A class AB amplifier is a push-pull audio output stage that biases its transistors slightly into conduction to eliminate crossover distortion while maintaining much higher efficiency than a pure class A design. In a real circuit, this topology changes the thermal management requirements and the output waveform fidelity, allowing high-power audio delivery without the massive heatsinks of Class A or the harsh zero-crossing glitches of Class B. Hobbyists and students commonly confuse it with Class A (which runs output devices fully on 100% of the time) and Class D (which uses high-frequency PWM switching rather than linear analog operation).
The Core Mechanics: Biasing Away Crossover Distortion
To understand Class AB, you first have to look at the fatal flaw of Class B. A standard Class B push-pull amplifier uses an NPN transistor for the positive half of the audio waveform and a PNP transistor for the negative half. However, silicon bipolar junction transistors (BJTs) require a base-emitter voltage ($V_{be}$) of roughly 0.6V to begin conducting. When the audio signal crosses the 0V zero-line, there is a brief window between +0.6V and -0.6V where neither transistor is turned on. This dead zone creates a harsh, jagged notch in the waveform known as crossover distortion, which sounds like a gritty buzz on high-frequency transients.
The Class AB topology fixes this by injecting a small, constant DC bias voltage between the bases of the two output transistors. Think of it like two water valves feeding a single pipe; instead of slamming one completely shut before opening the other (which causes a pressure spike or glitch), you leave both cracked open just a tiny bit during the handoff. This bias is typically generated by a string of diodes or a $V_{be}$ multiplier transistor circuit, forcing a small quiescent current ($I_q$) to flow through both output devices even when no audio signal is present. Because both devices remain slightly 'on' during the zero-crossing, the handoff is seamless.
Worked Numeric Example: Calculating Quiescent Power and Efficiency
Let us calculate the real-world thermal and power metrics for a discrete Class AB amplifier driving an 8-ohm loudspeaker. We will use a dual-rail power supply of $\pm 35V$ (70V total) and modern complementary BJTs like the ON Semi NJL3281D (NPN) and NJL1302D (PNP).
- Quiescent State (Idle): We set the bias trimpot to establish a quiescent current ($I_q$) of 50mA. The total supply voltage is 70V. The power dissipated as heat at idle is $P = V \times I = 70V \times 0.05A = 3.5W$ per channel. This is easily handled by a modest extruded aluminum heatsink.
- Maximum Signal State: At full output swing, the transistors cannot reach the absolute supply rails; they saturate about 2V short of the rails. Our peak output voltage is $35V - 2V = 33V$. The maximum RMS power delivered to the 8-ohm load is $P_{out} = V_{peak}^2 / (2 \times R_L) = 33^2 / 16 \approx 68W$.
- Efficiency Calculation: The theoretical maximum efficiency of a Class AB amp at full sine-wave output is identical to Class B: $\pi / 4$, or roughly 78.5%. In practice, accounting for emitter resistor losses and driver stage consumption, real-world peak efficiency sits around 65% to 70%. However, at typical 1W listening levels, efficiency plummets to under 10%, which is why Class AB amps run warm even at low volumes.
Where You Meet Class AB Amplifiers in Practice
Despite the rise of switching amplifiers, Class AB remains a staple in specific audio applications where linearity and low electromagnetic interference (EMI) are paramount.
- Mid-to-High-Tier AV Receivers and Integrated Amps: Brands like Marantz, Denon, and Yamaha still heavily utilize Class AB topologies for their front left/right stereo channels. The absence of high-frequency switching noise means simpler output filtering and a cleaner noise floor for sensitive analog sources like turntables.
- Guitar Amplifiers and Pedals: Solid-state guitar amps (like the classic Roland JC-120) use Class AB because of how it clips. When overdriven, a Class AB stage produces a softer, more symmetrical clipping profile with even-order harmonics that guitarists find musically pleasing, unlike the harsh aliasing of an overdriven Class D chip.
- Monolithic IC Power Amps: If you are building a DIY audio project, you will frequently encounter Class AB integrated circuits. The legendary Texas Instruments LM3886 (now largely legacy but still cloned), the STMicroelectronics TDA7294, and the LM1875 are all internally Class AB. They package the $V_{be}$ multiplier, driver stage, and thermal shutdown protection into a single TO-220 or TO-247 package.
Class AB vs. Class D in Modern Audio
When designing or buying audio gear in 2026, the choice almost always comes down to Class AB versus Class D. Here is how they stack up across critical engineering criteria.
| Criterion | Class AB Amplifier | Class D Amplifier |
|---|---|---|
| Idle Efficiency | Very Low (< 10% at 1W) | Very High (> 80% at 1W) |
| Output Filter Required | No (Direct to speaker) | Yes (LC low-pass filter to remove PWM carrier) |
| EMI / RFI Noise | Negligible | High (Requires careful PCB layout and shielding) |
| Clipping Behavior | Soft, symmetrical saturation | Hard clipping, potential for high-frequency artifacts |
| Typical Use Case | Audiophile 2-channel, studio monitors, guitar amps | Subwoofers, portable Bluetooth speakers, multi-channel AVRs |
For a deeper dive into the semiconductor physics governing these output stages, the All About Circuits semiconductor textbook provides excellent schematic breakdowns of the internal driver networks.
Frequently Asked Questions
Why does my class AB amplifier get hot even when no music is playing?
This is entirely normal and is a direct result of the quiescent bias current ($I_q$). To prevent crossover distortion, the amplifier intentionally forces a continuous DC current (typically 50mA to 150mA) through the output transistors and emitter resistors even when the input signal is zero. This current multiplied by your total supply voltage equals the idle heat dissipation. If the heatsink is too small, the amp will overheat at idle; always size your heatsink for the idle thermal load, not just the maximum musical peaks.
Can I replace Class AB output BJTs with MOSFETs?
Yes, but you cannot simply drop them in without modifying the bias network. Standard power BJTs turn on at roughly 0.6V ($V_{be}$), while standard vertical MOSFETs require a gate-source threshold voltage ($V_{gs}$) of 2.0V to 4.0V to begin conducting. If you swap to MOSFETs, your $V_{be}$ multiplier or diode string must be adjusted to provide this higher bias voltage. Alternatively, you can use Lateral MOSFETs (like the Exicon ECX10N20), which are specifically designed for audio and have a much lower, BJT-like threshold voltage and a negative temperature coefficient that inherently prevents thermal runaway.
How do I set the bias current on a DIY class AB amplifier kit?
You must measure the voltage drop across one of the output transistor's emitter resistors (usually a 0.22$\Omega$ or 0.33$\Omega$ 5W ceramic resistor). With the input shorted to ground and no speaker connected, power the amp through a bulb limiter or variac. Connect your multimeter in DC millivolt mode across the emitter resistor. Adjust the bias trimpot on the driver board until you read the target voltage. For example, if you want 100mA of bias current and have a 0.33$\Omega$ resistor, use Ohm's law ($V = I \times R$): $0.1A \times 0.33\Omega = 33mV$. Adjust the pot until the meter reads exactly 33mV, then let the amp warm up for 30 minutes and re-check, as the reading will drift as the heatsink reaches thermal equilibrium.
Is a class AB amplifier better than class D for a home theater receiver?
It depends on the channel. For the critical front Left/Right and Center channels where dialogue clarity and analog linearity matter most, high-end AV receivers still use Class AB. However, for surround and height channels (like Dolby Atmos setups with 9 to 15 channels), Class D is vastly superior. A 15-channel Class AB receiver would require a massive, heavy chassis and enormous heatsinks to dissipate the idle heat of 15 bias networks. Class D allows manufacturers to pack 15 amplification channels into a slim, cool-running chassis without tripping a standard 15A household breaker.
