A Class AB amplifier is a push-pull audio power stage where both output transistors are biased to conduct slightly more than half a cycle, eliminating the crossover distortion of Class B while avoiding the massive heat waste of Class A. In a practical circuit, this topology changes the thermal-to-fidelity trade-off, allowing you to deliver tens of watts of clean, low-THD audio without requiring the massive, fan-cooled heatsinks demanded by pure Class A designs.

Bench Rule of Thumb: If your audio output stage is running hot enough to burn your finger but doesn't require active cooling for a 50W output, you are almost certainly looking at a Class AB linear topology.

The Core Trade-Off: Linear Fidelity vs. Thermal Reality

To understand why the Class AB architecture dominates linear audio, you have to look at the failures of its predecessors. A pure Class A amplifier keeps its output transistor fully conducting at all times. It sounds incredible because the transistor never turns off, meaning there is zero switching or crossover distortion. However, its theoretical maximum efficiency is 25% (or 50% with an inductive load), meaning a 50W Class A amp will dissipate upwards of 150W as pure heat at idle.

Class B solves the heat problem by using a push-pull pair where one transistor handles the positive half of the AC waveform and the other handles the negative. Theoretical efficiency jumps to 78.5%. But because silicon transistors require about 0.6V to turn on, there is a dead zone around the zero-crossing point where neither transistor is conducting. This creates harsh crossover distortion that ruins high-fidelity audio.

Class AB is the engineered compromise. By injecting a small quiescent bias current ($I_q$) through both transistors even when no audio signal is present, both devices remain slightly "on" during the zero-crossing. This wipes out the crossover dead zone while keeping idle power dissipation to a manageable few watts. According to Analog Devices' amplifier class guide, this bias point is the critical variable that dictates both the Total Harmonic Distortion (THD) and the thermal baseline of the circuit.

Worked Numeric Example: Sizing a Heatsink for the LM3886

Let's move from theory to the workbench. You are building a stereo amp using the legendary Texas Instruments LM3886 Class AB IC. You have a split power supply of ±28V ($V_{CC} = 28V$) and are driving an $8\Omega$ speaker load. How big does your heatsink need to be?

Many beginners mistakenly calculate heatsink size based on the maximum power delivered to the speaker. This is wrong. You must size the heatsink for the maximum power dissipated by the IC itself, which occurs at roughly 40% of maximum output voltage.

  1. Calculate Maximum IC Power Dissipation ($P_{D(max)}$):
    The formula for a split-supply Class AB stage is $P_{D(max)} = \frac{2 \times V_{CC}^2}{\pi^2 \times R_L}$.
    $P_{D(max)} = \frac{2 \times 28^2}{9.87 \times 8} = \frac{1568}{78.96} \approx 19.86W$.
    Your IC will burn nearly 20 watts as heat under worst-case continuous sine wave conditions.
  2. Determine Required Thermal Resistance ($\theta_{JA}$):
    The LM3886 shuts down at a junction temperature ($T_J$) of 165°C. For reliability, we derate our target $T_J$ to 120°C. Assuming a worst-case chassis ambient temperature ($T_A$) of 40°C:
    $\theta_{JA} = \frac{T_J - T_A}{P_{D(max)}} = \frac{120 - 40}{19.86} = 4.02°C/W$.
  3. Subtract Internal Thermal Resistances:
    As noted in TI's LM3886 datasheet, the junction-to-case resistance ($\theta_{JC}$) is 1.0°C/W. A standard silicone thermal pad adds a case-to-sink resistance ($\theta_{CS}$) of about 0.5°C/W.
    Required Heatsink ($\theta_{SA}$) = $4.02 - 1.0 - 0.5 = 2.52°C/W$.

The Verdict: You need an extruded aluminum heatsink rated for 2.5°C/W or lower. If you use a 5.0°C/W heatsink, the IC will trigger its internal thermal protection and mute the audio during loud bass transients.

Where You Meet Class AB Amplifiers in Practice

While Class D (switching) amplifiers have taken over subwoofers and portable Bluetooth speakers due to their 90%+ efficiency, Class AB remains the standard where high-frequency linearity and low electromagnetic interference (EMI) are paramount.

  • Powered Studio Monitors: Bi-amped monitors like the JBL 305P MkII often use a Class D amp for the woofer (where efficiency matters and high-frequency switching noise is filtered by the driver's inductance) but use a Class AB amp for the tweeter to ensure pristine, noise-free high-frequency reproduction.
  • High-End AV Receivers: Brands like Marantz and Denon still utilize massive discrete Class AB output stages (using pairs of Sanken or ON Semi power transistors) in their flagship receivers to deliver high current into difficult 4-ohm speaker loads without the harshness of poorly filtered PWM.
  • Solid-State Guitar Amps: The Roland Jazz Chorus relies on a discrete Class AB output stage to deliver its signature clean headroom, which solid-state switching amps struggle to replicate without digital modeling.

Real-World Scenario Walkthrough: The Thermal Runaway Trap

Theory looks great on a schematic, but discrete Class AB amplifiers harbor a fatal physical trap if you ignore semiconductor physics. Here is a real-world bench failure.

Setup: A DIY builder constructs a 50W discrete push-pull amplifier using TIP31C (NPN) and TIP32C (PNP) output transistors. To set the Class AB bias, they place two 1N4148 signal diodes in series between the bases of the output transistors. This provides roughly 1.2V of forward bias, setting the quiescent idle current ($I_q$) to a safe 25mA per transistor. The output transistors are bolted to a large heatsink, but the bias diodes are left hanging in the air on the PCB.

Numbers: Silicon base-emitter junctions have a negative temperature coefficient of approximately -2mV/°C. As the output transistors heat up during a loud testing session, their junction temperature rises by 60°C. This means their required turn-on voltage ($V_{BE}$) drops by 120mV. However, the 1N4148 bias diodes on the PCB remain at room temperature, stubbornly supplying the original 1.2V bias.

Outcome: Because the output transistors now require 120mV less to turn on, but are still being fed the original voltage, the "extra" 120mV drives the base current exponentially higher. The idle current spikes from 25mA to 400mA. This massive current generates more heat, which drops the $V_{BE}$ further, which draws more current. Within four minutes, the amp emits a loud hiss, the output transistors short-circuit internally, and the power supply fuse blows.

What Went Wrong: The builder failed to implement thermal tracking. In a properly designed Class AB stage, the bias diodes (or a $V_{BE}$ multiplier transistor) must be physically mounted directly to the same heatsink as the output transistors. As the output transistors heat up and their $V_{BE}$ requirement drops, the heatsink warms the bias diodes, causing their forward voltage drop to decrease at the exact same -2mV/°C rate. This dynamic compensation holds the quiescent current perfectly stable regardless of temperature. For a deeper look at semiconductor thermal limits, the All About Circuits semiconductor textbook provides excellent baseline data on push-pull thermal dynamics.

Safety Warning: When debugging quiescent current on a discrete Class AB amp, always use a current-limited bench power supply or wire a 12V incandescent bulb in series with the main DC rail. If thermal runaway begins, the bulb will light up and limit the current, saving your output transistors from catastrophic short-circuit failure.

Frequently Asked Questions

What do people commonly confuse Class AB with?
Beginners frequently confuse Class AB with Class D. Class D is a switching amplifier that uses Pulse Width Modulation (PWM) to achieve 90%+ efficiency, requiring an LC low-pass filter at the output to remove high-frequency switching noise. Class AB is a linear amplifier; the output transistors act as variable resistors, directly tracking the input voltage waveform with no high-frequency switching or output filters required. They also confuse it with Class A, which conducts 100% of the time rather than the roughly 51% conduction angle of Class AB.

Is Class AB obsolete because of Class D? No. While Class D dominates portable, battery-powered, and subwoofer applications, Class AB remains critical in RF transmission, high-end studio monitoring, and medical imaging equipment. Class D generates significant Electromagnetic Interference (EMI) and requires complex dead-time management to prevent shoot-through currents. Class AB offers inherently low noise, zero switching artifacts, and excellent Safe Operating Area (SOA) characteristics for driving highly reactive loads.

How much quiescent current ($I_q$) should I set for a discrete Class AB amp?
For standard bipolar junction transistors (BJTs) in a home audio amp, an $I_q$ between 20mA and 50mA per output device is typical. This is enough to keep the devices out of the crossover dead-zone without generating excessive idle heat. MOSFET-based Class AB stages often require higher bias currents (100mA+) to push the bias point past the gate threshold voltage and into the linear region of the transfer curve.