Class D Amplifier Myths That Do Not Survive The Physics
Class D amplifiers sound digital, run hot, can’t control speakers, and leak radio noise into your room. The claim bundles four separate fears into one label. The strongest reason to drop it: at normal listening levels, switching artifacts sit far below the noise floor of any real room and speaker pair.
Key Takeaways
- The “digital sound” fear is a mix-up. Switching topology creates high-frequency ripple, but the LC filter ahead of the speaker turns that ripple into a smooth waveform. Your ears hear the filtered output, not the raw switching.
- Thermal myths trace back to first-generation boards. Early Class D units bolted big heatsinks because gate-drive traces were long and clumsy. Modern layouts with short, wide trace paths run cooler than Class AB rivals at the same wattage into the same speaker.
- The damping factor worry reads the wrong spec. At audible frequencies, the output impedance is set by the filter network, not by the raw switching stage. The number people cite is a DC measurement that tells you little about what your woofers actually feel at 60 Hz or 40 Hz.
The Claim, In Its Own Words
You hear it in forum threads, in YouTube comment sections, and in the “gotcha” lines of product reviews: “Class D amps have that digital, square-wave sound.” The variants run like this: “It sounds robotic at the top end.” “The switching noise leaks into the tweeter.” “You can hear the PWM in the cymbals.”
A second family of claims goes: “Class D runs so hot you need a heatsink the size of a frying pan.” And a third: “The damping factor is way lower, so the bass is loose and flabby.”
Each of these is a real fear wrapped around a real-sounding number. The trial below walks through where each one breaks.
The Source Of The Story
The digital-sound claim started in the early 1990s, when switching amplifiers were new to the consumer market. First-generation boards used wide PWM pulses and slow filter networks. Those units did produce audible switching-edge artifacts at high volume into low-impedance loads. A few early reviews noted a “gritty” upper-midrange texture, and the phrase “digital square-wave sound” stuck to the topology in the minds of reviewers who heard one or two rough examples.
The thermal fear came from the same batch of boards. Because gate-drive traces were long and components were packed close, the switching transistors actually ran hot. Manufacturers added large aluminum fins to the chassis to be safe. A photo of a Class D amp with a big heatsink became a visual shorthand: “Look how hot it runs.”
The damping-factor worry is a misread spec. Amplifier data sheets list a DC output impedance of a few milliohms. Someone compares that to a Class B amp’s 50 milliohms and concludes, “Lower impedance means less control.” But that DC number is one snapshot. The real control at audible frequencies is set by the LC filter, which sits between the switching stage and the speaker terminals. The claim looks at one link in the chain and calls it the whole chain.
What The Physics Actually Says
Think of a Class D amplifier like a very fast, very efficient water-pipe system. Instead of opening a valve to a precise halfway position (what a Class AB amp does with its output transistors sitting partway between on and off), the Class D system slams the valve fully open and fully closed many thousands of times per second. The pulse-width modulation (PWM) stage turns the smooth music signal into a train of fixed-height pulses whose width changes with the music. Wider pulses mean more voltage; narrower pulses mean less.
That raw pulse train is not what reaches your speaker. Between the switching stage and the output terminals sits an LC filter (inductor plus capacitor). The inductor resists sudden current changes; the capacitor smooths out voltage spikes. Together they act like a low-pass barrier: the very fast switching ripples (often well above 200 kHz) get chopped off, and only the smooth, slow-moving audio waveform passes through.
Here is where the “digital sound” claim breaks. The PWM switching artifacts live in a frequency band far above what the filter passes. At the output, the waveform looks smooth to any measurement and to your ears. The remaining distortion products from the filter are tiny, sitting well below the threshold of hearing in a normal room. Your speakers and the air in your room add their own noise floor, which buries those last traces of switching edge.
The thermal story follows the same logic. The gate-drive circuit is the small controller that tells each transistor exactly when to open and close. In modern designs, those traces are short and wide, so the transistors switch cleanly and waste very little energy as heat. The result: at equal wattage into the same 8-ohm or 4-ohm speaker, a well-laid-out Class D board typically runs cooler than its Class AB counterpart. The big heatsinks of the 1990s are no longer needed.
For the damping factor: the damping factor is the ratio of the speaker’s DC resistance to the amplifier’s output impedance at that same DC point. A higher ratio means the amp can pull current back through the voice coil more forcefully, slowing the cone faster. But the number on the data sheet is a DC (0 Hz) measurement. At 40 Hz or 80 Hz, the LC filter in front of the output changes the effective impedance dramatically. The filter dominates. So the raw switching-stage impedance that people quote is not the number that governs bass control in the audible range.
Where It Almost Holds
The claims are not pure fantasy. They describe conditions that existed or still exist at the edges:
- Poorly filtered or under-designed units. If the LC filter is undersized or the PWM frequency is too low, some switching-edge energy can creep into the upper treble. You might notice a slight “grain” in sustained cymbals or string harmonics. This is a design flaw, not a topology flaw. A properly designed filter at a switching frequency above 200 kHz pushes that energy where no human ear can reach it.
- Tight enclosures with long, unshielded speaker cables. If the speaker cable runs alongside the switching node inside a metal chassis without a braid or twist, a small amount of electromagnetic energy can couple into the cable. The effect is measurable on a lab oscilloscope but sits well below the noise of the room you listen in. It matters for a poorly laid-out budget unit; it does not matter for any amp with a metal enclosure and reasonable routing.
- Very low-impedance, high-current loads. Pushing a subwoofer that dips to 2 ohms under transient can stress the gate-drive and force the filter to work harder. Distortion may rise. But this is a load-matching issue that any topology faces; it is not unique to Class D.
The Verdict
Stop believing: “Class D amplifiers sound digital, run dangerously hot, and cannot properly control speaker cones.”
Carry this instead: A Class D amplifier is a switching power stage plus an LC filter. The filter is what you hear. At normal listening levels in a normal room, the switching artifacts are below your noise floor. Modern gate-drive layouts keep the board cooler than a Class AB rival at the same output. The DC damping figure on the spec sheet is one data point; the LC filter, not the raw switching stage, sets the impedance your woofers actually see at audible frequencies.
How The Story Travels
The claim persists for three reasons. First, the phrase “digital sound” is sticky. It pairs a buzzword (digital) with a physical sensation (square waves) and gives the listener something concrete to fear, even though the two words describe different layers of the circuit. Second, the thermal image is visual. A big heatsink is easy to photograph and easy to point at in a forum thread. Once the picture is in someone’s mind, the explanation that modern boards need no such fin is harder to slot in. Third, the original 1990s-era complaints were real for the units of that time. A few genuine artifacts were measured, a few reviews used the word “gritty,” and the label attached to the whole topology. Decades of repetition make the claim sound like a settled fact, even though the specific conditions that produced it no longer describe the units on the shelf today.
Related Questions
Does The Switching Frequency You Can Hear Matter If The Filter Is Above It?
The filter is set to pass audio and block everything above roughly 200 kHz. The PWM carrier rides in that blocked band. If the carrier were somehow low enough to reach the filter’s pass band, you would hear a hum or buzz at that frequency. In practice, manufacturers set the switching frequency well above the audible range precisely so the filter can do its job. What you hear is the smoothed output, not the carrier.
Is It True That Class D Amps Are Only Efficient At Low Volume?
This inverts the real behavior. A Class D stage wastes little energy because the transistors are either fully on or fully off. At low signal levels, the PWM pulses are narrow, but the conduction losses stay small. Efficiency actually drops slightly at the very highest power levels where the inductor and capacitor take more loss. For most of the volume range you listen at, a Class D amp draws noticeably less power from the wall than a Class AB amp producing the same output.
Will A Class D Amp Damage Tweeters Because Of High-Frequency Switching Noise?
The LC filter blocks the switching band before it reaches the speaker terminals. The tweeter only sees whatever the filter passes, which is the audio signal. You would need a failing filter (a shorted inductor or open capacitor) for switching energy to reach the driver. That is a component failure, not a topology characteristic. A healthy filter keeps the tweeter isolated from the switching node.
Do You Need To Match The Amp’s Damping Factor To The Speaker’s QTS Value?
The QTS value on a speaker data sheet (often around 0.4 to 0.7 for a woofer) is a mechanical and electrical property of the driver itself. The amplifier’s damping factor plays a small role in how the cone behaves after the signal stops. But the speaker’s own suspension and motor structure set most of the movement. No Class D amp has a damping factor so low that it would make a properly designed driver “flabby.” The claim confuses a minor correction with the main cause of bass control.
(Copyright hifireport.com. Republishing, scraping, or reselling this text is prohibited.)