DRV8323R: Unexplained current perturbation in 3 PWM mode

Hi Victorien,

1. Could there be another load that is leading to the extra current draw, like a resistor divider for example?

2. Could you tell me how and where, like with a block diagram, you are measuring the currents? 

Thank you,

Joshua

Hello Joshua,

1. The motor is currently running with no load.
2. I am measuring the current using an hardware trigerring on low-side sensing resistors. The measurement of the 3 channels is triggered at the center of the logic low states of the high side PWMs. At that particular moment, the complementary low side PWMs, are logic high state and centered. I already know that the 3 channel ADC sampling matches with the PWM frequency time constraint.
   
   In the figure bellow, I would trigger in the center of PWMxL. 
    

Hi Victorien,

Would you mind providing your schematic?

Thank you,

Joshua

Hi Joshua,
Unfortunately, I cannot provide you the schematics of my board but you can consider that it follows the typical implementation describes in the DRV8323Rx datasheet (p62). Note that my implementation has been largly tested using the 1xPWM mode of the DRV8323 and the motor was able to run smoothly.

It has been suggested to me that the cause of this behaviour is the fact that I am running the motor in open-loop. I was personally expecting the motor to be able to run, even in open-loop mode, up to few hundreds RPM at least. Here it is not the case.

Hi Victorien,

What type of probe are you using to measure the current across the low side sensing resistor?

Thank you,

Joshua

Hello Joshua,
The MCU uses internal ADCs and I use also external logic analyzer or external oscilloscope. All solutions measure the amplified output of the DRV8323 and not directly the voltage across the sensing resistors which is very small.

Hi Victorien,

Could you provide me a capture of the phase currents by using a current probe at the motor?

Thanks,

Joshua

Hello Joshua,
Sorry for the late response. Unfortunately, I don't have the tools to measure directly motor phase current except my low-side ADCs. What extra information would you expect from this measurement ?
Thank you,
Victorien.

Hi Victorien,

We wanted a capture of the phase current to see if the problem is present in the raw phase currents as well. This information can help to eliminate potential causes of the non-zero current values being seen, such as the DRV8323R's CSAs.

Since you lack the equipment to measure the phase current, could you measure the voltage across the shunt resistor and the shunt resistor value?

Thank you,

Joshua

Hello Joshua,

Shunt resistor value is 2mOhm. The driver measurement is centered for bidirectional currents. Voltages are measured on SOx pins.
Here is a set of measurement that I just did with the motor running in open-loop with PWM following phased sine waves. 
Find also attached a video for better view of the measurements.

EDIT : The observable plateaus/peaks around the signal's envelope seem to match with the duty cycle refresh of the PWMs. In my understanding, it cannot cause the current measurement you see bellow.

Zoom-in to see impact of gate ON/OFF states :

Those raw measurements give me the corresponding MCU's internal data. Time is ms, current is Ampere.

Which gives the following by zooming.

The internal currents' shape is so bad that it is not useable.

Let me know if you need further information.
Thank you in advance,
Victorien.

Hi Victorien,

There does not seem to be any problem with the currents based on the waveforms that you sent me. 

Would you mind double checking that when there is a non-zero sum that the ADCs were sampling at the correct time? If they did not sample when the low side gate was active this could lead to the non-zero sum.

Best regards,

Joshua