What Is a Class-D Amplifier's Carrier Frequency? (Switching Wave Explained)

A Class-D amplifier does not amplify sound the way older amplifiers do. Instead, it turns music into a rapid-fire chain of electrical pulses, then filters those pulses back into sound. The carrier frequency is the speed at which those pulses are created — and that speed shapes everything from heat output to sound quality.

The Light Dimmer That Plays Music

Before we touch any technical words, picture a light switch. Not a dimmer knob — a plain, old-fashioned light switch that is either fully on or fully off. Now imagine flicking that switch thousands of times every second. If you flick it very fast and keep the light on for longer stretches, the bulb looks brighter. If the on-stretches are shorter, the bulb looks dimmer. Your eyes smooth out the flickering and see a steady brightness.

A Class-D amplifier works on exactly the same principle. Its output transistors are not “halfway on” like the transistors in a traditional amplifier. They are either completely on (electricity flows with almost no resistance) or completely off (no electricity flows at all). This is why Class-D amplifiers run cool and waste so little energy — the transistors are not fighting the current. They just switch.

The speed of those switch-flips is called the **carrier frequency** (also called the **switching frequency**). It is the steady rhythm — the metronome — that the amplifier uses to chop the music signal into pulses. When you change the music from soft violin to loud drum hit, you are not changing the speed of the flicking. You are changing the width of each pulse: wider pulses for loud moments, narrower pulses for quiet moments. Engineers call this **pulse-width modulation**, or **PWM** for short.

How the Carrier Frequency Is Made

To understand where the carrier frequency comes from, think about a race track where one runner keeps a steady pace. That runner is your reference — consistently the same speed, not speeding up or slowing down. A second runner tries to keep up with the music, sometimes running fast, sometimes slow. At every instant, you compare the two runners’ positions. If the music runner is ahead, the output says “on.” If the music runner is behind, the output says “off.”

Inside the amplifier, this comparison is done by a tiny circuit called a **comparator**. One input of the comparator receives a seamless steady **triangle wave** (our steady runner) that goes up and down at the carrier frequency. The other input receives the music signal (our music runner). Every time the music signal is higher than the triangle wave, the comparator fires a pulse. Every time it is lower, the comparator goes silent.

The triangle wave runs somewhere between **250 kilohertz and 1.2 megahertz** — that is 250,000 to 1,200,000 cycles every single second. Written another way, the carrier produces a new pulse every 4 microseconds (at 250 kHz) down to every 0.8 microseconds (at 1.2 MHz). The music signal, by comparison, tops out at 20,000 cycles per second — painfully slow next to the carrier.

This huge speed gap is not an accident. The carrier frequency must outrun the fastest music note by a factor of at least ten to one. Otherwise, the pulses would not carry enough detail to reconstruct the music cleanly on the other side. Engineers call this the **Nyquist principle** — the sampling rate must be at least double the highest frequency you want to capture. But in practice, audio designers aim much higher than double. Ten to fifteen times higher is the working minimum. That is why 250 kHz is the starting point for full-bandwidth audio, and why better designs reach 400 kHz, 600 kHz, or beyond 1 MHz.

Cleaning Up: The Output Filter’s Job

After the comparator and switching stage have turned music into a high-voltage chain of pulses, you still cannot send that directly to a speaker. Speaker cones are physical objects — they cannot jump back and forth at 400,000 times a second. If you tried, the speaker would heat up, the voice coil could burn, and you would hear nothing except maybe a faint high-pitched whine.

The output of a Class-D amplifier passes through a **low-pass filter** — usually an **LC filter** made of an inductor (the L) and a capacitor (the C). A low-pass filter is like a bouncer at a club door: it lets low-frequency guests (your music, 20 Hz to 20 kHz) walk right through, but it stops high-frequency guests (the 400 kHz carrier and its harmonics) at the door.

The filter’s cutoff point — the frequency where the bouncer starts refusing entry — is set somewhere between the top of the audio band (20 kHz) and the carrier frequency. For a 400 kHz switching amplifier, the filter cutoff might sit around 50 to 80 kHz. That gives the filter plenty of room to pass all your music while starting to squash the carrier signal well before it reaches the speaker.

Designing this filter is one of the trickiest parts of building a Class-D amplifier. If the cutoff is too low, you lose sparkle and air in the treble. If the cutoff is too high, too much carrier energy leaks to the speaker. If the inductor is too small, the filter rings and adds distortion. If the inductor is too large, it gets hot and wastes the efficiency advantage you built the Class-D amp for in the first place.

Worked Example: What 400 kHz Actually Means

Let’s walk through a practical calculation so you can see how the carrier frequency shapes real amplifier design.

Suppose an engineer picks a carrier frequency of 400 kHz. Here is what that means step by step:

**Step 1: Switching Period**
The period is 1 divided by the frequency:
> **Calculation:** Period = 1 / 400,000 = 0.0000025 seconds (2.5 microseconds)

Every 2.5 microseconds, the amplifier makes one complete switch cycle — on then off, or off then on.

**Step 2: Time per Audio Cycle**
A 20 kHz treble note (the highest frequency humans can hear) has a period of:
> **Calculation:** Period = 1 / 20,000 = 0.00005 seconds (50 microseconds)

**Step 3: Pulses per Audio Cycle**
Divide the audio period by the switching period:
> **Calculation:** 50 microseconds / 2.5 microseconds = 20 pulses per audio cycle

That means for the highest note in your music, the amplifier carves the waveform out of 20 individual pulses. For lower notes, the number climbs dramatically. A 1 kHz midrange note — roughly the pitch of a female singer’s high belt — gets 400 pulses per cycle. A 50 Hz bass guitar note gets 8,000 pulses per cycle.

Now consider what happens if you drop to 200 kHz:
– 20 kHz treble: only 10 pulses per cycle
– 1 kHz midrange: 200 pulses per cycle
– 50 Hz bass: 4,000 pulses per cycle

With only 10 pulses to describe the finest treble detail, the amplifier has less information to work with. The output filter also must be set to a lower cutoff, which means it starts rolling off treble energy earlier. The sound does not necessarily become “bad,” but it may lose some openness and precision in the top octave compared to a higher switching design.

Now consider what happens if you jump to 1 MHz:
– 20 kHz treble: 50 pulses per cycle
– 1 kHz midrange: 1,000 pulses per cycle
– 50 Hz bass: 20,000 pulses per cycle

That sounds fantastic on paper. But every switching event wastes a tiny amount of energy as heat — charging and discharging the MOSFET’s internal capacitance. Double the switching frequency, and you roughly double that switching loss. At some point, around 800 kHz to 1.5 MHz for today’s silicon transistors, the heat penalty starts eating into the efficiency advantage that Class-D was designed to deliver.

A Tale of Two Listening Rooms

Let’s put these numbers into real listening situations so you can hear what they mean.

**Scene One: Jazz Trio in a Small Room**

You are in a 12-by-14-foot room with a hardwood floor and one bookshelf against the back wall. A recorded jazz trio plays through a pair of stand-mounted two-way speakers driven by a Class-D amplifier running a 400 kHz carrier. The double bass plucks a walking line around 60 to 80 Hz. The ride cymbal shimmers with energy up near 8 to 12 kHz.

Because the carrier is 400 kHz (20 times the highest audio frequency), the switching noise lives far above what any tweeter can reproduce and far above what your ears can detect. The LC output filter has its cutoff set at 60 kHz — well above 20 kHz — so the cymbal’s sparkle passes through without being dulled. The bass notes come through with tight punch because the amplifier wastes very little power as heat even during sustained low-frequency passages.

Now swap to a budget amplifier switching at only 200 kHz. The filter cutoff gets pushed down to 35 kHz to keep the carrier energy in check. The treble loses a tiny bit of air — not dramatically different, but the cymbals feel slightly less open, less ethereal. A careful listener might notice, especially on a recording they know well.

**Scene Two: Rock Band in a Garage**

You are listening to a live rock recording in a converted garage with concrete floors and almost no sound treatment. The kick drum punches at 55 Hz, the snare cracks around 200 Hz with lots of harmonic energy up to 6 kHz, and distorted electric guitars fill the 200 Hz to 4 kHz range with harmonic density.

In this environment, the carrier frequency matters less than you might think. The room’s slap echo and the sheer loudness swamp the subtle treble differences between amplifiers. The most important thing the Class-D amplifier does here is stay cool while delivering sustained high power. A 600 kHz design and a 300 kHz design would be very hard to tell apart in this untreated space.

The lesson: the carrier frequency’s audible impact shows up most clearly in quiet rooms with good speakers and clean recordings. In rough acoustic conditions or at party volumes, it fades behind far bigger factors like room acoustics and speaker placement.

The Trade-offs: Why Faster Is Not Always Better

Every engineering choice in audio is a compromise, and carrier frequency is no exception. Here is what designers juggle when they pick a switching speed:

**Dead time** is one of those hidden gremlins that makes high switching frequencies harder than they look. When a MOSFET turns off, it does not happen instantly. There is a tiny gap — a few billionths of a second — where neither transistor in a pair is conducting. Engineers intentionally insert this gap to prevent both transistors from being on at once, which would short-circuit the power supply (a destructive event they call **shoot-through**).

But dead time creates distortion. While both transistors are off, the output voltage floats — it is not being actively driven. When the carrier frequency goes up, the switching events happen more often, so dead time occupies a larger share of each cycle. At 1 MHz, getting dead time right demands extremely fast transistors and very precise gate-driver circuits. One common workaround: using **gallium nitride** (GaN) transistors, which switch faster than silicon and produce less dead-time distortion. But GaN is harder to design with and more sensitive to circuit board layout.

When the Carrier Frequency Stops Being the Bottleneck

Carrier frequency gets a lot of attention in marketing materials, but in many real-world setups, it is not the limiting factor for sound quality. Here are the cases where chasing a higher switching frequency gives you no audible benefit:

**Your speakers roll off above 16 kHz.** Most home loudspeakers, especially affordable bookshelf models, naturally lose output above 15 to 18 kHz. If your tweeter cannot reproduce 20 kHz, then an amplifier whose filter starts rolling off at 18 kHz (common for 250 kHz switching designs) sounds the same as one flat to 20 kHz.

**Your music has little treble energy.** If you listen mostly to classic rock, folk, or spoken-word content, the highest octave (10 to 20 kHz) carries very little musical information. The carrier frequency’s effect on that octave becomes irrelevant.

**You listen at conversation-level volumes.** At low listening levels, the amplifier’s inherent noise floor — the faint hiss you hear with your ear against the tweeter — matters more than treble extension. A well-designed 300 kHz amplifier with a quiet noise floor can sound cleaner than a noisy 1 MHz design.

**Your room has no acoustic treatment.** Bare walls and hard floors create reflections that smear treble detail far more than any filter-rolloff difference between switching frequencies. Put absorption panels at first-reflection points before worrying about amplifier topology.

Try This at Home: Can You Hear the Carrier?

You can run a simple experiment with almost any Class-D amplifier and your own ears. You will need a music track with strong, delicate treble — a solo acoustic guitar recording with finger squeaks and string noise, or a close-miked cymbal passage — and a pair of speakers or headphones you know well.

**What to do:**
1. Play the track at a comfortable listening volume.
2. Sit in your normal listening position. Close your eyes. Focus on the very highest sounds — the shimmer, the air, the sense of space around the instruments.
3. Now play the same track but reduce the volume until the music is barely audible — just above a whisper.
4. At this very low level, put your ear close to the tweeter (carefully — do not touch it). Listen for any faint, steady high-pitched whine or buzzing that does not change with the music. This is residual carrier noise bleeding through the output filter.
5. Repeat with a different Class-D amplifier if you have one. Compare the residual noise between them.

**What you are hearing:** At normal listening volumes, the music completely masks any carrier residue. The test only works at very low volumes because the carrier noise stays at a fixed level regardless of volume — it is consistently there, humming away at the switching frequency. A well-designed amplifier with a properly tuned output filter will produce almost no audible residue even with your ear right on the tweeter.

**Important safety note:** Do not remove the amplifier cover or probe any internal connections. The power supply capacitors inside can hold a dangerous charge even after the amplifier is unplugged.

Common Questions

Does a higher carrier frequency mean better sound quality?

Not by itself. A higher carrier frequency makes output-filter design easier and can preserve more treble extension, but it also increases switching losses and electromagnetic interference. The full design — power supply quality, feedback-loop design, circuit board layout, and filter tuning — matters more than the raw carrier number. A well-built 300 kHz amplifier can sound cleaner and more musical than a poorly executed 1 MHz design.

Why do some Class-D amplifiers make a faint high-pitched whine?

That whine is likely residual carrier-frequency energy leaking past the output filter or radiating from the power supply. At normal listening distances and volumes, your music should mask it completely. If the whine is loud enough to hear during quiet passages, the output filter may be poorly tuned, the circuit board layout may be picking up switching noise, or the power supply may be oscillating.

Can dogs hear the carrier frequency from a Class-D amplifier?

Possibly, but unlikely. The carrier frequency runs at 250,000 Hz or higher — far beyond any animal’s hearing range. What a dog might detect is not the carrier itself but lower-frequency artifacts: a poorly filtered 400 kHz switching stage can produce harmonic ringing at 40 kHz or 80 kHz, which falls within canine hearing (up to about 45 kHz). A well-designed output filter eliminates these artifacts before they reach the speaker terminals.

Does switching frequency affect bass performance?

Carrier frequency has almost no direct impact on bass reproduction. Bass notes live between 20 Hz and 200 Hz — even a 250 kHz carrier provides over a thousand pulses per bass cycle, which is far more than enough. Where switching frequency indirectly helps bass is through overall efficiency: a cooler-running amplifier sustains high current into low-impedance bass drivers without sagging, which keeps bass tight and controlled during demanding passages.

Will a Class-D amplifier with a 1 MHz carrier run hotter than one at 400 kHz?

Generally, yes. Every switching event wastes a small amount of energy charging and discharging the transistor’s internal capacitance and dealing with the brief moment when the transistor is neither fully on nor fully off. A 1 MHz amplifier switches two and a half times more often than a 400 kHz design, so switching losses are roughly two and a half times higher. The difference may only be a few watts, but in a compact chassis, those extra watts can raise the internal temperature noticeably.

Is the carrier frequency related to the “digital” in Class-D?

No. Despite the “D” in the name, a basic Class-D amplifier is entirely analog — the D simply follows A, B, and C in the historical naming order. The carrier is an analog triangle wave, not a digital clock. Some modern Class-D designs do include a digital signal processing chip before the modulation stage, but the core PWM conversion remains an analog process. The “digital amplifier” label that sometimes appears in marketing is misleading.