Strain on motor with higher voltage and PWM


I recently purchased two pololu micrometal HP gearmotors,

I understand that these motors are recommended to run at 6V but can run at 9V though shortening their life through increased strain.

If they are run with PWM, is the strain / wear-and-tear on the motor proportional to the duty cycle?

eg. is the strain at 9V with 66% duty cycle about the same as at 6V with 100%?

Would the answer be the same at a low PWM frequency (say 100Hz) and high frequency (say 20kHz)?

As background, in my planned project the motors will most of the time run at 9V, 66% duty cycle, but will occasionally need a short burst (e.g. less than a second) of extra power, at 9V 100%, and I’m hoping that means that most of the time they are not being pushed to the limit.



If you limit the duty cycle, it is effectively the same as limiting the supply voltage and would limit the strain. As you suggested, a 9V supply limited to 66% duty cycle is effectively the same as operating with a 6V power supply limited to 100% duty cycle. How low you can set the PWM frequencies and still use duty cycle to emulate a lower applied voltage to the motor will depend on your specific motor. 100Hz is a pretty slow PWM frequency for motor drivers, but it might still be fine. You could test it with your motor by monitoring the output voltage with an oscilloscope or multimeter.

- Grant

Thanks for confirming that Grant.

Grant hints at this point:

The lower the duty cycle you are using to emulate things (e.g, getting “6V” from 60V at 10% cycle) and the lower the frequency you are using (e.g, 100hz vs 500hz vs 20kHz+), the less you are accurately emulating a lower voltage.In turn the more you are going to have a ripple in your output voltage that will increase the RMS current going through the windings and potentially overheat the brushes or motor.

This is because to my understanding the motor is acting as an inductor that smooths out the current (inductors don’t want a change in current), and the higher frequency the more blocking it is, and the small duty cycles are asking for a greater amount of smoothing.

Thanks Tomek.
I understand how important your point is now - I wasn’t getting good results with a low PWM frequency (490Hz), but from 10kHz onwards it worked well. For anyone that’s interested, this was explored on this thread.