After a lot of research and simulation tries, I cannot determine how a brushed DC motor would act if given a high PWM frequency, ignoring any losses in the controller due to switching.

What I know so far: When a brushed DC motor recieves low frequency PWM, there will be a current ripple and a reduction in effeciency. Increasing the frequency reduces ripple amplitude, and it is generally adviseable to go above 20 kHz if driving with high frequency to avoid audible noise. A basic model of a DC motor is an inductor, a resistor, and back EMF.

However, I cannot find information for what happens at higher frequncies. I believe that the motor also has a capacitance (other than any capacitor that may be added to reduce EMI). So, at what frequency range can a typical motor be driven efficiently and without reduced speed and torque?

If even in the MHz region the capacitance is so small that it is negligible , then I would expect the DC component to not be attenuated, while the AC component of the PWM signal would become negligible. This would possibly allow PWM drive without significant current ripple, needing only small ceramic capacitors and thus a smaller driver board, ignoring switching losses.

I was not able to simulate this, and I dont have the equipment to test it. I would appreciate any help. Thank you.

  • \$\begingroup\$ (1) I don't know how to simulate, so I just test the real thing: 5~12V DC motor. My experience is that the workable PWM frequency range usually around 500Hz to 5kHz, A good starting point is 1kHz, (2) Reference: electronics.stackexchange.com/questions/564733/… \$\endgroup\$
    – tlfong01
    Sep 19, 2022 at 6:04
  • \$\begingroup\$ Torque ripple should be negligible above a few kHz,and the motor inductance would eliminate the need for a smoothing capacitor except for EMI issues. A very low PWM frequency might be useful to provide a sort of hammer drill effect. \$\endgroup\$
    – PStechPaul
    Sep 19, 2022 at 7:15
  • 1
    \$\begingroup\$ Slight increase in core eddy current losses, more EMI. \$\endgroup\$
    – winny
    Sep 19, 2022 at 8:00
  • \$\begingroup\$ @PStechPaul According to this, voltage ripple is still an issue that is solved via an input capacitor: modularcircuits.com/blog/articles/h-bridge-secrets/… \$\endgroup\$
    – Anas Malas
    Sep 19, 2022 at 12:33
  • 1
    \$\begingroup\$ Back in college I discovered that in tiny servo motors for RC cars, when we took out the electronics and drove the motors directly, if the PWM frequency was too high, the motors generated less torque at the same duty cycles compared to lower frequencies. I don't recall the exact frequencies we used though. \$\endgroup\$
    – Aaron
    Sep 19, 2022 at 15:08

1 Answer 1


I believe that the motor also has a capacitance (other than any capacitor that may be added to reduce EMI).

Internal capacitance is generally very low, resulting in a resonant frequency in the MHz range. However brushed DC motors typically have a 0.05 to 0.1 μF capacitor wired across them to suppress EMI, and the output capacitance of the switching transistor(s) can be several nF.

So, at what frequency range can a typical motor be driven efficiently and without reduced speed and torque?

It mostly depends on the type of motor, but there can be wide variations between different models even within the same type. The majority of brushed DC motors have a laminated iron armature core which increases inductance but has eddy current loss at high frequencies. This works well at frequencies of a few kHz. Coreless motors have much lower inductance so they need higher PWM frequency (>30 kHz) to smooth out the motor current.

Although high frequency PWM (defined as high enough to cause continuous current flow) is more efficient, low frequency PWM often produces a more linear throttle response at low speed by reducing 'stiction' (it basically vibrates the motor to make it start and run more reliably at lower voltage). This is very dependent on the particular motor and mechanism it is attached to, as well as the torque load.

Another factor is the type of PWM applied. 'Slow decay' (also called 'synchronous rectification' or 'active freewheeling') shorts out the motor when PWM is off to keep the current going, while 'fast decay' just uses a diode to recirculate back-emf current. Fast decay causes ringing at low PWM ratios when the current is discontinuous, and can make the response sensitive to capacitance.

Here is some test data showing the quite complex interaction between PWM frequency and speed response of a cheap 130 size brushed DC gear-motor:-

enter image description here

And here are the results for a typical N20 size gear-motor:-

enter image description here

In both cases the throttle response with 'fast decay' (diode freewheeling) is terrible. At 27 kHz the motor doesn't start until the PWM ratio reaches 65-70%.

  • \$\begingroup\$ Excellent resource! Somehow I had missed this specific page. However, I wonder what the hysterisis on this is. What if the driver was configured to start at 100% duty cycle which is then reduced to the desired duty after the inertia is overcome and the motor spins. Would the motor then keep spinning even below the duty cycle that would have started it? \$\endgroup\$
    – Anas Malas
    Sep 19, 2022 at 12:16
  • 1
    \$\begingroup\$ Probably. But if load increases (or voltage drops) even momentarily the motor may stall and not restart. \$\endgroup\$ Sep 19, 2022 at 19:37

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.