A Class AB power amplifier is a hybrid electronic circuit that combines the continuous conduction (low distortion) of Class A for small signals with the push-pull efficiency of Class B for large signals. By keeping both output transistors slightly turned on when no audio signal is present, this topology solves the most glaring flaw of pure Class B designs while avoiding the massive heat output of pure Class A.
In a real circuit, implementing this topology fundamentally changes your thermal management and biasing network requirements. You can no longer just bolt transistors to a heatsink; you must introduce a temperature-compensated bias spreader (usually a VBE multiplier) to prevent thermal runaway. The most common confusion among hobbyists is assuming 'AB' implies a 50/50 operational split between Class A and Class B, or conflating it with Class D (which is a high-frequency switching topology, not a linear one). In reality, a Class AB amplifier operates in Class B for 95% of its power output, only utilizing Class A operation for the few millivolts around the zero-crossing point.
The Core Concept: Biasing and the Zero-Crossing Handoff
To understand why this topology exists, you have to look at the zero-crossing point of an AC audio waveform. In a pure Class B push-pull amplifier, the NPN transistor handles the positive half of the wave, and the PNP transistor handles the negative half. Because silicon transistors require about 0.6V to 0.7V at the base-emitter junction to begin conducting, there is a 'dead zone' where neither transistor is on. This creates crossover distortion, which sounds like a harsh, fizzy static on quiet audio passages.
The Shift-Change Analogy: Think of two workers carrying a heavy log. In pure Class B, Worker 1 drops the log the exact millisecond Worker 2 grabs it, causing a jarring bump (crossover distortion). In Class AB, they overlap their grip for a few inches, ensuring a perfectly smooth handoff before one lets go.
To create this overlap, we inject a small DC bias voltage between the bases of the two output transistors. This is typically achieved using a VBE multiplier circuit—a transistor and two resistors placed between the driver bases. By adjusting the resistor ratio, you set a bias voltage of roughly 1.2V to 1.4V, which is just enough to keep both output devices in a state of slight conduction (quiescent current) even when the input signal is exactly 0V.
The Math: Quiescent Current and Efficiency in a Real Circuit
Let us run a worked numeric example for a typical solid-state audio amplifier designed to deliver 100W RMS into an 8-ohm load.
- Target Peak Voltage: To get 100W into 8Ω, we need an RMS voltage of 28.28V (since P = V²/R). The peak voltage is 28.28V × √2, which equals 40V.
- Power Supply Rails: To account for transistor saturation voltages and emitter resistor drops, we use ±45V DC rails (total 90V across the output stage).
- Quiescent Current (Iq): We set the bias trimmer to allow 50mA of idle current flowing through the output transistors.
- Quiescent Power Dissipation: 90V × 0.05A = 4.5 Watts of heat generated per channel, even with no music playing.
At full 100W output, the theoretical maximum efficiency of this Class AB design is 78.5%. In practice, accounting for driver stage losses and the voltage drop across the 0.22Ω emitter resistors, real-world efficiency sits around 60% to 65%. This means to output 100W of acoustic/electrical power, the amplifier draws roughly 160W from the wall, dissipating 60W as heat into the aluminum extrusion.
Warning: Thermal Runaway
As silicon transistors heat up, their VBE requirement drops. If your bias circuit is not thermally coupled to the heatsink, the amplifier will draw more idle current as it gets hot, which makes it hotter, drawing even more current. This positive feedback loop will destroy your output stage in seconds. Always mount the bias transistor directly to the main heatsink, as detailed in Rod Elliott's thermal stability guidelines.
Where You Meet Class AB in Practice
Despite the rise of Class D switching amplifiers, Class AB remains a staple in specific applications where low noise and high linearity are non-negotiable.
| Application | Why Class AB is Used | Common Examples |
|---|---|---|
| AV Receivers & Audiophile Amps | Zero high-frequency switching noise; excellent high-frequency phase linearity. | Denon AVR-X series, Yamaha A-S series, classic Marantz models. |
| Op-Amp Output Stages | Provides low-impedance drive without the dead-zone glitching of Class B. | LM386, NE5532, TDA2030, LM3886 (Gainclone). |
| RF Linear Amplifiers | Preserves the amplitude envelope of SSB/AM radio signals without PWM artifacts. | Ham radio HF linear amps using LDMOS or BJTs. |
| Studio Monitor Speakers | Predictable thermal behavior and flat frequency response without output LC filters. | Active studio monitors (older Genelec, KRK Rokit G3). |
If you are tearing down a heavy piece of audio equipment with a massive copper-wound toroidal transformer and large finned heatsinks, you are almost certainly looking at a Class AB power amplifier.
Frequently Asked Questions About Class AB Power Amplifiers
Is a Class AB power amplifier better than Class D for audio quality?
Historically, yes, but the gap has nearly closed. Class AB inherently avoids the high-frequency electromagnetic interference (EMI) and output filter phase-shift issues that plagued early Class D designs. However, modern Class D chips (like the TI TPA3255 or Hypex Ncore modules) achieve THD+N figures below 0.001%, rivaling or beating average Class AB implementations. Choose Class AB when you need absolute RF silence in a studio environment or are driving highly reactive, difficult speaker loads without complex compensation networks. Choose Class D when efficiency, compact size, and battery operation are priorities.
How do I set the bias current on a Class AB power amplifier kit?
You must set the quiescent current (Iq) using a digital multimeter. Power the amp through a current-limited bench supply or a dim-bulb tester to protect against shorts. Set your multimeter to DC millivolts and measure the voltage drop across one of the output transistor's emitter resistors (typically 0.22Ω or 0.33Ω). Using Ohm's Law (I = V/R), a reading of 11mV across a 0.22Ω resistor indicates exactly 50mA of bias current. Adjust the VBE multiplier trimmer potentiometer slowly—wait 60 seconds between adjustments for the thermal mass to stabilize—until you hit the manufacturer's target, usually between 10mV and 25mV.
Why does my Class AB power amplifier get so hot even with no music playing?
This is normal physics, not necessarily a fault. As calculated in our 100W example, a typical Class AB amp dissipates 3W to 6W of heat per channel at idle just to maintain the bias overlap. In a multi-channel AV receiver with 7 or 9 channels, that is 30W to 50W of continuous heat radiating into the chassis with zero audio output. However, if the heatsink becomes too hot to touch (exceeding 60°C) at idle, your bias current may be set too high, or the thermal paste between the transistor and the heatsink has dried out and failed.
What is the difference between a Class AB power amplifier and a Class H amplifier?
Class H is actually an enhancement built on top of a Class AB foundation. A standard Class AB amp runs its output transistors from fixed high-voltage rails (e.g., ±50V), meaning when outputting a small 5W signal, the transistors are dropping massive amounts of excess voltage as heat. A Class H amplifier uses a tracking power supply or switched rail tiers to dynamically lower the supply voltage when the audio signal is small, and instantly switches to the high-voltage rails for loud transients. The output stage is still operating in Class AB, but the Class H rail-switching boosts overall efficiency from ~60% up to 80% or more.
