How to Choose the Right PWM Frequency for Your Application

Pulse-width modulation (PWM) is one of those topics that seems simple on the surface—you toggle a pin high and low, adjust the duty cycle, and suddenly you have speed control, brightness control, or position control. But once you move past blinking an LED and start driving real hardware, especially micro servo motors, the question of frequency stops being academic and starts being the difference between a smooth, quiet mechanism and a jittery, buzzing mess that overheats and dies early.

This article walks through how to reason about PWM frequency for your specific application, with a heavy focus on micro servos, because they are one of the most common—and most misunderstood—loads that hobbyists and product designers encounter. We'll cover what frequency actually does, why servos are special, how to read datasheets without getting fooled, and how to test your way to a reliable design.

Why PWM Frequency Matters More Than You Think

PWM frequency is the rate at which the signal repeats. A 50 Hz signal completes 50 cycles per second, meaning each cycle lasts 20 milliseconds. A 20 kHz signal completes 20,000 cycles per second, with each cycle lasting 50 microseconds.

That single number—the frequency—determines several things at once:

• How the load interprets the signal. Some devices care only about duty cycle; others care about the timing of edges.
• Audible noise. Frequencies in the human hearing range (roughly 20 Hz to 20 kHz) can produce whining, buzzing, or whistling from motors and inductors.
• Switching losses. Every edge in a MOSFET or driver circuit dissipates a little energy. Higher frequencies mean more edges per second and more heat.
• Resolution trade-offs. On a fixed-clock microcontroller, higher PWM frequency usually means fewer discrete duty-cycle steps.
• Filtering requirements. If you're converting PWM to an analog voltage, the frequency determines how aggressive your low-pass filter must be.

For a simple LED dimmer, you might pick 1 kHz and never think about it again. For a micro servo, the wrong frequency will cause visible twitching, audible chatter, and reduced lifespan. The stakes go up fast.

The Micro Servo Motor: A Special Case

How a Micro Servo Actually Reads Your Signal

A standard analog micro servo—think of the 9g-class units used in small robots, camera gimbals, and RC planes—expects a 50 Hz PWM signal with a pulse width between roughly 1 ms and 2 ms. A 1.5 ms pulse commands the center position; 1 ms drives it to one extreme and 2 ms to the other. Some servos accept a wider range, such as 0.5 ms to 2.5 ms, for extended travel.

Inside the servo, a small control IC compares the incoming pulse width to the feedback from a potentiometer connected to the output shaft. The difference drives a tiny DC motor through an H-bridge until the error is zero. This is a closed-loop system, and it is exquisitely sensitive to the timing of the input pulses.

Why 50 Hz Is the Traditional Standard

The 50 Hz convention comes from the early days of radio control, where multiple servo channels were multiplexed onto a single receiver output. Each channel got a slice of a 20 ms frame. The standard stuck because it worked, and because the analog control loops inside servos were tuned around it.

Here's the key insight that trips up a lot of people: the servo doesn't just measure pulse width—it also expects the frame to arrive at a predictable rate. If you send pulses at 300 Hz, the control loop sees updates six times faster than it was designed for. Some digital servos handle this fine and even benefit from it. Many analog micro servos do not. They may jitter, draw excessive current, or oscillate around the target position.

Digital vs. Analog Micro Servos

Digital micro servos are increasingly common and often advertise support for higher update rates—300 Hz or even 333 Hz. The internal controller samples the input faster and drives the motor with a higher-frequency PWM of its own, which produces tighter deadband and better holding torque.

But "supports higher rates" does not mean "requires higher rates." A digital servo rated for 333 Hz will usually still work at 50 Hz. The reverse is not true: feeding 333 Hz to an analog servo designed for 50 Hz is asking for trouble.

If you're designing a product around a micro servo, check the datasheet for the maximum refresh rate and treat it as a hard ceiling, not a target.

Reading the Datasheet Without Getting Fooled

Datasheets for hobby servos are notoriously thin. A typical micro servo listing gives you torque, speed, voltage, weight, and dimensions—and maybe a line about "operating frequency: 50 Hz." That's often the only guidance you'll get.

Here's what to look for and what to be skeptical about:

The "Operating Frequency" Line

If the datasheet says 50 Hz, assume 50 Hz is the safe default. If it says "50–330 Hz," you have room to experiment, but start at 50 Hz and work upward while monitoring behavior.

Speed and Torque Ratings

These are almost always measured at a specific voltage and, implicitly, at the standard refresh rate. If you double the refresh rate, you may see slightly better small-signal response but also more heat. Don't assume the published torque figure holds at 300 Hz.

Deadband Specifications

Deadband is the range of pulse widths the servo treats as "close enough" to the target. A tight deadband (say, 5 microseconds) is great for precision but demands a clean, stable PWM signal. If your microcontroller's PWM has significant jitter, a tight deadband servo will hunt constantly.

Matching PWM Frequency to Your Controller

The Resolution Problem

Most microcontrollers generate PWM from a timer with a fixed clock. If your timer runs at 16 MHz and you want 50 Hz, you have 320,000 clock ticks per cycle—plenty of resolution. If you want 20 kHz, you have 800 ticks per cycle, which is still workable but starts to constrain fine duty-cycle control.

For servos, this matters because the useful pulse range is narrow. A 1 ms to 2 ms window at 50 Hz represents a duty cycle of 5% to 10%. If your timer only offers 8-bit resolution (256 steps across the full period), you get about 13 steps across the servo's entire travel range. That's visibly coarse. You want at least 12-bit resolution, ideally 16-bit, for smooth servo motion.

This is one reason 50 Hz is convenient: it gives you the most resolution headroom on a fixed-clock timer.

Jitter and Clock Stability

Servo control is a timing-sensitive job. If your PWM peripheral derives from an unstable clock, or if your firmware occasionally delays timer updates, the servo will twitch. Higher frequencies amplify jitter because each pulse is shorter, so a fixed absolute timing error becomes a larger fraction of the pulse width.

Practical advice: use a hardware PWM peripheral, not a software loop, and avoid sharing the timer with other tasks that might change its configuration mid-flight.

Audible Noise and the 20 kHz Question

One of the most common reasons people want to raise PWM frequency is to escape the audible range. Motors and inductors driven at 1–8 kHz can produce an annoying whine. Pushing above 20 kHz makes the switching inaudible to most humans.

For micro servos, this logic mostly applies to the motor drive inside the servo, which you don't control. The 50 Hz command signal itself is far below the audible range and produces no sound on its own. So raising your command frequency to 20 kHz doesn't make a servo quieter—it just violates the servo's input spec and may cause malfunction.

Where the 20 kHz rule does apply is when you're driving the servo's power rail through a switching regulator, or when you're using PWM to control a separate DC motor or LED array alongside the servo. In those cases, pushing switching frequencies above 20 kHz is a standard trick for silent operation.

Testing Your Way to a Confident Design

Specs and theory only get you so far. Here's a practical test sequence for any servo-driven design:

Step 1: Start at the Spec

Run the servo at 50 Hz with a clean 1.5 ms center pulse. Confirm it holds position without visible movement. Listen for chatter. Measure current draw at rest.

Step 2: Sweep the Frequency

If the datasheet allows higher rates, increase in steps—100 Hz, 200 Hz, 300 Hz—and at each step, command small position changes and observe. Watch for oscillation, increased current, or audible noise. Stop at the first sign of degradation.

Step 3: Stress the Extremes

Command full travel repeatedly and measure temperature rise on the servo case. Higher refresh rates increase the rate at which the internal H-bridge switches, which means more heat. A servo that's fine at 50 Hz may run noticeably warmer at 300 Hz.

Step 4: Check Under Load

Micro servos behave very differently under load than free. Attach your actual mechanical load and repeat the sweep. Load-induced current spikes combined with a high refresh rate can push a small servo past its thermal limits.

Step 5: Verify Long-Term Stability

Run the system for an hour or more at your chosen frequency and log behavior. Servos that seem fine for thirty seconds sometimes develop drift or chatter over longer periods, especially as they warm up.

Common Mistakes and How to Avoid Them

• Assuming faster is better. Higher refresh rates can improve responsiveness in digital servos, but they are not universally beneficial and can damage analog units.
• Ignoring resolution loss. Raising frequency on a fixed-clock timer reduces duty-cycle resolution. For servo work, this often matters more than the frequency itself.
• Using software PWM for servos. Bit-banged PWM is fine for LEDs but introduces jitter that servos hate.
• Forgetting the power supply. A servo that twitches at a new frequency may actually be browning out because the supply can't handle the current spikes. Check the rails with a scope.
• Mixing servo and motor PWM on the same timer. If you're driving a DC motor at 20 kHz and a servo at 50 Hz, give them separate timers. Sharing one rarely ends well.

A Quick Decision Framework

When you're staring at a new design and wondering what frequency to pick, run through this checklist:

1. What does the datasheet specify? Default to it.
2. Is the servo analog or digital? Analog: stay at 50 Hz. Digital: 50 Hz is safe; higher rates are optional and should be validated.
3. Does your timer have enough resolution at your chosen frequency? Aim for 12-bit or better across the servo's pulse range.
4. Is anything else on the same timer? If so, separate them.
5. Have you tested under real load and over time? If not, you haven't finished choosing.

PWM frequency is one of those parameters that rewards a little patience up front. Pick it deliberately, test it honestly, and your micro servo will reward you with smooth motion, low noise, and a long service life—instead of the dreaded buzz-and-twitch that sends so many projects back to the bench.