• State
  • Locked Locked
  • Replies 4 replies
  • Subscribers 83 subscribers
  • Views 1327 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

LAUNCHXL-F28069M: Noise at high RPM

user3473222
Intellectual 300 points

Part Number: LAUNCHXL-F28069M
Other Parts Discussed in Thread: MOTORWARE, DRV8323

Hello,

I am trying to run a BLDC motor at speeds close to maximum (it is required for application).. Identified motor parameters are: Rs - 0,0504 Ohm, Ls - 36,6 uH, Flux 0,02018 VpHz, No of pole pairs - 14. The driver I use is DRV8323RS booster pack.

When motor reaches high speeds, it stars to make a clicking noise. It sounds like a sand in the bearings (but actualy motor is new, and it is the same on several different motors). The noise is also seen on the measured current ( I am putting measured current to DAC so it can be monitored on osciloscope) - looks like a disturbance of sinewave (it is seen on attached video).

After multiple experiments I noticed that this noise and maximum RPMs depend on PWM frequency:

When PWM frequency   = 10 kHz and PWM_ticks_per_PWM_tick = 1, Motor reaches 5179 RPM and almost no noise can be heard or sean on current waveform.

When PWM frequency   = 20 kHz and PWM_ticks_per_PWM_tick = 2, Motor reches only 4808 RPM and strong noise can be heard and monitored (this case is seen on video).

When PWM frequency   = 20 kHz and PWM_ticks_per_PWM_tick = 1, Motor can reach 5069 RPM but the noise is simmilar like in previous case.

It were the only changes I made. For this experiment I was using Motorware lab 3b so all motor parameters were allredy determined and identical in every run.

What could be the reason of this noise, and what should I try to do in order to get rid of this disturbances?

Waveforms seen on osciloscope: CH1 (yellow) -estimated angle on DAC output; CH1 (blue) - measured current (ADC) of one of three phases and put to DAC output; CH3 (purple) - PWM waveform on phase low-side gate; CH4 (green) - duration of main ISR (at the beginning of ISR GPIO is set high and at the end of ISR, before return statement, it is set low. The piece of carpet on the video is used to apply some frictional load to motor - because at no-load current is weak and can not be waveform can not be seen clearly.

Regards,

Vladislav

  • 0
    TI__Guru** 116930 points
    It seems the maximum electric frequency of motor is very high, you need to use a higher pwm and controller frequency for this motor. And it seems you was running the motor without any load, and the current is very small which is difficult to estimate the angle correctly during high speed.
  • 0 in reply to Yanming Luo
    Intellectual 300 points
    Thanks for the answer!
    I had the same impression. However the higher PWM frequency I set, the stronger noise I hear. Yes I admit, that when motor isn't loaded current signal is really weak, but when I apply some load (for example add some friction), I still observe disturbances in measured current.
    I made the assumption, that at higher PWM frequency there is less time to do all I and V measurements. Could the PWM-ADC timing be the issue? If I understand correctly, ADC conversion is triggered t the middle of Low PWM state. But how about propagation delays and rise times? Should I try any time offset for the ADC COS trigger?
    I have tried lab 10,where current reconstruction is implemented but the result is very similar...
  • 0 in reply to user3473222
    TI__Guru** 116930 points
    Did you try to tune the speed/current PI regulator with different parameters for high and low speed? You don't need to use OVM if the input voltage is enough for motor running with high speed. What's the dead time of the inverter?
  • 0 in reply to Yanming Luo
    Intellectual 300 points
    Yeas I tried various values for Ki and Kp for both speed and currend PI controllers. Sometimes I observed oscillations in speed (and sound) of the motor, but it looked completely different from the noise I heard.
    The dead time of the DRV8323 is set to be 100 ns.