A class A/B amplifier is a linear electronic circuit that combines the low-distortion continuous conduction of Class A at low signal levels with the higher power efficiency of Class B at high signal levels by applying a small quiescent bias current to its output transistors. What this topology changes in a real circuit is the thermal management requirement and idle power draw, trading the extreme heat of pure Class A and the high-frequency switching noise of Class D for a predictable, manageable middle ground. Beginners commonly confuse it with Class D (which uses high-frequency PWM switching) or Class G/H (which dynamically switches power supply rails to improve efficiency).
The Core Trade-Off: Quiescent Current vs. Crossover Distortion
To understand the class A/B amplifier, you must first look at the flaw in a pure Class B design. A Class B push-pull output stage uses two complementary transistors (NPN and PNP). One handles the positive half of the audio waveform, and the other handles the negative half. The problem is that silicon bipolar junction transistors (BJTs) require a base-emitter voltage (Vbe threshold of ~0.6V) to begin conducting.
When the audio signal crosses the zero-volt axis, neither transistor is fully turned on. This creates a dead zone known as crossover distortion, which manifests as a harsh, notch-like artifact in the audio output. Think of a two-lane highway merging into a tunnel where both lanes close for a split second during the handoff; traffic (the audio signal) jolts to a halt and surges forward, creating a ripple.
The class A/B amplifier solves this by injecting a small, continuous bias current (quiescent current, or Iq) through both transistors even when there is no input signal. This keeps both devices slightly "on" and operating in their linear region during the zero-crossing, effectively erasing the dead zone. The trade-off is that this continuous current flow generates heat, requiring substantial heatsinking even when the amplifier is idling.
Worked Numeric Example: Biasing and Thermal Dissipation
Let us calculate the real-world thermal and power parameters for a discrete class A/B amplifier output stage using the ON Semiconductor MJL21193/MJL21194 complementary power transistor pair.
- Power Supply: Dual ±40V DC rails (80V total rail-to-rail).
- Load: 8-ohm loudspeaker.
- Quiescent Bias Current (Iq): 50mA.
First, we calculate the theoretical maximum RMS output power. The peak voltage is 40V. The RMS voltage is 40V / √2 = 28.28Vrms. Using the power formula P = V² / R, we get 28.28² / 8 = 100W RMS. While a pure Class B stage has a theoretical maximum efficiency of 78.5%, our class A/B stage will operate closer to 65-70% efficiency at full power due to the bias current and Vce(sat) voltage drops across the transistors.
Now, we calculate the idle thermal dissipation, which dictates your heatsink sizing. With an Iq of 50mA and a total rail voltage of 80V, the idle power dissipation is:
Pq = Vtotal × Iq = 80V × 0.05A = 4W per transistor.
Since there are two output devices, the stage dissipates 8W of heat before a single note of music plays. If a designer decides to push the bias to 200mA to achieve a "more Class A-like" sound signature, the idle dissipation jumps to 32W. This requires massive extruded aluminum heatsinks and careful thermal compensation to prevent thermal runaway.
As output transistors heat up, their Vbe threshold drops, causing them to draw more bias current, which generates more heat in a destructive feedback loop. A class A/B amplifier must use a Vbe multiplier transistor mounted directly to the heatsink to dynamically reduce the bias voltage as the chassis temperature rises.
Where You Meet Class A/B Amplifiers in Practice
Despite the rise of switching amplifiers, the class A/B topology remains a staple in specific high-performance audio applications:
- Mid-to-High-Tier AV Receivers: Brands like Denon and Marantz frequently use discrete class A/B output stages in their AVR-X series. These amplifiers must deliver high instantaneous current into complex, low-impedance speaker loads (often dipping to 3 ohms at certain frequencies) without the electromagnetic interference (EMI) filtering headaches associated with Class D.
- Critical Studio Monitoring: While portable PA systems have universally adopted Class D for weight savings, critical listening environments often rely on class A/B designs for their ultra-low noise floors and predictable phase response. For deeper technical analysis on why this topology remains the benchmark for low-feedback stability, refer to the Elliott Sound Products amplifier class guide.
- DIY Audio and Chipamps: The legendary Texas Instruments LM3886 (and modern discrete equivalents built around op-amps like the OPA1612) remains highly popular in the DIY community. Its internal class A/B architecture provides excellent thermal protection and straightforward compensation networks.
Class A/B vs. Class D and Class G/H
Understanding how class A/B stacks up against other common topologies helps you choose the right architecture for your build or purchase.
| Topology | Max Theoretical Efficiency | Crossover Distortion | EMI / Filtering Needs | Typical Use Case |
|---|---|---|---|---|
| Class A/B | ~78.5% (Practical: 65%) | Eliminated via bias | Very Low | AVRs, Studio Monitors, DIY Hi-Fi |
| Class D | ~90-95% | N/A (Switching noise instead) | High (Requires LC output filters) | Portable PA, Subwoofers, Soundbars |
| Class G/H | ~85-90% | Eliminated via bias + rail switching | Moderate | Pro Audio Touring Amps, High-end AVRs |
Choose Class A/B when you need a low-noise floor, simple output filtering, and excellent high-frequency linearity without worrying about switching artifacts. Choose Class D when thermal constraints, weight, and battery life are the primary design drivers.
Frequently Asked Questions
Why does my class A/B amplifier get so hot even with no music playing?
This is entirely normal and is a direct result of the quiescent bias current (Iq) required to eliminate crossover distortion. A typical 100W class A/B amplifier will dissipate between 10W and 25W of heat at idle. The heatsinks are sized to handle this continuous thermal load while keeping the transistor junction temperatures safely below 85°C. If the heatsinks are cool to the touch at idle, the amplifier is likely biased too low and may exhibit crossover notch distortion.
Is a class A/B amplifier better than class D for high-fidelity audio?
Historically, class A/B was the undisputed king of high-fidelity audio due to its lack of high-frequency switching noise and superior phase linearity. However, modern Class D architectures utilizing advanced feedback loops and high switching frequencies (such as Hypex Ncore or Purifi Eigentakt modules) now routinely surpass class A/B in Total Harmonic Distortion plus Noise (THD+N) measurements. Today, class A/B wins on simplicity, lower EMI, and predictable behavior with highly reactive speaker loads, rather than raw distortion metrics.
How do I measure the correct bias voltage on a class A/B output stage?
Never measure bias current directly in-circuit without breaking the power rail. Instead, measure the voltage drop across the emitter resistors of the output transistors. For example, if your design uses 0.22-ohm emitter resistors and targets an Iq of 50mA, use Ohm's Law (V = I × R) to find your target voltage: 0.05A × 0.22Ω = 11mV. Use a high-quality 4.5-digit multimeter set to the millivolt range, measure across the emitter resistor, and adjust the Vbe multiplier trimmer potentiometer until you hit exactly 11mV. Always allow the amplifier to thermally stabilize for 20 minutes before taking the final measurement, as the reading will drift as the heatsink warms up.
