DRV8306: The ABC three-phase output waveform of the drive circuit designed by DRV8306 is wrong, the motor speed is slow, and the duty cycle cannot be raised to more than 50%

The following two signals are the control signal PWM and the A-phase voltage waveform. As shown in the figure, the yellow signal refers to the A-phase voltage waveform, which is A trapezoidal wave. The other green signal represents the PWM waveform.

I found that even if I gave the PWM_PIN of DRV8306 a 100% duty cycle PWM signal, the motor speed was still slow, about 60% of the normal state. Therefore, I collected the three-phase voltage waveform and PWM waveform of the drive circuit. On the whole, the three-phase waveform of the driving circuit presents a trapezoidal wave, which seems to be correct. But when I enlarged the waveform, I noticed something was wrong.

The following diagram has a PWM rating of 40KHZ and a duty cycle of about 30%, It can be seen that every two pulses can be synchronized with the PWM pulse. But in between these two pulses there is always a pulse that drops instantaneously after passing the rising edge. This led me to expect a duty cycle of 30% to control the motor speed, but in reality it was only around 15%.

The same situation is shown in the figure below. The PWM duty cycle in the figure below is about 80%. 

This is even more true when the PWM is 100% duty cycle. At this time, the A-phase modulation waveform of the drive circuit has completely changed into a 50% duty cycle square wave.

I think this is also the reason why my motor speed is slow. May I ask the experts of TI, what causes this phenomenon, and how to solve this problem.

Here is my schematic diagram of the circuit

4212.DRV8306原理图.pdf

Cherry Zhou：

Hi,

Thank you for the query and please allow some time for our experts to get back to you.

Thanks

,

happy：

Thank you for your reply,I added some experiments again. and I have some more information to offer you.

I feel like the situation is a little bit like the pulse-by-pulse Current Limit.

The following figure shows the comparison of GHX-GND and PWM waveform, This is similar to the ABC three-phase voltage waveform.

Below is a comparison of GLX – GND and PWM waveforms,This image shows PWM signals in yellow and GLX signals in green.

the combination of the above two diagrams fits well with:"If the current limit activates, the high-side FET is disabled until the beginning of the next PWM cycle. Because the synchronous rectification is always enabled, when the current limit activates, the low-side FET is activated while the high-side FET is disabled."

But the sampling resistance I used was very small, only 5 milliohm, And the motor running current at full speed is only 0.5A. 

It should not cause a Current Limit.

,

happy：

Have you found anything wrong？This problem has caused me a lot of trouble.

,

Cherry Zhou：

Hi,

Sorry for the late action and the expert will provide you a response by Thursday at the latest.

Thanks

,

happy：

Thank you very much for your help during this period. I have solved this problem, which is the configuration problem of nBRAKE in my own program