The One-Sentence Definition and What It Actually Changes

An amp class categorizes how an amplifier's output transistors are biased and conduct current relative to the input signal cycle, directly determining its efficiency and distortion profile. When you select an amplifier topology, you aren't just choosing a sound signature; you are fundamentally dictating the thermal management, power supply headroom, and physical footprint of your circuit.

What it changes in a real circuit: The amp class dictates your quiescent current draw (idle power), the physical size of your heatsinks, the complexity of your power supply unit (PSU), and the type of output filtering required before the signal reaches the load.

Think of the output transistors like water valves controlling flow to a turbine. A Class A valve is always wide open, bleeding off excess water as heat even when no turbine work is needed. A Class D valve rapidly snaps fully open and fully closed, wasting almost no water but requiring a downstream filter to smooth the pulsing flow.

People routinely confuse "Class D" with "Digital." Class D is strictly an analog switched-mode topology; the "D" simply followed "C" alphabetically, not conceptually. Another common trap is assuming a higher letter means a generational upgrade (e.g., Class D is "better" than Class A). In reality, it is a lateral move trading thermal efficiency for high-frequency switching noise and output filter complexity. For a deeper look at the underlying semiconductor physics, the Analog Devices MT-047 Tutorial remains the benchmark reference for conduction angles and biasing.

The Math: Efficiency, Heat, and a 100W Worked Example

Let's calculate the thermal load for a 100W RMS audio amplifier driving a 4-ohm subwoofer. This is where the theoretical conduction angles hit the workbench.

  • Class A (360° conduction): Max theoretical efficiency is 25% (practically closer to 20%). To get 100W of acoustic power, the PSU must deliver ~500W. The output transistors must dissipate 400W of continuous heat. You need a massive heatsink, roughly the size of a car battery, and a heavily regulated linear power supply.
  • Class AB (>180° conduction): Max theoretical efficiency is 78.5% (practically ~60%). For 100W out, the PSU delivers ~166W. Heat dissipation drops to 66W. This is manageable with a standard extruded aluminum finned heatsink and a standard toroidal transformer.
  • Class D (Switched-mode PWM): Practical efficiency sits around 90-95%. For 100W out, the PSU delivers ~105W. Heat dissipation is a mere 5W. You can mount the IC directly to the PCB with a tiny stamped metal clip, and run it off a compact SMPS (Switched-Mode Power Supply).
Bench Tip: When designing a Class AB board, always calculate your heatsink thermal resistance ($\theta_{SA}$) using the worst-case dissipation, which actually occurs at roughly 1/3 to 1/2 of maximum power output, not at full clipping.

Where You Meet Amp Classes in Practice

You will encounter these topologies across wildly different price points and form factors. Here is where each class dominates the current market:

  • Guitar Pedals & Studio Preamps (Class A): Discrete op-amps like the API 2520 or Neve 1073 clones run in Class A. Designers accept the massive heat and low efficiency to achieve zero crossover distortion and a specific profile of even-order harmonic saturation when driven hard.
  • AV Receivers & Powered Studio Monitors (Class AB): Chips like the TI LM4780 or STMicroelectronics TDA7294 are the workhorses here. They offer a balance of low THD+N (Total Harmonic Distortion plus Noise) and manageable heat in multi-channel chassis where pure Class A would melt the plastic fascias.
  • Subwoofers & Portable Bluetooth Speakers (Class D): Modules based on the TI TPA3116D2 or Hypex Ncore NC252MP dominate this space. When battery life, compact enclosure volume, or massive wattage (1000W+ plate amps) is the primary constraint, Class D is the only physically viable option.

The Amp Class Decision Tree: Which One to Pick

Stop debating forum posts and use this decision matrix to select your topology and specific silicon for your next build.

Application Constraint Recommended Class Concrete IC / Module Pick
Audiophile 2-channel near-field listening; heat and PSU cost are irrelevant; lowest possible THD is mandatory. Class A DIY Nelson Pass Aleph-J topology or TI LME49600 buffered discrete design.
Multi-channel home theater or powered studio monitors; need low noise floor without melting a compact chassis. Class AB Texas Instruments LM4780 (parallelable) or STMicroelectronics TDA7293.
High-power subwoofers, PA systems, or portable battery-powered audio; maximum wattage per cubic inch required. Class D Texas Instruments TPA3255 (up to 2x315W) or Hypex NC252MP module.
High-end AV receivers needing AB sound quality but D-class efficiency; willing to design complex multi-rail PSUs. Class G / H Audio Partnership (Arcam) proprietary designs or custom rail-switching discrete builds.

Common Myths and Misconceptions

Q: Does Class D sound "cold" or "digital" compared to Class AB?
A: This is a legacy myth from early 2000s Class D silicon (like first-gen Tripath chips). Modern Class D ICs, particularly those utilizing advanced feedback loops like the TI Purifi or Hypex Ncore architectures, routinely achieve THD+N figures below 0.001% and flat frequency response up to 20kHz. Any "coldness" is usually the result of poorly designed LC output reconstruction filters interacting with complex speaker impedance curves, not the switching topology itself. Consult the Texas Instruments Audio Amplifiers Overview for modern measurement data that debunks this.

Q: What exactly are Class G and Class H?
A: They are not fundamentally new output stages; they are clever power supply tricks applied to Class AB amplifiers. Class G uses multiple fixed voltage rails and switches between them depending on signal demand. Class H continuously modulates the supply rail voltage to track the audio signal. Both drastically reduce the heat of a Class AB amp without introducing the high-frequency switching noise of Class D, but they require highly complex, expensive power supply designs.

Q: Can I parallel Class D amplifier ICs for more power?
A: Generally, no. Unlike Class AB chips (which often feature built-in paralleling pins to average out offset voltages), paralleling Class D outputs usually results in catastrophic failure. The high-frequency PWM carrier signals will beat against each other, causing massive circulating currents between the output stages that will instantly fry the MOSFETs. If you need more power from Class D, buy a higher-rated single IC or use a bridge-tied load (BTL) configuration if the datasheet supports it.

The Default Recommendation: For 90% of DIY makers building powered speakers, subwoofers, or desktop audio systems today, default to a Class D module based on the Texas Instruments TPA3255. The thermal savings, elimination of massive toroidal transformers, and drastic reduction in PCB space overwhelmingly outweigh the negligible THD differences at audible frequencies. Reserve Class AB for specific low-power, ultra-low-noise preamp stages, and leave Class A to those who enjoy heating their listening rooms in the winter.