An amplifier's class (A, B, or AB) defines the portion of the input signal cycle during which its output transistors conduct current, directly dictating its efficiency, heat dissipation, and audio distortion profile. If you are designing an audio stage or selecting a power amp module, this single letter determines whether you need a massive extruded aluminum heatsink or just a small PCB copper pour. Understanding these topologies is the difference between a board that runs cool and one that melts your solder joints.
Conduction Angles and the Spec Sheet Breakdown
The fundamental difference between linear amplifier classes comes down to biasing. In a Class A topology, the output device is biased so that it conducts current for the entire 360° of the input signal cycle. It never turns off. Think of Class A as a single security guard working a 24-hour shift; they are always on duty, which means zero handoff delays but massive fatigue (heat).
In a Class B topology, the circuit uses a push-pull configuration where one transistor handles the positive half of the waveform (180°) and another handles the negative half. At idle, with zero input signal, both transistors are completely off. This eliminates idle heat but introduces a new problem: crossover distortion, which occurs when the signal crosses the 0V threshold and one transistor turns off before the other fully turns on.
| Class | Conduction Angle | Max Theoretical Efficiency | Quiescent (Idle) Current | Primary Distortion Type | Typical Heat Sink Requirement |
|---|---|---|---|---|---|
| Class A | 360° | 50% (Push-Pull) | Very High (Max) | Thermal drift / Clipping | Massive (Sized for idle power) |
| Class B | 180° | 78.5% | Zero | Crossover (Dead-zone) | Small (Sized for ~40% output) |
| Class AB | 181° - 359° | 50% - 78.5% | Low (mA range) | Minor Crossover / Thermal | Moderate |
| Class C | < 180° | > 80% | Zero | Severe (Requires LC tank) | Small |
Note: Class D (switching) is excluded from this table as it relies on PWM modulation rather than linear conduction angles. For a deeper dive into switching topologies, refer to this Texas Instruments application report on amplifier classes.
Worked Numeric Example: Sizing the Heat Sink for 50W Output
To see what this topology choice changes in a real circuit, let us design a linear audio amplifier to deliver 50W RMS into an 8Ω load. We will calculate the required thermal resistance ($\theta_{SA}$) for the heatsink in both Class A (push-pull) and Class B configurations.
The Class A Scenario (50% Max Efficiency)
To get 50W of audio output at maximum theoretical efficiency, the power supply must deliver 100W of DC power.
- At full output: 100W in - 50W out = 50W dissipated as heat.
- At idle (no music playing): 100W in - 0W out = 100W dissipated as heat.
Assuming a maximum silicon junction temperature ($T_j$) of 150°C and a worst-case ambient room temperature ($T_a$) of 25°C, our allowable temperature rise ($\Delta T$) is 125°C.
Required Thermal Resistance: $\theta_{SA} = \Delta T / P_{diss} = 125°C / 100W = 1.25°C/W.
A thermal resistance of 1.25°C/W requires a massive, heavy finned aluminum extrusion, likely spanning the entire width of a standard 17-inch audio chassis, or forced-air cooling.
The Class B Scenario (78.5% Max Efficiency)
At full 50W output, the supply delivers roughly 63.7W. The heat dissipated at full power is only 13.7W. However, in Class B, maximum transistor dissipation does not occur at full power or at idle; it peaks when the output voltage swing is roughly 63% of maximum.
Using the standard dissipation formula for Class B ($P_{diss(max)} = 2 \times V_{cc}^2 / (\pi^2 \times R_L)$), and knowing that $V_{cc}^2 / R_L = 100$ for a 50W output, the maximum total heat dissipation is $200 / 9.87 = 20.26W.
- At idle: 0W dissipated.
- At peak internal dissipation: 20.26W dissipated.
Required Thermal Resistance: $\theta_{SA} = 125°C / 20.26W = 6.16°C/W.
A 6.16°C/W heatsink is a small, inexpensive piece of extruded aluminum that can easily be mounted directly to the PCB. This is the exact reason Class B (and its cousin, Class AB) dominates consumer and professional audio.
Where You Meet This in Practice and Common Confusions
If you are tearing down gear or reading spec sheets, you will rarely see 'Pure Class B' in audio. Here is where these topologies actually live and what people get wrong about them.
Real-World Applications
- Class A: Found in high-end audiophile headphone amplifiers (like the Pass Labs XA series) and RF low-noise amplifiers (LNAs) where absolute minimum distortion is required and power efficiency is irrelevant. You will also find it in the input voltage-gain stages of almost all amplifiers, where signal levels are low and heat is negligible.
- Class AB: The undisputed king of analog audio. By adding a small bias current (usually 20mA to 100mA) to a Class B circuit, both transistors stay slightly 'on' around the 0V crossover point. This eliminates the dead-zone distortion while retaining most of the Class B efficiency. Professional PA amplifiers, AV receivers, and guitar amp solid-state stages use Class AB.
What People Commonly Confuse
Many hobbyists and even some marketing departments label amplifiers as 'Class B' when they are actually Class AB. Pure Class B has a severe crossover dead-zone (typically ~0.6V for BJTs or ~2V for MOSFETs) that sounds like harsh buzzing on low-level signals. If an audio amp sounds clean at low volumes, it is biased into Class AB, not pure Class B.
Do not confuse the linear conduction angles of A/B/AB with Class D. Class D is a switching topology that uses Pulse Width Modulation (PWM) to toggle MOSFETs fully on or fully off, achieving >90% efficiency. Class D relies on LC output filters to reconstruct the audio waveform, whereas Class A/B/AB rely on the transistors operating in their linear (active) region. For a solid breakdown of linear vs switching behavior, Electronics Tutorials provides excellent schematic comparisons.
Frequently Asked Questions
Does a Class A amp actually sound better than Class AB?
Objectively, a well-designed Class AB amp with high slew rate and adequate bias current will measure identically to Class A on modern audio analyzers (THD+N below 0.001%). Subjectively, Class A amps avoid 'crossover artifacts' entirely and tend to run in single-ended configurations that produce even-order harmonic distortion, which the human ear finds pleasing. However, you are paying a massive premium in electricity and heatsink weight for a marginal psychological and acoustic benefit.
Why do Class A amps take 30 minutes to 'warm up' and sound good?
This is purely a thermal issue. Class A circuits generate immense heat, which changes the $V_{BE}$ (base-emitter voltage) of bipolar transistors or the $V_{GS}$ of MOSFETs. As the silicon heats up, the bias current drifts. High-end Class A amps use complex thermal-tracking bias circuits (like the Vbe multiplier) mounted directly to the output heatsink. The amp only reaches its designed quiescent current—and optimal linear operating point—once the massive heatsink reaches thermal equilibrium, which can easily take 30 to 45 minutes.
Can I modify a Class AB amp to run in pure Class A?
You can increase the bias current by adjusting the Vbe multiplier trimpot, but you will likely destroy the output stage. The existing heatsinks are sized for Class AB dissipation (e.g., 20W). If you bias the amp to draw 5A of idle current at 40V rails to achieve 'Class A' operation, you are dumping 200W of heat into a heatsink rated for 20W. The thermal runaway will destroy the output transistors in seconds unless you completely rebuild the thermal management system.
