Why Class D Amplifier Efficiency Is So High
A Class D amplifier is a power stage that drives a speaker using a pair of fast switches instead of a dimmer-like resistor. It sits at the very end of the signal chain, right before the speaker wire, replacing the linear output transistors found in older designs.
- Switch, not dimmer: The output stage acts like a pair of relays that snap open or shut, so almost no energy turns into heat the way a linear transistor does.
- The filter does the smoothing: A small inductor and capacitor between the raw switching pulses and the speaker shape square waves into a clean audio waveform.
- Efficiency above ninety percent: The only real power loss lands in the brief crossover moment when one switch turns off and the other turns on.
- Physical limits set the ceiling: Switching speed, dead time, and electromagnetic pickup in the crossover decide how smooth the output stays at every wattage.
What It Means, Plainly
Think of a light switch on your wall. When it is on, current flows with almost no resistance. When it is off, no current flows at all. In neither position does the switch itself waste energy as heat. A Class D output stage works the same way. It uses two transistors (often called MOSFETs, the fast modern version of the older bipolar transistor) wired in a pair. The upper switch connects the speaker wire to the power rail. The lower switch connects it to ground. They alternate: when the upper one is fully open, the lower one is fully shut, and vice versa. The result is a string of sharp square-wave pulses.
Those pulses are not what the speaker wants. A filter network — a small inductor (a coil that resists sudden changes in current) and a capacitor (a component that resists sudden changes in voltage) — sits between the raw switching and the speaker terminals. Together they smooth the jagged pulses into a clean, rounded audio waveform. The speaker sees a normal-looking signal, even though the amp behind it is just flicking two switches many thousands of times per second.
Because the switches spend their time either fully open (no voltage across them, so no power lost) or fully shut (no current through them, so again no power lost), the output stage wastes very little energy as heat. That is the core idea.
Where The Name Comes From
The word Class in amplifier naming traces back to the 1930s, when engineers described output transistors by how much of the waveform they conducted for. Class A meant the transistor ran for the full 360 degrees of a cycle; it was usually partially on, like a dimmer left somewhere between full and nothing. Class B cut that in half: each of two transistors handled 180 degrees. Class C went even shorter, conducting for less than half a cycle.
Then engineers asked a different question: what if the transistor is not a dimmer at all, but a snap-on, snap-off switch? The conduction-angle idea no longer fits, because a switch is either fully open or fully shut. There is no in-between position. The next available letter in the sequence was D, so the topology got that tag. The “D” does not encode an angle. It encodes a different philosophy: the output device is a relay (a fast, electronic switch), not a variable resistor. The name exists because the design logic changed from “control how much current flows” to “turn it on and off so fast that the average looks like a smooth signal.”
What It Is, And What It Is Not
What it is: A specific output-stage topology. Two transistors act as hard-on, hard-off switches. A filter inductor and capacitor smooth the resulting square waves into an audio signal. The topology sits inside a unit that may also have a preamp, a digital signal processor, and a power supply; those are separate blocks.
What it is not:
- Not a whole amplifier. “Class D” describes only the final output stage. The rest of the unit — input stage, equalizer, DAC, power supply — can be anything.
- Not a quality grade. A poorly laid-out Class D stage with a weak filter and long dead time will sound grittier than a well-tuned Class A stage. The topology changes the heat, not the ceiling on musicality.
- Not the same as “digital.” The input signal to a Class D stage can be fully analog. The switching happens in the analog domain. Calling it “digital” confuses the signal format with the output topology.
- Not “no heat.” The crossover moment, where both switches are briefly off, still dissipates some power. The amp does warm up; it just runs far cooler than a linear design at the same output level.
Confusables, Separated
| Look-alike term | What it actually means | The one-line tell |
|---|---|---|
| Class A output | One transistor conducts the full cycle, acting as a variable resistor | Dimmer, not switch |
| Class B or AB output | Two transistors each handle one half of the cycle, still in a partially conducting state | Two dimmers sharing the job, still linear |
| Pulse-width modulation (PWM) | The timing method that varies pulse width to encode signal amplitude | A tool inside a Class D design, not the topology itself |
| DSP amplifier | A unit that processes the signal digitally before it reaches the output stage | Describes the front end, not the output stage |
| Amplifier output impedance | How hard it is for a speaker to draw current from the amp | An electrical parameter, not a switching topology |
The Audible Consequence
Bass. The filter inductor and capacitor must be large enough to absorb the switching ripple at the lowest notes. If the filter is marginal, you may hear a faint grainy texture under the deepest kicks or the thump of a double bass. The square-wave edges leak through just enough to make low notes feel slightly rough instead of round. A well-sized filter keeps that ripple far below the ear’s sensitivity.
Vocals and mids. The crossover moment — the brief window when the upper switch is off and the lower switch is not yet on — is called dead time. In the midrange, roughly 2 to 5 kHz, dead time shows up as a thin, sibilant edge on consonants like “s” and “t.” Quieter passages reveal it most, because the tiny gap is a larger slice of the waveform when the signal is small. Good designs shrink dead time to a few nanoseconds, but it is not zero.
High frequencies. The switching frequency and its harmonics sit well above 20 kHz, so you do not hear the pulses themselves. What you can hear, if the layout is poor, is electromagnetic pickup: the fast voltage jumps at the filter inductor couple into nearby wires and act like a tiny antenna. Cymbals and hi-hats may carry a faint etch or sparkle that a linear amp would not add. Clean wiring, short filter loops, and proper shielding keep this to a minimum.
The practical limits on switch speed, dead time, and electromagnetic pickup in the crossover are what keep a Class D amp from gliding smoothly at every wattage. They are physical constraints, not design mistakes.
What Readers Ask
Why Does My Class D Amp Barely Get Warm While My Old Linear One Runs Hot?
In a linear amp, the output transistor sits somewhere between full on and full off for most of the time. That in-between state is where voltage across the transistor and current through it multiply into waste heat. A Class D output is either fully on or fully off, so that multiplication almost rarely happens. The filter inductor and capacitor handle the smoothing instead. The result is an efficiency above ninety percent, versus maybe sixty to seventy percent for a linear design, so far less energy ends up as warmth on the chassis.
Is A Class D Amp The Same As A “Digital” Amp?
No. “Digital” describes the input signal format, a stream of ones and zeros from a DAC. A Class D stage can accept that digital input, or it can accept a plain analog voltage from a CD player. The Class D label is about the output-stage topology, not the signal format. Many solid-state amps with digital front ends still use a linear Class A or AB output. The two ideas are independent.
What Is Dead Time And Why Does It Matter To The Sound?
Dead time is the tiny gap, often a few nanoseconds, when the upper switch has just closed and the lower switch has not yet opened, or vice versa. Both switches are off for that sliver. During the gap, no current reaches the filter, and the inductor drives whatever current it is holding on its own. In bass-heavy passages the effect is negligible. In quiet midrange detail, it can shave a hair of energy off the peaks of consonants, making “s” and “ch” sound a touch thinner. Shaping the gate timing shrinks the gap.
Can I Hear The Difference Between A Well-Built And A Poorly Built Class D Amp?
Yes, and the gap is mostly in the mids and highs. A well-designed stage has a filter large enough to tame ripple, a gate driver that minimizes dead time, and a layout that keeps switching noise from radiating into the speaker cable. A poorly built one may show a grainy bass texture, a sibilant edge on vocals, and a faint etch on cymbals. At high listening volumes the differences are smaller; at low volumes, where the signal is closer to the floor, the imperfections stand out.
Does A Class D Amp Damage Speakers More Easily Because Of The Switching Spikes?
The filter inductor and capacitor are sized specifically to keep the raw square-wave pulses from reaching the speaker voice coil. What the speaker sees is a smooth waveform, not spikes. The concern is not the switching itself but whether the filter is rated for the current and whether the output stage can handle a short-circuited or very low-impedance speaker without the switches getting overstressed. In normal use with a properly matched speaker, the switching topology does not add extra risk to the driver.
(Originally published on hifireport.com. Unauthorized scraping or republishing is prohibited.)