An amplifier class defines the biasing and conduction angle of its output transistors, dictating exactly how much of the input signal cycle the active devices are turned on. In a real circuit or installation, the chosen class fundamentally changes your power supply sizing, heatsink mass, PCB copper weight, and overall battery life. Beginners commonly confuse amplifier classes (A, B, AB, D) with amplifier topologies (single-ended, push-pull, bridged), or falsely assume that 'Class D' stands for 'Digital' rather than its actual switching/pulse-width modulated operation.
The Core Amplifier Classes: Bias, Conduction, and Efficiency
Every amplifier must balance linearity (how accurately it reproduces the input) against efficiency (how much DC power it converts to AC output versus waste heat). The class of operation sets the baseline for this compromise by determining the quiescent current—the current flowing through the output devices when no audio signal is present.
| Class | Conduction Angle | Theoretical Max Efficiency | Typical Real-World Efficiency | Primary Trade-off |
|---|---|---|---|---|
| Class A | 360° (Full cycle) | 50% (Transformer-coupled) | 15% - 30% | Massive heat generation; zero crossover distortion. |
| Class B | 180° (Half cycle) | 78.5% | 50% - 70% | Severe crossover distortion; rarely used alone in audio. |
| Class AB | 180° to 360° | 78.5% | 50% - 65% | Compromise between Class A linearity and Class B efficiency. |
| Class D | Switching (PWM) | 100% | 85% - 95% | Requires output LC filters; EMI management is critical. |
While Class A keeps transistors fully biased on at all times to avoid the non-linear transition zones of silicon, Class AB introduces a small bias voltage to keep both push-pull devices slightly on during the zero-crossing. Class D abandons the linear region entirely, using output MOSFETs as high-speed switches operating at 300 kHz to 1.2 MHz, modulating the pulse width to encode the audio signal before a passive LC filter reconstructs the analog waveform.
Worked Thermal Example: Sizing a Heatsink for Class-AB vs Class-D
To understand what amplifier classes change on your workbench, let us calculate the thermal management required for a 50W RMS output into an 8-ohm load, comparing a traditional Class-AB chip (like the legendary LM3886 or TDA7294) against a modern Class-D IC (like the TI TPA3116D2).
Class-AB Thermal Math
Assume a realistic maximum efficiency of 60% for our Class-AB amplifier at full power.
- Power drawn from supply: 50W / 0.60 = 83.3W
- Heat dissipated (Pd): 83.3W - 50W = 33.3W
Now, we calculate the required heatsink thermal resistance ($\theta_{SA}$). We want to keep the silicon junction temperature ($T_J$) below 125°C (derated from the 150°C absolute max for reliability) in a 40°C ambient environment ($T_A$). Assuming a junction-to-case thermal resistance ($\theta_{JC}$) of 1.5°C/W:
$\theta_{SA} = [(T_J - T_A) / P_d] - \theta_{JC}$
$\theta_{SA} = [(125 - 40) / 33.3] - 1.5 = 2.55 - 1.5 = 1.05°C/W$
A thermal resistance of 1.05°C/W requires a large, heavy, and expensive extruded aluminum fin array, often necessitating forced air cooling if enclosed in a chassis.
Class-D Thermal Math
Assume a realistic efficiency of 90% for our Class-D amplifier.
- Power drawn from supply: 50W / 0.90 = 55.5W
- Heat dissipated (Pd): 55.5W - 50W = 5.5W
Using the same thermal formula:
$\theta_{SA} = [(125 - 40) / 5.5] - 1.5 = 15.45 - 1.5 = 13.95°C/W$
A requirement of 13.95°C/W is easily met by the exposed thermal pad on the IC soldered directly to a standard 2oz copper PCB pour, or a tiny $0.50 stamped metal clip. This is why modern Bluetooth speakers and soundbars can output massive wattage from sealed plastic enclosures without melting.
Where You Meet Amplifier Classes in Practice
You will encounter these classes in distinct corners of the electronics market, driven by the physical constraints of the end product.
- Class A: Reserved for boutique, ultra-high-end audiophile gear (e.g., Pass Labs XA series) and RF low-noise amplifiers (LNAs) where signal purity and noise floor are paramount, and power consumption is irrelevant. Expect massive transformer banks and chassis that double as space heaters.
- Class AB: The workhorse of traditional AV receivers, guitar amplifiers, and the output stages of nearly every operational amplifier (op-amp) you use in analog signal conditioning. It provides excellent linearity without the extreme thermal penalty of Class A.
- Class D: Dominates battery-powered devices, active studio monitors, subwoofers, and smart speakers. High-end implementations using GaN (Gallium Nitride) FETs or advanced modulation like Purifi Eigentakt and Hypex Ncore NC400 modules have effectively eliminated the 'harsh treble' stereotype of early Class-D designs, offering THD+N figures below 0.0005%.
- Class G / Class H: Often found in professional PA amplifiers. These are essentially Class-AB amplifiers with intelligent power supplies. Class G switches between discrete voltage rails based on signal demand, while Class H continuously tracks the audio envelope to modulate the supply voltage, minimizing the voltage drop across the output transistors and boosting efficiency to near Class-D levels.
For a deeper dive into the switching mechanics of modern audio ICs, review the Texas Instruments Class-D Audio Amplifier overview, which details dead-time insertion and EMI mitigation techniques critical for PCB layout.
Frequently Asked Questions About Amplifier Classes
Is Class D actually digital audio?
No. The 'D' in Class D does not stand for Digital; it was simply the next letter in the alphabet after Class C (which is used in RF oscillators, not linear audio). A Class-D amplifier is fundamentally an analog switching circuit. The input is an analog waveform, and the output of the LC filter is an analog waveform. The pulse-width modulation (PWM) process in the middle is a high-frequency analog switching technique, not a binary digital data stream like S/PDIF or I2S, though many modern Class-D chips integrate a digital front-end (DAC and DSP) for convenience.
Why do Class A amplifiers get so hot even with no music playing?
Because of quiescent bias current. In a Class-A design, the output transistors are biased heavily into their linear region so they never turn off. If a Class-A amplifier is designed to deliver 20W into 8 ohms, it might draw 50W to 80W continuously from the wall just sitting idle. The transistors must dissipate that full power as heat at all times to ensure that when a transient peak hits, the silicon is already conducting and perfectly linear, avoiding crossover distortion entirely.
What is the difference between Class G and Class H amplifiers?
Both aim to improve the poor efficiency of Class-AB by reducing the voltage dropped across the output transistors during low-level signals. Class G uses multiple fixed power supply rails (e.g., ±35V and ±70V) and switches between them using diodes or secondary transistor banks when the signal exceeds the lower rail. Class H uses a single supply rail that is continuously modulated (tracked) by the input signal envelope, keeping the supply voltage just slightly above the instantaneous audio peak. Class H is more complex to design but offers smoother transitions and better efficiency.
Can I use a Class D amplifier for full-range hi-fi audio?
Absolutely, provided you select a modern, high-performance module. Early Class-D chips (like the first-generation Tripath or basic TI TPA3118) suffered from high output impedance and load-dependent frequency response, which altered the treble depending on the speaker's impedance curve. Modern implementations using post-filter feedback or advanced self-oscillating topologies (like the theoretical models discussed in semiconductor texts and realized in Hypex Ncore or Purifi Eigentakt modules) exhibit flat frequency response into complex loads, ultra-low noise floors, and THD figures that rival or beat the best Class-AB designs on the market.
