DRV5015-Q1: Output stuck at high or low, can recover

Hey Frank,

Thanks for providing those details.  I am going to forward that to our design team to see if they may be able to provide any insight.  I will try to give you a status update before the end of the day.

In the mean time, can you clarify that GLASS_HALLx_DN_PBx corresponds to the DRV5015 out pin?

Hi Patrick,

Thanks for the discussion on webex. I summarized the latest information here:

• We checked the layout of the supply caps, they are on the bottom layer, the same location as the DRV5015 on the top layer. Please comments if this layout would be a risk for noise at input.

• One thing that I noticed regarding to the layout, the pink trace is the power supply (3.3V), it connects to left part first and then connect to part on right.
• According to the issue description, only the DRV5015 part on the left has problem. The 5015 on the right always good.
• Maybe the magnetic flux over these two DRV5015 are different?
• The 3.3V power supply rail waveform is the purple curve in previous waveform; it is steady 3.2V when the issue happened.
• Here is a better resolution waveform here. The output waveform shows there is PWM when problem happened, but looks like it was pulled high externally.

Regarding to these, we have some questions:

1. Please comment if this layout and caps placement can still cause the noise.
2. If the magnetic flux is not enough, will DRV5015 output a middle voltage between 0V and VCC? Or it only has 0V and VCC two output levels?
3. Is there any kind of condition can cause the DRV5015 latch or stuck? Or is there any possibility can make the internal FET cannot turn on?
4. We tried to change some condition after replicate the issue. If we change the motor or move a magnates above the DRV5015, it can recover. Does change the magnetic field over device can make it recover?
5. Regarding to the waveform, do you think it might be external strong pull up cause the output high?
6. Please kindly review the schematic, does the 3.3V and 4.7k pull ups ok?

Regards,

Frank

Hello Frank,

Thanks for providing those details.  I will be looking further into this issue tomorrow and try to give you an update then.

Hey Frank,

I still have not gotten feedback from our design team.  However, in light of some of the details provided by you, I have a few thoughts on what might be happening.  I am thinking that one device is properly positioned while the other is not.  By that I mean the max and min magnetic field values observed by both devices is not the same. I suspect the left device in particular probably sees smaller max and min values.  As you ramp up the motor to full speed you approach closer to the bandwidth of the part, where the Vout/Vin = 1/2.  This could ultimately lead to the voltage feeding into the internal comparator falling a bit shy of the threshold. If you are consistently seeing this across multiple boards, then the tolerance on the lot of devices your customer bought might be tight enough, that you could swap the left with the right device and see the error at the same spot on the board.  Alternatively, if you are able to drive the motor without the DRV5015 feedback, you might try placing a linear sensor like the DRV5056 and observing what kind of output it has at the higher pwm frequencies.

Regarding the questions you had above.

1. The second capture looks cleaner, thereby suggesting the prior capture might have aspects that could have been improved regarding the scope or probe.  I think the cap placement is ok.

2. if Magnetic flux does not change sufficiently as the polarities toggle, then device may remain stuck in last state.

3.  Waiting for design team to comment on whether they are aware of any way that such a latch up could occur due to design.

4. I think slow movement allows for sufficient signal to trigger the internal comparator.

5. at this point I dont think the pull up is the issue.

6.4.7k pull up and 3.3V supply seem fine.