A Class AB amplifier is a hybrid audio power circuit that combines the low-distortion linear operation of Class A with the high-efficiency push-pull topology of Class B by biasing both output transistors slightly into their conduction zone to eliminate crossover distortion. When you choose this topology for a real circuit or installation, it fundamentally changes your thermal management strategy and mandates a temperature-compensated biasing network—trading a small percentage of theoretical efficiency for pristine audio linearity and the elimination of high-frequency switching noise.
The Core Mechanics and Amplifier Class Comparison
In a standard push-pull output stage, an NPN transistor handles the positive half of the AC audio waveform, while a PNP transistor handles the negative half. In a pure Class B design, these transistors are biased exactly at cutoff. Because silicon bipolar junction transistors (BJTs) require approximately 0.6V across the base-emitter junction to begin conducting, a pure Class B amp creates a 'dead zone' when the audio signal crosses the 0V threshold. This dead zone results in severe crossover distortion, which sounds like harsh fuzz on low-level signals and generates high-frequency harmonics that can destroy tweeters.
The Class AB topology solves this by applying a small DC bias voltage between the bases of the two output transistors. This holds both devices in a slightly conducting state (quiescent current, or $I_q$) even when no audio signal is present. As the signal crosses zero, one transistor smoothly takes over from the other without a dead zone. To understand where Class AB sits in the broader ecosystem of audio power design, review the comparison matrix below.
| Amplifier Class | Conduction Angle | Theoretical Max Efficiency | Typical Real-World Efficiency | Crossover Distortion | Primary Use Case |
|---|---|---|---|---|---|
| Class A | 360° (Always ON) | 25% (Single-ended) / 50% (Push-pull) | 15% - 30% | None | Boutique audiophile amps, preamp stages |
| Class B | 180° (Half-wave) | 78.5% | 60% - 70% | Severe (Unusable for audio) | RF transmitters, legacy paging systems |
| Class AB | 181° - 359° | 78.5% | 50% - 65% | Eliminated (with proper bias) | AV receivers, studio monitors, guitar amps |
| Class D | Switching (PWM) | 100% | 85% - 95% | N/A (Switching noise instead) | Subwoofers, portable PA, budget soundbars |
Source: Amplifier class fundamentals via Electronics Tutorials and All About Circuits.
Worked Numeric Example: Quiescent Bias and Thermal Dissipation
To see how Class AB changes real-world circuit design, let us calculate the quiescent bias and thermal dissipation for a 100W RMS amplifier driving an $8\Omega$ load. We will use a dual power supply with $\pm 40V_{DC}$ rails and ON Semiconductor's MJL21193 (NPN) and MJL21194 (PNP) complementary power transistors in TO-264 packages.
Step 1: Calculate Class AB Quiescent Power Dissipation ($P_q$)
At idle, there is no audio signal, so the emitter voltage is essentially 0V. The voltage across the collector-emitter junction ($V_{CE}$) of the NPN transistor is the full positive rail: 40V.
$P_q = V_{CE} \times I_q$
$P_q = 40V \times 0.05A = 2.0W$ per transistor.
The total idle heat dissipation for the push-pull pair is 4.0W. This requires only a modest extruded aluminum heatsink.
Step 2: Compare to a Pure Class A Topology
To deliver 100W RMS into $8\Omega$, the peak output current must be $I_{peak} = \sqrt{(100W \times 8\Omega) / 8\Omega} \times \sqrt{2} \approx 5A$. A push-pull Class A amplifier must bias its output transistors at least at half the peak current to avoid cutoff, meaning $I_q$ must be 2.5A.
$P_{q(ClassA)} = 40V \times 2.5A = 100W$ per transistor.
The total idle heat dissipation is 200W—enough to boil water, requiring massive forced-air cooling or enormous thermal mass, despite delivering the exact same audio output as the Class AB design.
Step 3: The Thermal Runaway Hazard
This 4.0W Class AB dissipation introduces a critical installation variable: thermal tracking. The base-emitter voltage ($V_{BE}$) of a silicon BJT drops by approximately $-2mV/^\circ C$. If your heatsink temperature rises by $30^\circ C$ during operation, the required $V_{BE}$ drops by $60mV$. If your bias circuit does not automatically reduce the drive voltage by that exact $60mV$, the quiescent current will skyrocket from 50mA to potentially 500mA, causing thermal runaway and catastrophic transistor failure. This is why every practical Class AB amp requires a $V_{BE}$ multiplier transistor physically bolted to the output heatsink to monitor and compensate for temperature shifts in real-time.
Where You Meet Class AB in Practice
Despite the relentless march of switching technologies, the Class AB topology remains the backbone of professional and high-end consumer audio in 2026. You will encounter it in these specific applications:
- Flagship AV Receivers: While budget and mid-range receivers have migrated to Class D to meet strict energy consumption mandates, flagship models (like the Denon AVR-A1H or Marantz Cinema series) still utilize massive discrete Class AB output stages. Class AB handles complex, low-impedance speaker loads (dipping to $2\Omega$ or $3\Omega$ with heavy phase shifts) with greater phase-margin stability. It also avoids the aggressive LC output filters required by Class D, which can alter high-frequency response depending on the specific speaker's impedance curve.
- High-End Studio Monitors: While nearfield monitors have largely adopted Class D for cost and thermal reasons, main monitors and high-end midfields (such as the ATC SCM50 or Focal Trio11) rely on Class AB. The topology ensures zero intermodulation distortion in the critical midrange frequencies, providing the transient accuracy required for professional mixing.
- Solid-State Instrument Amplifiers: The Roland Jazz Chorus (JC-120) remains an industry benchmark for clean guitar tones. It relies entirely on a discrete Class AB topology to deliver massive clean headroom without the harsh, artificial clipping characteristics inherent in switching amplifiers when pushed to their limits.
Common Confusions: What Class AB Is Not
When specifying or repairing audio gear, Class AB is frequently conflated with other amplifier designations. Clearing up these confusions is vital for proper troubleshooting and system design.
Confusion 1: Class AB vs. Class D
Class D is a switching amplifier that uses Pulse Width Modulation (PWM). It operates its output MOSFETs fully ON or fully OFF, achieving >90% efficiency but generating a high-frequency switching carrier (typically 300kHz to 600kHz) that must be stripped out by a passive LC low-pass filter before reaching the speaker. Class AB operates its output devices in the linear (active) region. There is no switching carrier, no PWM, and no output inductor. If you are debugging high-frequency tweeter burnout, a failing Class D output filter is a prime suspect; in a Class AB amp, you would look for DC offset or high-frequency parasitic oscillation instead.
Confusion 2: Class AB vs. Class G and Class H
Class G and Class H are not separate fundamental output topologies; they are power-supply efficiency enhancements built on top of a Class AB core.
A Class G amplifier uses multiple fixed voltage rails (e.g., $\pm 35V$ and $\pm 70V$) and switches between them based on the audio signal amplitude, reducing the voltage drop across the output transistors during low-level signals.
A Class H amplifier continuously modulates the supply rail voltage using a tracking switching power supply to perfectly mirror the audio envelope. In both cases, the actual output stage delivering current to the speaker remains a Class AB push-pull circuit. If a Class H amp blows an output transistor, you troubleshoot the output stage exactly as you would a standard Class AB amp, but you must also verify the rail-tracking power supply is functioning correctly.
Frequently Asked Questions
Why does my Class AB amplifier need a warm-up period before it sounds 'right'?
When a Class AB amplifier is first powered on, the output transistors and the heatsink are at ambient room temperature. As the $V_{BE}$ multiplier and output devices warm up over the first 15 to 30 minutes, the thermal compensation circuit adjusts the bias voltage. Many designers intentionally set the cold $I_q$ slightly lower than the optimal thermal equilibrium point to prevent thermal runaway during the warm-up phase. The amp reaches its lowest distortion specification only once the heatsink stabilizes at its designed operating temperature (often $45^\circ C$ to $55^\circ C$).
Can I use MOSFETs instead of BJTs for a Class AB output stage?
Yes, but the biasing requirements change drastically. Standard vertical MOSFETs have a positive temperature coefficient at low currents, which can lead to thermal runaway just like BJTs. However, lateral MOSFETs (such as the Exicon ECX10N20/ECX10P20 pair) are specifically designed for audio. They exhibit a zero temperature coefficient at higher currents and lack the secondary breakdown limitations of BJTs. If you swap a BJT output stage for lateral MOSFETs, you must redesign the $V_{BE}$ multiplier to provide the higher gate threshold voltage (typically $\pm 2V$ to $\pm 3V$ total bias, compared to $\pm 1.2V$ for BJTs) and account for the higher input capacitance of the MOSFET gates.
How do I measure crossover distortion on the bench?
Standard THD (Total Harmonic Distortion) meters often mask crossover distortion because it occurs at low power levels. To measure it, inject a 1kHz sine wave into the amplifier and use an oscilloscope to monitor the output across a dummy load. Zoom in heavily on the zero-crossing point (the X-axis). If the bias is set correctly, the waveform will be a smooth, continuous curve. If the $I_q$ is too low, you will see a distinct 'notch' or flattening at the zero-crossing. For precise quantification, perform an FFT analysis at 1W output power and look for a spike in the 3rd and 5th harmonics, which are the primary signatures of crossover non-linearity.
