Corporate Vertical: Power Amplifiers and Integrated Circuit Design Summary

Power Amplifiers: Principles and Characteristics

• Definition and Purpose: Power amplifiers are designed to deliver large amounts of power to a specific load.
• Key Characteristics:
  • Voltage and Current: They must be capable of generating high voltages and/or delivering large currents.
  • Output Impedance: They must possess a low output impedance.
  • Power Conversion Efficiency: High efficiency in the output stage is critical to prevent energy from being wasted as heat in the amplifier rather than being delivered to the load.
  • Linearity: For applications like audio amplification, low Total Harmonic Distortion ($THD$) is required for linear signal amplification.

Power Transistor Technology and Limitations

• Device Types: Power amplifiers typically utilize Bipolar Junction Transistors ($BJT$) or Metal-Oxide-Semiconductor Field-Effect Transistors ($MOSFET$).
• BJT and MOSFET Limitations:
  • Maximum Rated Current ($I_{max}$): Measuring in Amperes; exceeding this leads to melting wires.
  • Maximum Rated Voltage ($V_{max}$): Measuring in hundreds of Volts; exceeding this causes reverse pn-junction avalanche breakdown.
  • Maximum Rated Power ($P_{max}$): Measuring in Watts to tens of Watts; exceeding this causes the semiconductor junctions to exceed maximum temperature, leading to breakage.
• Second Breakdown: A limitation unique to $BJTs$. It occurs at higher voltages and fairly high currents, even within the other maximum boundaries. It is caused by non-uniform current density. Local areas heat up, resistance decreases, and current increases further until the area melts. This can happen in milliseconds and is terminal for the device.
• Safe Operating Area ($SOA$): This is the region defined by voltage and current limits on linear or logarithmic scales where the transistor can operate safely. For DC conditions, $P_T \approx V_{CE} I_C$.
• Transistor Examples:
  • TIP41C (NPN) / TIP42C (PNP): Complementary power transistors in a $TO-220$ package. They are used for general-purpose circuits, audio amplifiers, and linear or switching applications.

BJT vs. MOSFET Performance

• MOSFET Advantages:
  • Faster switching times.
  • Constant gain and response time over a wider range.
  • No second breakdown.
  • Easily connected in parallel.
  • Reduced parameter variation with temperature changes.

Classification of Amplifiers

• Basis of Classification: Amplifiers are classified by the percentage of the signal cycle during which the output transistors are conducting.
• Class-A: The transistor conducts for the entire cycle ($360^\circ$).
• Class-B: Each transistor conducts for exactly half of the cycle ($180^\circ$). Utilizes a complementary "push-pull" configuration.
• Class-AB: Transistors are slightly biased and conduct for slightly more than half of the cycle ($>180^\circ$). This reduces crossover distortion.
• Class-C: The transistor conducts for less than half of the cycle ($<180^\circ$).
• Other Classes: Class $D$, $E$, $F$, $G$, and $H$ use output transistors as switches and digital electronics to provide very high efficiency at the cost of higher distortion.

Class-A Amplifier Efficiency and Power

• Biasing: For maximum output swing and assuming $V_{CE(sat)} = 0\,V$:
  • $V_{CEQ} = \frac{V_{CC}}{2} = I_{CQ} R_L$
• Instantaneous Power ($P_Q$): $P_Q = v_{CE} i_C$. For a sinusoidal current $i_C = I_{CQ}(1 + \sin(\omega t))$ and voltage $v_{CE} = \frac{V_{CC}}{2}(1 - \sin(\omega t))$:
  • $P_Q = \frac{V_{CC} I_{CQ}}{2} (1 - \sin^2(\omega t))$
• Maximum Power Dissipation: Occurs when there is no AC signal (idle at Q-point):
  • $P_{Q,max} = \frac{V_{CC} I_{CQ}}{2}$
• Conversion Efficiency ($\eta$):
  • $\eta = \frac{\text{Average AC power to load } (\bar{P}_L)}{\text{Average supply power } (\bar{P}_S)}$
  • $\bar{P}_{L,max} = \frac{V_{CC} I_{CQ}}{4}$
  • $\bar{P}_S = V_{CC} I_{CQ}$
  • $\eta_{max} = 0.25 = 25\%$

Class-B Amplifier and Power Conversion Efficiency

• Operation: Consists of a complementary pair (e.g., NPN/PNP).
  • $v_I > 0$: NPN turns on, supplying current to the load.
  • $v_I < 0$: PNP turns on, sinking current from the load.
• Crossover Distortion: In a real Class-B stage with $V_{BE(on)} = 0.7\,V$, neither transistor conducts when $-0.7\,V \le v_I \le 0.7\,V$, causing dead zones in the output.
• Class-B Power Analysis:
  • $\bar{P}_L = \frac{V_p^2}{2 R_L}$
  • $\bar{P}_{Qn} = \frac{V_{CC} V_p}{\pi R_L} - \frac{V_p^2}{4 R_L}$
  • Peak efficiency at maximum swing ($V_p = V_{CC}$):
  • $\eta_{max} = \frac{\pi}{4} \approx 78.5\%$
• Pros/Cons vs Class-A:
  • Advantages: High efficiency, little power dissipation at the Q-point, more power to the load.
  • Disadvantages: Bipolar supply required (dual rails), requires matched NPN/PNP, crossover distortion.

Class-AB Output Stage and Biasing

• Quiescent Bias: Designed to reduce crossover distortion by applying a small bias so transistors conduct slightly when $v_I = 0$.
• Biasing Methods:
  • Diode Biasing: Two diodes in series ($D_1$, $D_2$) create $V_{BB} \approx 2 V_{BE(on)}$. Bias current must ensure diodes stay on during peak base current.
  • $V_{BE}$-Multiplier: A transistor ($Q_1$) circuit where $V_{BB} = V_{BE1} (1 + \frac{R_1}{R_2})$. This allows for adjustable and stable biasing.
• Thermal Stability: As temperature increases, $V_{BE(on)}$ decreases, which increases $I_{CQ}$, leading to further heating. This positive feedback is called thermal runaway.

Input Buffer Transistors and Darlington Pairs

• Input Buffers: Provide high input resistance to the preamplifier and generate necessary offset voltages.
• Darlington Pairs: Connects two transistors as a single device to significantly increase gain ($\beta \approx \beta_1 \beta_2$). This reduces the required bias current from the preceding stages (e.g., from $22.4\,mA$ to $203\,\mu A$ in specific designs) and increases input resistance ($R_{in}$).

Thermal Resistance and Heat Management

• Circuit Analogy:
  • Power Dissipation ($P$) $\equiv$ Current Source ($I$)
  • Thermal Resistance ($\theta$) $\equiv$ Electrical Resistance ($R$)
  • Temperature Difference ($\Delta T$) $\equiv$ Voltage Difference ($V$)
  • $\Delta T = P \times \theta$
• Thermal Nodes:
  • Junction ($T_j$): Device interior temperature.
  • Case ($T_c$): External package temperature.
  • Heat Sink ($T_{sink}$): Temperature of the cooling hardware.
  • Ambient ($T_{ambient}$): Environment temperature.
• Thermal Path: $T_j - T_{ambient} = P_D (\theta_{j-c} + \theta_{c-s} + \theta_{s-a})$.
• $\theta_{j-c}$: Fixed thermal resistance from junction to case.
• $\theta_{c-s}$: Resistance between case and sink; minimized using thermal grease (silicone paste).
• $\theta_{s-a}$: Resistance from sink to ambient; lowered by increasing sink surface area/fins.

Transistor Current Sources

• Current Mirror (Two-Transistor):
  • Utilizes a reference resistor ($R_1$) to set $I_{REF}$.
  • $I_{REF} = \frac{V_{CC} - V_{BE1} - V_{EE}}{R_1}$.
  • $I_O = \frac{I_{REF}}{1 + \frac{2}{\beta}}$.
  • Output resistance $R_o = r_{o2}$. Finite $V_A$ (Early Effect) makes $I_O$ sensitive to $V_{CE}$.
• Three-Transistor Current Source:
  • Reduces the influence of $\beta$ by adding a third transistor at the base junction.
  • $I_O = \frac{I_{REF}}{1 + \frac{2}{\beta(\beta_3 + 1)}}$.
• Wilson Current Source:
  • Provides higher output stability and much larger output resistance ($R_o = \frac{\beta r_{o3}}{2}$).
• Widlar Current Source:
  • Allows $I_{REF}$ and $I_O$ to differ significantly using an emitter resistor ($R_E$).
  • Formula: $V_T \ln(\frac{I_{REF}}{I_O}) \approx I_O R_E$.
  • Enables low output currents using relatively small resistors, saving IC die area.

Differential Amplifiers

• Function: Amplifies small-signal differences while rejecting noise common to both inputs.
• Definitions:
  • Differential-mode voltage ($v_d$): $v_1 - v_2$
  • Common-mode voltage ($v_{cm}$): $\frac{v_1 + v_2}{2}$
  • Output voltage: $v_o = A_d v_d + A_{cm} v_{cm}$
• Common-Mode Rejection Ratio ($CMRR$):
  • $CMRR = \left| \frac{A_d}{A_{cm}} \right|$
  • $CMRR_{dB} = 20 \log_{10} \left| \frac{A_d}{A_{cm}} \right|$
  • Higher $CMRR$ (e.g., $>80\,dB$) represents better noise rejection.
• Symmetrical Half-Circuits:
  • Common-Mode Half-Circuit: Used when $v_{cm}$ is applied to both inputs. Emitter node effectively sees double the shared current source resistance ($2R_o$).
  • Differential-Mode Half-Circuit: Used when a differential signal is applied. The emitter node acts as a signal ground (virtual ground).
• Input Resistance:
  • Differential: $R_{id} = 2r_\pi$
  • Common-mode: $R_{icm} = \frac{r_\pi}{2} + (\beta + 1) R_o$

Active Filters and Sallen-Key Biquads

• Active vs Passive:
  • Passive: No power needed, low noise, cannot amplify, difficult for low frequencies (requires massive inductors/capacitors).
  • Active: Use op-amps, no inductors, easy to scale for low frequencies, provides gain, but requires power and is noisier.
• Biquad Transfer Function: $H(s) = \frac{a_2 s^2 + a_1 s + a_0}{s^2 + b_1 s + b_0} = \frac{a_2 s^2 + a_1 s + a_0}{s^2 + (\frac{\omega_0}{Q})s + \omega_0^2}$.
• Sallen-Key Topology:
  • Low-pass ($HLP$): $HLP(s) = \frac{G \omega_0^2}{s^2 + (\frac{\omega_0}{Q})s + \omega_0^2}$.
  • High-pass ($HHP$): $HHP(s) = \frac{G s^2}{s^2 + (\frac{\omega_0}{Q})s + \omega_0^2}$.
  • Band-pass ($HBP$): $HBP(s) = \frac{G (\frac{\omega_0}{Q})s}{s^2 + (\frac{\omega_0}{Q})s + \omega_0^2}$.
• Butterworth Filters: Characterized by a flat passband. Factorized functions are used based on order ($n$) for a normalized cutoff of $1\,rad/s$.
• Scaling:
  • Frequency Scaling ($k_f$): Divide capacitors/inductors by $k_f$.
  • Impedance Scaling ($k_z$): Multiply resistors/inductors by $k_z$; divide capacitors by $k_z$.
• Sensitivity Analysis (Monte Carlo): Simulates random variations in components within tolerances (e.g., $\pm 10\%$ for resistors, $\pm 20\%$ for capacitors) to predict real-world performance.

Practical Operational Amplifiers

• Ideal Op-Amp Assumptions: Infinite gain ($A_{OL}$), infinite input resistance ($R_i$), zero output resistance ($R_o$), infinite bandwidth, and zero offsets.
• Practical Parameters:
  • Finite Open-Loop Gain: Typically around $10^5$. Reduces actual closed-loop gain.
  • Voltage Rails: Input/output signals cannot swing all the way to $V_{CC}$ or $V_{EE}$.
  • Output Current Limit: Usually limited to about $10\,mA$.
• Gain-Bandwidth Product ($GBP$):
  • Trade-off between gain and bandwidth: $f_T = A_{CLO} \times f_{3dB}$.
• Slew Rate ($SR$): Maximum rate of change of output voltage ($\frac{dv_o}{dt}$).
  • Full-power Bandwidth ($FPBW$): Frequency where the op-amp becomes slew rate limited: $f_{max} = \frac{SR}{2 \pi V_{PO}}$.
• DC Non-Idealities:
  • Input Offset Voltage ($V_{OS}$): Differential voltage needed to force $v_o = 0$. Modeled as a source at the input.
  • Input Bias Current ($I_B$): Average current into the terminals: $I_B = \frac{I_{B1} + I_{B2}}{2}$.
  • Input Offset Current ($I_{OS}$): Difference between terminal currents: $I_{OS} = |I_{B1} - I_{B2}|$.
• Compensation:
  • Bias Current Compensation: Adding a resistor $R_3 = R_1 || R_2$ at the non-inverting terminal to balance voltage drops.
  • Offset Voltage Compensation: Using resistor networks/potentiometers to inject small correcting voltages at the inputs.