How is the PWM frequency chosen in motor drivers?
PWM frequency is the balance between quietness and heat: raising it cuts the hum and the current ripple, and raises the switching loss and the heating.
Choosing the PWM frequency is a joint decision between three effects: audible hum, the ripple in the motor current and the switching loss in the driver. As the frequency rises the first two improve and the third gets worse. In applications where people stand close by, going above the hearing band is the right choice; where the heating is already marginal, staying at a lower frequency gives thermal relief.
Weigh the three effects together when deciding
- Acoustics: The switching frequency creates mechanical vibration in the motor winding. At low values such as 1–2 kHz this is an audible hum; above 20 kHz it is typically inaudible.
- Current ripple: As the period shortens, the motor inductance smooths the current better; the ripple falls, and with it the extra heating and the torque ripple.
- Switching loss: Energy is spent in the MOSFET at every on-off event, and this switching loss grows in direct proportion to the frequency. If the cooling is inadequate the board overheats.
What changes in practice from one frequency band to another
| Band | Audibility | Current ripple | Switching loss |
|---|---|---|---|
| 1–2 kHz | Pronounced hum | High | Low |
| 5–10 kHz | Shrill, still audible | Medium | Medium |
| 20–25 kHz | Typically inaudible | Low | High |
The table is not an absolute limit, it shows a direction. Two motors can behave differently in the same band, because the ripple depends on the inductance of the motor: small motors with low winding inductance want a higher PWM frequency for the same quietness, while large, high-inductance motors draw a smooth current at a low frequency too. The threshold of audibility is not a sharp line either: it shifts from person to person, with age and with the background noise. A hum that goes unheard in a workshop can be a nuisance inside a quiet caravan at night.
A common mistake: trying to solve every hum with frequency
The most frequent error in the field is to treat every sound you hear as a PWM problem. Raising the frequency stops the sound but makes the driver hotter; if you are already thermally marginal, you have stopped the sound and bought a fault. What is more, the source of the hum is often not electronic: a loose mounting foot, an unbalanced rotor, a bearing that has dried out and a sheet panel going into resonance all produce a similar sound. Before you decide, uncouple the motor from the load and see whether the same sound persists; if the sound changes with the load, the problem is mechanical.
Where this choice does not apply
Frequency is not a user-accessible parameter on every driver. The switching frequency of the AXI DC drivers (ST42, KS250, PT500) is not stated in the product data and it is not a parameter the user can set. On all three models speed and ramp are open to the user (with a pot on the KS250 and PT500); the current limit is set with one pot per motor on the ST42 and with separate A1/A2 pots per direction on the PT500. The values above are a general engineering reference and must not be attributed to the products. The rule also reverses on inductive loads that are not motors: when driving a proportional solenoid, a low frequency and dither are used deliberately, because there the aim is not quietness but breaking the friction hysteresis.
What to do
If you are using a driver whose frequency you can set, decide first on the noise tolerance of the application: go above the hearing band in cabins, greenhouses, doorways and other places where people stand close by, and accept that there is no such obligation in a machine room. Then watch the driver's temperature not at idle but on the real duty cycle and under load; no driver heats up in a single short trial. Change one step at a time and note both the sound and the temperature at every step, or you will not be able to tell which setting fixed what. If the heating increases, take the frequency back down and work through the other causes of heating in order; frequency is the last item on that list, not the first.
The full topic: DC Motor Driver Selection Guide