• State
  • Locked Locked
  • Replies 6 replies
  • Subscribers 82 subscribers
  • Views 1314 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

DRV8301-HC-C2-KIT: InstaSpin BLDC BEMF Waveform shape

tyson jensen
Prodigy 230 points

Part Number: DRV8301-HC-C2-KIT

Why does the BEMF waveform become very ugly around 7k RPM?  The motor still spins but my worry is it will stop commutation shortly because it no longer has this nice square wave which can be integrated and time the commutation from.  I don't see this if running with just an inertial disk instead of a propeller.

I've played with Vint threshold but doesn't seem to help.   

I've changed the BEMF filter cap to 0.015uF based on calculations in the documents.

I've attached a FLV file that can be viewed with VLC player of the CCS window and waveforms while ramping the RPM.

Setup:
DRV8301-HC-C2 kit Rev D
CC2803x ISO DIMM Rev 1.3 control card
CCS 6.2 and latest control suite install.

PWM: 41kHz
ISR:    41kHz

Motor:
KDE600XF-530-G3
Poles = 10, verified by counting magnets
Voltage = 22V - 52V
Current = 88A continuous, 170A peak
Kv = 530 RPM/V
L = 15uH - measured by LCR
R = 17mOhm

14 x 7 prop on static thrust stand

  • 0
    TI__Guru** 113150 points
    Did you use oscilloscope measure the phase current and voltage waveform? For high speed, the datalog waveform will be not good for there is no enough points in buffer.
    Adjust the variable "Int_Threshold" to a correct value to improve the commutation.
  • 0 in reply to Yanming Luo
    Prodigy 230 points

    I've run the motor at 12-15k RPM with just a disk on the motor instead of a propeller and the waveform looks fine at those speeds.  I don't think it's a sampling problem based on that result.

    Maybe it's noise getting into those ADC lines that run from the drive board to the processor board? Which causes sampling errors and bad waveforms.

    I've looked at the PWM signals and Commutation Step counter and it looks like it's operating correctly.

    The Phase Voltage is noisy even at low RPMs but it's tough to know because the signals are routed on the board so long they can pick up noise and leads from the scope probe.

    Does the Phase current matter? I'm only using speed control.  The code doesn't appear to make use of the summed phase current measurements unless you use the current controller.

  • 0 in reply to tyson jensen
    TI__Guru** 113150 points
    Is the same hardware board and motor? If yes, the high current maybe make noise on voltage sampling for the motor run with propeller.
    Did you check the pwm output during high speed running? Have any current waveform with current probe of oscilloscope for analysis?
    Speed close loop control is enough for most BLDC control, don't need current close loop.
  • 0 in reply to Yanming Luo
    Prodigy 230 points

    I took a scope shot of one phase current and a couple PWM signal of high side and low side of phase A.

    You can see the BEMF dip in voltage corresponding to a dip in the phase current.  I would think this current would ramp up without the dip.

    The commutation integration Vint threshold gets noisy because of these dips in BEMF voltage.  This causes the timing to change and not a reliable commutation at higher RPMs. 

    Any other thoughts with what is going on? 

  • 0 in reply to tyson jensen
    Prodigy 230 points

    The other question that comes to mind is what is causing this double spike of phase current?

    Everything i've read talks about the current waveform looking square.  The average current is about 1.5A in this picture above which matches what is read by the power supply. Why does it start to look exponential? 

    Here's another scope shot merged from many shots with 2 sample averaged data to clean it up to get the full electrical cycle of 6 steps and the phase current and BEMF.

  • 0 in reply to tyson jensen
    Prodigy 230 points
    Please can anyone provide some insight?