.
- Sensitivity to Phase Noise: A phase perturbation $\Delta \phi$ caused by component variations produces frequency shift $\Delta f_0 \approx \frac{\Delta \phi}{d\phi / d\omega}$.
- High-$Q$ Stabilization: Networks with high phase slope $\frac{d\phi}{d\omega}$ (such as Quartz crystals and high-$Q$ LC tanks) dramatically increase $S_F$, minimizing frequency drift.

<!--
Frequency stability depends directly on how fast the loop phase changes with frequency. A steep phase slope dphi/domega means that any small phase shift caused by temperature or noise will result in only a microscopic change in operating frequency.
-->

---

# Application Case Study: RC Phase-Shift Oscillator

- Feedback Network Transfer Function: 3-stage high-pass RC ladder transfer factor $\beta(s) = \frac{s^3 R^3 C^3}{s^3 R^3 C^3 + 6 s^2 R^2 C^2 + 5 s R C + 1}$.
- Evaluating on $j\omega$ Axis: $\beta(j\omega) = \frac{-j \omega^3 R^3 C^3}{(1 - 6 \omega^2 R^2 C^2) + j (5 \omega R C - \omega^3 R^3 C^3)}$.
- Imposing Phase Condition: Phase shift is $180^\circ$ when imaginary part of denominator equals zero: $5 \omega_0 R C - \omega_0^3 R^3 C^3 = 0 \implies \omega_0 = \frac{1}{\sqrt{6} R C}$.
- Imposing Magnitude Condition: Substituting $\omega_0$ into real part gives $\beta(j\omega_0) = -\frac{1}{29}$. To satisfy Barkhausen criterion $A\beta = 1$, inverting amplifier gain must be $A_v = -29$ ($|A_v| = 29$).

<!--
Let's apply Barkhausen's criterion to the 3-stage RC phase-shift oscillator. By setting the imaginary part of the feedback factor to zero, we find the oscillation frequency omega_0 = 1/(RC sqrt(6)). Evaluating the magnitude at this frequency gives beta = -1/29, which means the inverting amplifier must provide a voltage gain of -29.
-->

---

# Application Case Study: Wien-Bridge Oscillator

- Feedback Network Transfer Function: Lead-lag Wien bridge network $\beta(s) = \frac{Z_p}{Z_s + Z_p} = \frac{sRC}{s^2R^2C^2 + 3sRC + 1}$.
- Evaluating on $j\omega$ Axis: $\beta(j\omega) = \frac{j\omega RC}{(1 - \omega^2 R^2 C^2) + j 3 \omega RC}$.
- Phase Zero-Crossing Condition: Real term in denominator vanishes at $\omega_0^2 R^2 C^2 = 1 \implies \omega_0 = \frac{1}{RC} \implies f_0 = \frac{1}{2\pi RC}$.
- Magnitude Criterion Evaluation: At $\omega_0$, $\beta(j\omega_0) = \frac{j\omega_0 RC}{j 3 \omega_0 RC} = \frac{1}{3} \angle 0^\circ$. Non-inverting op-amp voltage gain $A_v = 1 + \frac{R_f}{R_1}$ must equal $3$ ($\frac{R_f}{R_1} = 2$).

<!--
Applying Barkhausen's criteria to the Wien-Bridge oscillator yields omega_0 = 1/RC. At this frequency, the phase shift is 0 degrees and the feedback magnitude is 1/3. Hence, the non-inverting amplifier must have a gain of 3.
-->

---

# Application Case Study: Colpitts LC Oscillator

- Circuit Model: Transistor amplifier coupled with LC tank consisting of split capacitors $C_1, C_2$ and inductor $L$.
- Tank Impedance & Resonance: Loop reactance vanishes at $\omega_0 L - \frac{1}{\omega_0 C_1} - \frac{1}{\omega_0 C_2} = 0 \implies \omega_0 = \frac{1}{\sqrt{L C_{eq}}}$, where $C_{eq} = \frac{C_1 C_2}{C_1 + C_2}$.
- Feedback Factor Calculation: Voltage division across capacitive divider yields $\beta = \frac{V_f}{V_{out}} = \frac{X_{C1}}{X_{C2}} = \frac{C_2}{C_1}$ (with $180^\circ$ phase inversion via common-emitter configuration).
- Barkhausen Gain Threshold: Transistor voltage gain requirement $|A_v| \ge \frac{C_1}{C_2}$, or transconductance $g_m R_L' \ge \frac{C_1}{C_2}$.

<!--
In the Colpitts oscillator, split capacitors C1 and C2 form the feedback voltage divider. Applying Barkhausen's criteria shows that the resonant frequency is determined by the equivalent capacitance Ceq = C1*C2/(C1+C2) and the required amplifier gain is C1/C2.
-->

---

# Non-linear Dynamic Limit Cycles & Transconductance Compression

- Phase-Space Representation: Trajectory in $(v, \frac{dv}{dt})$ state space evolves from origin (noise seed) into a stable closed orbit called a limit cycle.
- BJT Large-Signal Transconductance: Collector current under large AC base drive $V_m$: $I_C(t) = I_Q e^{\frac{V_m \cos\omega t}{V_T}} \implies G_m(V_m) = \frac{g_m}{V_m / V_T} \frac{2 I_1(V_m/V_T)}{I_0(V_m/V_T)}$.
- Dynamic Self-Balancing: As peak amplitude $V_m$ grows, large-signal transconductance $G_m(V_m)$ continuously decreases until $G_m(V_m) R_L \beta = 1.0$.
- Sinusoidal Purity Tradeoff: Operating near threshold ($A_{start} \beta \approx 1.1$) yields low distortion sinusoidal output; excessive startup gain ($A_{start} \beta \gg 1$) forces severe non-linear clipping.

<!--
Steady-state amplitude in Barkhausen oscillators is governed by limit cycles in non-linear dynamics. As oscillation amplitude V_m expands, the active transistor's large-signal transconductance G_m decreases until loop gain exactly equals 1.
-->

---

# Limitations of Barkhausen Criterion & Nyquist Encirclement Test

- Linearization Limitation: Barkhausen criterion assumes linear small-signal steady-state conditions; fails to predict non-linear latch-up or multi-frequency squegging.
- Necessity vs Sufficiency: $A\beta = 1 \angle 0^\circ$ is a necessary condition for linear systems, but not always sufficient for complex multi-stage circuits with multiple phase crossings.
- Nyquist Stability Test: Plotting polar locus of $A(j\omega)\beta(j\omega)$ from $\omega = -\infty$ to $+\infty$; sustained oscillation requires Nyquist plot to pass through or encircle critical point $(1, j0)$.
- Gain & Phase Margins: Oscillators are engineered with negative gain margin at $f_0$ during startup to guarantee instability, transitioning to 0 dB margin at steady state.

<!--
It's important to understand the limits of Barkhausen's criteria. While it provides necessary conditions, rigorous stability analysis for complex multi-loop circuits requires the Nyquist stability test to verify encirclement of the point (1, j0).
-->

---

# Phase Noise & Spectral Purity Model

- Spectral Line Broadening: Ideal oscillator produces delta function spectrum $\delta(f - f_0)$; actual oscillator has phase noise sidebands due to thermal and flicker noise.
- Leeson's Phase Noise Model: $L(f_m) = 10 \log_{10} \left[ \frac{1}{2} \left( 1 + \left( \frac{f_0}{2 Q_L f_m} \right)^2 \right) \left( 1 + \frac{f_c}{f_m} \right) \frac{F k T}{P_{sig}} \right]$ in dBc/Hz.
- Key Parameters: $f_m$ is offset frequency from carrier, $Q_L$ is loaded quality factor, $f_c$ is flicker noise corner frequency, $F$ is amplifier noise figure, $P_{sig}$ is signal power.
- Design Rule for High Spectral Purity: Maximize resonator quality factor $Q_L$ and power dissipation $P_{sig}$ while minimizing amplifier noise figure $F$.

<!--
Real oscillators exhibit phase noise around the carrier frequency. Leeson's model shows that to minimize phase noise, we must maximize the loaded Q factor of the resonator tank and drive the oscillator with adequate signal power.
-->

---

# Summary: Key Takeaways on Barkhausen's Criteria

- Barkhausen Conditions: Sustained oscillation requires loop gain magnitude $|A\beta| = 1$ and total loop phase shift $\angle A\beta = 0^\circ$ or $360^\circ$.
- Startup vs Steady-State: System requires small-signal $|A\beta| > 1$ for startup from thermal noise, settling to $|A\beta| = 1$ in steady state via gain compression.
- Frequency Determination: Operating frequency $f_0$ is set by phase condition $\phi(f_0) = 0^\circ$; phase slope $\frac{d\phi}{d\omega}$ dictates frequency stability $S_F$.
- Practical Applications: Validated across RC Phase-Shift ($f_0 = \frac{1}{2\pi RC\sqrt{6}}, A_v = 29$), Wien-Bridge ($f_0 = \frac{1}{2\pi RC}, A_v = 3$), and Colpitts ($f_0 = \frac{1}{2\pi\sqrt{LC_{eq}}}, A_v = \frac{C_1}{C_2}$).

<!--
In summary, Barkhausen's criteria provide the essential foundation for oscillator analysis and design. Mastering magnitude and phase conditions allows us to analyze any RC, LC, or Crystal oscillator topology.
-->