Switching Amplifier Theory: How Class D PWM Circuits Work

A switching amplifier (universally known as a Class D amplifier) is an electronic circuit that converts an analog input signal into a high-frequency pulse-width modulated (PWM) square wave, using power MOSFETs as on/off switches rather than linear variable resistors to drive a load. By keeping the output transistors either fully on (saturation) or fully off (cutoff), the circuit minimizes the voltage drop across the transistors while current is flowing, drastically reducing wasted heat. This fundamentally changes thermal management and power supply sizing in real installations, eliminating massive extruded aluminum heat sinks and heavy copper transformers, and shrinking a 500W rack amplifier into a 4-ounce PCB.

The Core Mechanism: PWM and the LC Filter

To understand how a switching amplifier operates, you have to look past the audio signal and look at the switching frequency. The process relies on comparing a low-frequency analog audio wave against a high-frequency triangle wave.

The signal chain follows a strict numbered sequence:

1. Modulation: An internal oscillator generates a triangle wave, typically between 300 kHz and 600 kHz. A comparator checks the analog audio input against this triangle wave.
2. PWM Generation: When the audio voltage is higher than the triangle wave, the comparator outputs a logic HIGH. When it drops below, it outputs a LOW. This creates a PWM square wave where the duty cycle directly mirrors the audio amplitude.
3. Power Stage: This low-power PWM signal drives the gates of a half-bridge (or full-bridge) of power MOSFETs. The MOSFETs switch the high-voltage DC rail to the output, creating a high-power, high-voltage PWM square wave.
4. Demodulation (LC Filter): An inductor-capacitor (LC) low-pass filter sits between the MOSFETs and the speaker. It blocks the 400 kHz+ switching frequency and passes the reconstructed analog audio frequencies (20 Hz - 20 kHz) to the load.

The Math: Linear vs. Switching Efficiency

The primary reason the industry adopted the switching amplifier topology is efficiency. Let us run a worked numeric example to see what this means for your power supply and thermal design.

Assume you are designing an amplifier to deliver 100W RMS into a 4-ohm subwoofer. We will compare a traditional Class AB linear amplifier (typically 55% efficient at full power) against a modern Class D switching amplifier (typically 92% efficient).

In the Class AB design, 81.8 watts of energy is burned off as heat. You need a heavy heat sink and a power supply capable of delivering nearly 182 watts continuously. In the Class D design, only 8.7 watts is lost to heat. The power supply only needs to provide 109 watts, and the MOSFETs can often be cooled by the ambient airflow inside the enclosure without any dedicated heat sinking.

Where You Meet Switching Amplifiers in Practice

You are likely already using switching amplifiers daily, even if you do not realize it. They have entirely conquered specific market segments due to their size and thermal advantages:

• Consumer Soundbars and Smart Speakers: Devices like the Sonos Era 300 use Class D amplifiers (often custom silicon or Texas Instruments TAS58xx series) because they must fit multiple high-power channels into sealed, unventilated plastic enclosures.
• Active PA Speakers: Modern powered speakers, like the QSC K12.2, use Class D modules paired with Switch Mode Power Supplies (SMPS). This allows a 2000W peak speaker to weigh under 40 lbs, whereas a 1990s equivalent would weigh over 100 lbs.
• DIY and Maker Projects: The TPA3116D2 is a ubiquitous, low-cost (under $15) Class D IC found on countless generic amplifier boards. It is the go-to chip for hobbyists building custom active crossovers or garage audio systems.
• Non-Audio Haptic Drivers: Because a switching amplifier is essentially a high-frequency AC voltage source, engineers frequently repurpose Class D audio ICs to drive piezoelectric actuators, ultrasonic transducers, and precision haptic feedback motors.

Real-World Scenario: The TPA3116D2 Subwoofer Build and the EMI Trap

Switching amplifiers are highly efficient, but their high-frequency square waves make them notorious for Electromagnetic Interference (EMI) if the output filter is poorly designed. Here is a real-world bench scenario that illustrates this.

The Setup: A DIY builder constructs a 2.1 active speaker system using a generic TPA3116D2 2x50W board, bridged to drive a single 4-ohm 10-inch subwoofer. The board is powered by a 24V 5A laptop power brick. The switching frequency is configured to 600 kHz via an external timing resistor.

The Numbers: The board's output filter uses a cheap, unshielded 10µH drum inductor and a 0.68µF film capacitor. Using the LC cutoff frequency formula ($f_c = \frac{1}{2\pi\sqrt{LC}}$), the filter's cutoff is roughly 61 kHz.

The Outcome: The subwoofer played loudly and cleanly. However, a Software Defined Radio (SDR) on the desk three feet away showed a massive noise floor spike, and a nearby ESP32 microcontroller's ADC was returning erratic, noisy sensor readings.

What Went Wrong: The 61 kHz cutoff frequency was too close to the audio band and too far below the 600 kHz switching frequency, failing to adequately attenuate the high-frequency harmonics. Worse, the unshielded drum inductor was acting as an antenna, radiating the 600 kHz fundamental and its odd harmonics directly into the air.

The Fix: The builder swapped the unshielded inductor for a 15µH shielded ferrite core inductor and added a common-mode choke on the speaker output wires. This dropped the radiated EMI by over 40dB, clearing up the SDR and stabilizing the ESP32's ADC without altering the audio quality.

Common Confusions: Digital Amps and Switching Power Supplies

When discussing switching amplifiers, two major points of confusion frequently arise on forums and in datasheets.

Confusion 1: 'Digital' Amplifiers
Marketing departments love to label Class D amplifiers as 'Digital Amplifiers.' This is technically false. The PWM modulation stage in a standard Class D amp (like the TPA3116D2) is an entirely analog process comparing two continuous voltage waveforms. It does not process binary logic, DSP, or I2S data streams. The only time a switching amplifier is truly 'digital' is when it integrates a digital input stage (like the Texas Instruments TAS5805M, which accepts I2S audio and handles the PWM conversion in the digital domain before hitting the analog power stage).

Confusion 2: Switching Power Supplies (SMPS)
People often confuse switching amplifiers with the Switch Mode Power Supplies that power them. While both use high-frequency MOSFET switching and inductors, their goals are opposite. An SMPS switches a DC voltage to regulate a steady, constant DC output voltage regardless of the load. A switching amplifier switches a DC voltage to regulate a constantly changing AC current that perfectly tracks an audio input waveform.

Frequently Asked Questions

For deeper technical analysis on modulation schemes and filter design, refer to the Texas Instruments Class-D Audio Amplifier basics application note and the comprehensive topology breakdowns available on All About Circuits.