What is a recommended PWM frequency to send to the DRV8801? What is the max it can take? I realize that switching losses go up as the PWM frequency increases, but ideally I’d like to run them at a high enough PWM frequency such that there is no audible “PWM whine”.
I will be running motors at up to about 1 amp each.
The electrical characteristics table in the datasheet for the DRV8801 IC seems to imply that 50kHz is the upper limit. We typically recommend using a PWM frequency of about 20kHz, which should be inaudible to most people.
Thanks. I’m using an Arduino, so my options aren’t unlimited. Arduino supports (rounded) - 31kHz, 4kHz, 1kHz, or 500 Hz (the default). I’ll try the 31K and see how that works.
Ok, I did some experiments with this. I found that the voltage output, vs the duty cycle, was pretty non-linear at higher PWM frequencies. So I might stick with a lower frequency.
But I also noticed something I don’t understand, that I’m hoping you can explain. Even though the voltage was roughly the same, at lower PWM frequencies, the motor RPMs were MUCH faster. I’m running 24 volt motors, and at an input duty cycle value of 101 (out of the possible 255), I was sending roughly 9v to the motors. At a PWM frequency of 31kHz, the motors didn’t move. At 4kHz, they moved, but slowly. And at 500Hz, they spun very fast. And the voltage measured at the 500Hz frequency was actually lower than that of the 31kHz frequency (8.92v for the 500Hz, vs 9.17 for the 31kHz).
All frequencies were verified on an oscilloscope, and all voltages verified on a voltmeter.
So I know you don’t know my full setup, but if you could shed some light on what type of phenomenon is going on here, I’d appreciate it! There is something going on that I can’t explain based on what my understanding is of how powering a DC motor via PWM works - clearly I don’t fully understand it!
The response of a motor to PWM depends a great deal on the inductance of the windings. It appears that your motors have high inductance, which means that the winding current rises slowly with time. At high switching frequency, the average current is very low. See this page for a more complete explanation and recommendations precisionmicrodrives.com/app … on-control
Thanks! That explains it.