An amplificateur classe (amplifier class) defines how the active output devices—transistors or tubes—conduct current relative to the input signal cycle, dictating the fundamental trade-off between power efficiency and signal fidelity. In a real circuit, this classification changes your DC biasing network, the physical mass of your heatsinks, and whether you must design an LC low-pass output filter to strip switching noise. Many hobbyists and even some audio engineers confuse the letter designations with digital logic or assume a higher letter implies newer, better technology; in reality, 'Class D' is an analog switching topology, and a poorly executed Class A will sound objectively worse than a well-engineered Class D.
The Core Trade-Off: Conduction Angles and Efficiency
The defining metric of any amplifier class is its conduction angle—the portion of the 360-degree input waveform during which the output device actively conducts current. Linear topologies (Class A, B, AB) act as variable resistors, burning excess voltage as heat. Switching topologies (Class D) act as rapid on/off switches, relying on pulse-width modulation (PWM) and external filters to reconstruct the analog waveform.
| Class | Conduction Angle | Max Theoretical Efficiency | Typical Real-World Efficiency | Common IC / Topology | Primary Use Case |
|---|---|---|---|---|---|
| A | 360° (Always on) | 25% (Transformer) / 50% (Inductive) | 15% - 25% | Discrete (JLH1969, Pass ACA) | High-end audiophile, low-power preamps |
| B | 180° (Push-pull) | 78.5% | 50% - 60% | Rarely used alone (crossover distortion) | Theoretical baseline for Class AB |
| AB | >180° and <360° | 78.5% | 45% - 65% | LM3886, TDA7294, Discrete BJTs | AV receivers, studio monitors, guitar amps |
| D | Switching (PWM) | 100% | 85% - 95% | TI TPA3116D2, ICEpower, Purifi | Active subs, portable BT speakers, pro PA |
| G | AB with rail switching | ~80% | 65% - 75% | Discrete + tracking PSU, STA516 | High-power studio mains, pro audio racks |
When evaluating modern silicon, GaN (Gallium Nitride) FETs are rapidly replacing silicon MOSFETs in Class D output stages. Because GaN devices switch with virtually zero reverse recovery charge, they allow Class D amplifiers to push switching frequencies past 1 MHz. This drastically shrinks the required output LC filter inductors, reducing PCB footprint and improving high-frequency phase response.
Worked Example: Heatsink Sizing for a 50W RMS Output
To understand what the amplifier class changes on your workbench, let's calculate the required heatsink thermal resistance ($\theta_{SA}$) for a 50W RMS continuous sine-wave output into an 8-ohm load. We will compare a traditional Class AB design using discrete MJL21193/MJL21194 complementary BJTs against a Class D design using the TI TPA3116D2.
Assumptions:
- Maximum junction temperature ($T_{J(max)}$): 150°C
- Ambient enclosure temperature ($T_A$): 40°C
- Junction-to-case thermal resistance ($\theta_{JC}$): 1.5°C/W
- Case-to-sink thermal resistance ($\theta_{CS}$, with thermal pad): 0.5°C/W
1. Class AB Calculation (Assuming 50% practical efficiency at continuous full power):
$$P_{OUT} = 50W$$ $$P_{IN} = P_{OUT} / 0.50 = 100W$$ $$P_{DISS} = P_{IN} - P_{OUT} = 50W$$
The formula for required heatsink thermal resistance is:
$$\theta_{SA} = \frac{T_{J(max)} - T_A}{P_{DISS}} - (\theta_{JC} + \theta_{CS})$$ $$\theta_{SA} = \frac{150 - 40}{50} - (1.5 + 0.5) = 2.2 - 2.0 = \mathbf{0.2°C/W}$$
2. Class D Calculation (Assuming 90% practical efficiency):
$$P_{OUT} = 50W$$ $$P_{IN} = P_{OUT} / 0.90 = 55.5W$$ $$P_{DISS} = P_{IN} - P_{OUT} = 5.5W$$
$$\theta_{SA} = \frac{150 - 40}{5.5} - 2.0 = 20.0 - 2.0 = \mathbf{18.0°C/W}$$
A thermal resistance of 18.0°C/W is easily achieved by a small, stamp-sized extruded heatsink, or simply by utilizing a 2-ounce copper pour on the PCB itself. This single mathematical reality is why Class D has entirely conquered the portable audio and active subwoofer markets.
Where You Meet This in Practice
You will encounter different amplifier classes based strictly on the physical and economic constraints of the end product:
- Home Audio & Audiophile: Class A and Class AB still dominate high-end discrete designs (e.g., Pass Labs, Benchmark AHB2). The massive transformers and heatsinks are accepted as part of the product's physical aesthetic and perceived value. Crossover distortion in Class AB is mitigated by heavy negative feedback and high bias currents.
- Active Subwoofers: Plate amplifiers built into the back of subwoofer enclosures are almost exclusively Class D (e.g., Dayton Audio SPA series). A Class AB amp inside a sealed wooden box would quickly trigger its thermal protection circuit due to trapped ambient heat.
- Battery-Powered Portables: Bluetooth speakers and portable PA systems rely on Class D ICs like the TPA3116D2 or NXP's TFA9894 smart amps. The high efficiency directly translates to longer battery life and allows the use of small, cheap lithium-ion packs.
- RF Transmitters & HAM Radio: While audio uses A, AB, and D, RF applications frequently use Class C. Class C conducts for less than 180° of the cycle, resulting in massive distortion for audio, but when paired with a high-Q LC resonant tank circuit, it reconstructs a clean sine wave at a single frequency with up to 85% efficiency.
FAQ: Clearing Up Amplifier Class Confusion
Is Class D a 'digital' amplifier?
No. The 'D' does not stand for Digital; it was simply the next available letter after Class C. Class D is an analog switching topology. The input is an analog audio signal that is compared against a high-frequency triangle wave to generate a PWM signal. While the control of a modern Class D amp might involve a DSP chip, the amplification stage itself is strictly analog switching.
Does Class A always sound better than Class D?
No. A poorly designed Class A amplifier will suffer from thermal drift, high noise floors, and rolled-off high frequencies due to output capacitor limitations. Conversely, modern Class D modules like the Purifi Eigentakt or Hypex Ncore achieve THD+N (Total Harmonic Distortion plus Noise) figures below 0.0005%, which is vastly superior to the noise floor of most Class A designs. The 'Class A sounds warmer' argument is usually a description of intentional even-order harmonic distortion, not superior fidelity.
What is the difference between Class G and Class H?
Both are variations of Class AB designed to improve efficiency. Class G uses multiple fixed power supply rails (e.g., ±35V and ±70V). The amplifier runs on the lower rail for small signals, switching to the higher rail only during loud transients. Class H uses a continuously variable (tracking) power supply that modulates the rail voltage in real-time to stay just slightly above the instantaneous audio signal voltage. Both reduce the voltage drop across the output transistors, cutting heat dissipation by roughly 30% compared to standard Class AB.
