How Class D Amplifiers Use PWM to Exceed 90% Efficiency

This article provides a comprehensive overview of how Class D switching amplifiers achieve over 90% power efficiency in modern audio electronics. By abandoning the traditional continuous linear operation of older amplifier designs, Class D technology utilizes high-frequency pulse-width modulation (PWM) to rapidly switch output transistors between fully on and fully off states. The following sections break down the mechanics of PWM conversion, the role of MOSFET switching in minimizing thermal power dissipation, and the final low-pass filtering stage that reconstructs the amplified analog audio signal for the speaker.

The Limitation of Traditional Linear Amplifiers

To understand the efficiency of Class D amplifiers, it is necessary to look at why traditional amplifiers waste power. Class A, B, and AB amplifiers operate in a linear mode. Their output transistors act like variable resistors, constantly regulating voltage to mirror the incoming analog audio signal.

Because these transistors are often partially on, they simultaneously experience both significant voltage across them and current flowing through them. In electronics, power dissipation (heat) is calculated as Voltage multiplied by Current (\(P = V \times I\)). Consequently, linear amplifiers burn off 30% to 75% of their drawn power as heat, requiring large, heavy heat sinks.

High-Frequency Pulse-Width Modulation (PWM)

Class D amplifiers solve the heat problem by never operating their transistors in a partially open, resistive state. Instead, they convert the analog audio input into a stream of digital-like pulses using Pulse-Width Modulation (PWM).

This process requires two components:

  1. The Input Audio Signal: The standard analog waveform.
  2. A High-Frequency Carrier Wave: A triangle or sawtooth waveform generated internally, typically running between 250 kHz and 1.5 MHz—far above the 20 kHz limit of human hearing.

A circuit component called a comparator continuously compares the voltage of the audio signal against the voltage of the high-frequency triangle wave. When the audio signal’s voltage is higher than the triangle wave, the comparator outputs a high voltage. When the audio signal is lower, it outputs a low voltage.

The result is a square wave where the width of each pulse corresponds directly to the amplitude of the original audio signal at that exact microsecond. High audio peaks create wide pulses; quiet audio passages create narrow pulses.

The Switching Stage and Power Retention

Once the audio is encoded into a PWM signal, it is sent to the switching output stage, which relies on MOSFETs (Metal-Oxide-Semiconductor Field-Effect Transistors). Unlike the transistors in linear amplifiers, these MOSFETs act as binary switches. They are either completely off or completely on.

This binary operation is the core reason Class D amplifiers exceed 90% efficiency:

  • When the switch is OFF: Resistance is infinite, meaning current (\(I\)) is zero. Even though voltage (\(V\)) is high, \(V \times 0 = 0\) watts of power lost.
  • When the switch is ON: The MOSFET has near-zero resistance, meaning the voltage drop (\(V\)) across the transistor is nearly zero. Even though current (\(I\)) is high, \(0 \times I = 0\) watts of power lost.

In reality, transistors are not perfect. A tiny amount of power is lost due to the slight resistance of the MOSFET when on (conduction loss) and the brief nanosecond it takes the transistor to transition from on to off (switching loss). However, because the transistors spend almost no time in the transitional state, heat generation is negligible, allowing over 90% of the power supply's energy to go directly toward the audio output.

Reconstructing the Audio Signal

A loudspeaker cannot properly play a high-frequency PWM square wave, and feeding megahertz-level frequencies into a speaker coil would waste energy and generate electromagnetic interference (EMI). The final step in a Class D amplifier is converting the high-power pulses back into a smooth analog waveform.

Before reaching the speaker terminals, the amplified PWM signal passes through a passive low-pass filter, consisting of inductors and capacitors (an LC filter). This filter blocks the high-frequency switching carrier (e.g., 500 kHz) while allowing the low-frequency audio band (20 Hz to 20 kHz) to pass through unhindered.

The inductor resists rapid changes in current, effectively averaging out the wide and narrow pulses. What emerges from the filter is an exact, high-voltage replica of the original analog audio input, ready to drive the speaker with minimal wasted energy.