DRV8353: DRV8353

Ahmed Iftikhar

Prodigy 20 points

Part Number: DRV8353

Hello.

I am testing DRV8353H in 1xPWM mode.

VM = 15V

VDRAIIN = 24V

For testing purpose, I am using function generator to generate the PWM at 10KHz.

The bulk capacitance at MOSFET supply is around 300uF, motor is taking <1amp current at no load during free run.

I power-up DRV circuit first and then turn-ON the PWM - everything is fine and motor starts to spin.

DRV circuit power down first and then PWM turned-off - no problem

But if motor is running and PWM is turned off before DRV circuit, after 2 to 3 times, following problem occurs.

VGLS goes down to ~4 volts.

DRV current increases to ~120mA.

IC starts to get hot.

I have to replace the IC to get the circuit working.

Why is this happening?

Is BEMF destroying the IC? It has the path to go to the MOSFET supply.

Regards.

over 2 years ago

0 Eli Sorenti-Burns over 2 years ago

TI__Intellectual 1350 points

Hi Ahmed,

Thank you for posting on E2E. Your issue may be the same as the one Peter faced in the thread (+) DRV8353: DRV8353 - Damaged after freerun when stop PWM - Motor drivers forum - Motor drivers - TI E2E support forums. Have you tried the solution listed there? Please let me know if trying this method solves your issue.

Best,

Eli

0 Ahmed Iftikhar over 2 years ago in reply to Eli Sorenti-Burns

Prodigy 20 points

Hi Eli.

I have tried to stop the motor gradually and the problem doesn't occur. How gradual this should be?

My ultimate use of the circuit is for position control of a shaft (having feedback sensor) coupled with motor.

In the above mentioned thread it is mentioned that "The problem is that the rotating energy has no place to go and therefore the IC gets damaged".

Doesn't the stored energy go to the MOSFET supply only (through MOSFET diodes), and if the voltage remains within the VDRAIN limit, the IC should not be damaged?

Or the IC can be damaged in some other way that I am missing?

One more thing please, how fast can I toggle the direction bit of the DRV?

0 Ahmed Iftikhar over 2 years ago in reply to Ahmed Iftikhar

Prodigy 20 points

I have monitored VDRAIN and it remains well within its limit.

Regards.

0 Eli Sorenti-Burns over 2 years ago in reply to Ahmed Iftikhar

TI__Intellectual 1350 points

Hi Ahmed,

Ahmed Iftikhar said:
I have tried to stop the motor gradually and the problem doesn't occur. How gradual this should be?

I believe the fastest way to do this would be to brake the motor first and then turn it off. In the previous thread, Adam mentioned that either slowing the PWM down or braking the driver before turning it off could solve this.

Ahmed Iftikhar said:
Doesn't the stored energy go to the MOSFET supply only (through MOSFET diodes), and if the voltage remains within the VDRAIN limit, the IC should not be damaged?

The voltage could potentially damage the SHx pins depending on its magnitude. Please see Figure 32 in the data sheet to see the internal gate output architecture.

Ahmed Iftikhar said:
One more thing please, how fast can I toggle the direction bit of the DRV?

We don't have this value specified in the data sheet. I'll have to run some tests in the lab and can get you a response by the end of the week. What is the minimum time you would need to toggle?

Let me know if this answers your questions.

Best,

Eli

0 Ahmed Iftikhar over 2 years ago in reply to Eli Sorenti-Burns

Prodigy 20 points

Hello Eli.

I wanted to share some more data with you.

I have checked 02 faulty DRV iCs with multimeter and following is the outcome.

Resistance between GHx/GND, GLx/GND and SHx/GND (both polarities) is in the range of kilo-ohms (greater than 150K in each case).

DVDD is ~5V,

VCP is ~VDRAIN+10V.

Only VGLS goes down to ~4V. At power-up the FAULT pin goes high for 10ms and then stays low and IC starts to get hot.

One more thing, I powered-up the DRV circuit first and then applied PWM with 50% duticycle directly, then exactly the same problem occurred. Motor stopped after taking a jerk and VGLS got down to around 4 volts. There is no problem whenever I start the motor at lower PWM dutycycle (e.g at 20% duty cycle) and go up gradually.

Is there any limitation to increase the duty cycle gradually too?

When motor is running all wave forms look normal on oscilloscope.

Eli Sorenti-Burns said:
We don't have this value specified in the data sheet. I'll have to run some tests in the lab and can get you a response by the end of the week. What is the minimum time you would need to toggle?

Minimum time to toggle the direction bit is 2ms.

0 Eli Sorenti-Burns over 2 years ago in reply to Ahmed Iftikhar

TI__Intellectual 1350 points

Hi Ahmed,

Thank you for the update.

Eli Sorenti-Burns said:
I'll have to run some tests in the lab and can get you a response by the end of the week. What is the minimum time you would need to toggle?

Apologies, I'll need a some more time to test this. I'll get back to you about it by the end of next week. 2ms should be fine but I'll double check.

Ahmed Iftikhar said:
At power-up the FAULT pin goes high for 10ms and then stays low and IC starts to get hot.

(+) [FAQ] Debugging faults with Hardware variant devices - Motor drivers forum - Motor drivers - TI E2E support forums is a good resource to utilize when looking at faults in DRV8353H. Are the gate drivers active or in Hi-Z after this fault occurs? It's likely a VGLS undervoltage fault. Section 9 of (2) [FAQ] Quick Guide to Debugging Common Issues in BLDC motor drivers - Motor drivers forum - Motor drivers - TI E2E support forums has some solutions to this issue.

Ahmed Iftikhar said:
Is there any limitation to increase the duty cycle gradually too?

I'll align with my team and get back to you on this.

Best,

Eli

0 Eli Sorenti-Burns over 2 years ago in reply to Eli Sorenti-Burns

TI__Intellectual 1350 points

Hi Ahmed,

Thank you for your patience. 2ms is fine. The electrical signals will toggle within nanoseconds, however the MOSFETs may take longer to discharge. The main limiting factor is how long it takes for the rotor to change momentum. Still checking about the duty cycle.

Best,

Eli

0 Eli Sorenti-Burns over 2 years ago in reply to Eli Sorenti-Burns

TI__Intellectual 1350 points

Hi Ahmed,

I'd like to clarify a few things before digging deeper into this debug.

Ahmed Iftikhar said:
Motor stopped after taking a jerk and VGLS got down to around 4 volts.

What do you mean by 'taking a jerk'?

Ahmed Iftikhar said:
Is there any limitation to increase the duty cycle gradually too?

This depends on motor load. Example: If you have very small load on the motor and you slowly increase or increase it fast, the DRV will perform fine. However, if you have a large load and a fast transition then this can produce some back pumping. Please watch the last half of the video Stages of Motor Control located here: Precision labs series: Brushless-DC motors | TI.com. Finding the limitations of this will require some testing on your end. Additionally, here is an article on back pumping that will give you some ideas on how to modulate your duty cycle.

Best,

Eli

+1 Eli Sorenti-Burns over 2 years ago in reply to Eli Sorenti-Burns

TI__Intellectual 1350 points

Hi Ahmed,

Please mark the thread as resolved if this has solved your issue. I recommend doing some testing regarding the limitation on increasing duty cycle as it will depend on your motor load. Feel free to post again on E2E if you have any other questions.

Best,

Eli

0 Ahmed Iftikhar over 2 years ago in reply to Eli Sorenti-Burns

Prodigy 20 points

Hello Eli.

I am doing different tests right now. I'll update if I have some other questions.

Thanks.

Regards.