DRV10987: Issue at Open to Closed loop transition..

Ernest Cho

Part Number: DRV10987

Hi team,

My customer is producing DRV10987 for an year, but recently they have issue regarding initial motor stopping at the open to closed loop transition.

And I found below blog post for open loop and it mentioned "start lose sync and stall". I think this is my customer issue.
But the thing is they are using for year without this issue.

https://e2e.ti.com/blogs_/b/industrial_strength/posts/integrated-intelligence-part-2-startup-reliability

"During an open loop, the motor is driving without any information about the position of the rotor. This phase of open-loop commutation, also known as blind commutation, is very important because it is directly related to system reliability. If blind commutation is not configured correctly, the motor will start, lose sync and stall."

My question is what could make this issue? like board noise, missed or not enough time of BEMF calculation?

FYI,

Issue rate is 23units /100ku.

(belows are evaluated for testing)

SPEED PWM duty is 33% as a starting point. --> Once decrease it like 11% it work well .

customer VCC is 26.3V which is little bit higher. --> If decrease VCC with external P/S like 24 or 23V then it work well.

Rated speed is 200Hz and Op2ClsThr = 25.6Hz. --> If increase Op2ClsThr like 51.2Hz then it work well.

Unfortunately, it is not seen any of fault flag. device is restarted after 5 sec.

Best Regards,

Ernest Cho.

over 2 years ago

+1 Isaac Lara over 2 years ago

TI__Mastermind 19640 points

Hey Ernest,

Thank you for your question, it seems like this is being handled over email so I will close this thread.

To note I am not sure if all these experimental changes were all done at the same time or performed separately, but increasing the Op2ClsThr to 51.2Hz gives the motor more time in the open loop before it goes into close loop operation. Perhaps the extra time given allows the motor to catch up reflecting good BEMF at the time of the handoff. A lower starting point might have a similar effect, typically when the motor stops when going from the open to close loop transitions decreasing the acceleration or increasing the open loop current tend to be a possible solution to this problem. I hope this helps!

Best,

Isaac

0 Ernest Cho over 2 years ago in reply to Isaac Lara

TI__Intellectual 2961 points

Hi Isaac,

Thank you for your feedback. Your comments help to me.

Isaac Lara said:
decreasing the acceleration or increasing the open loop current tend to be a possible solution to this problem.

My experiments performed separately. And any modification cause decreasing acceleration have effect of this issue, but increasing open loop current does not have effect to me.

One more quick question is : Why the fault flag did not set for this issue? I read FaultRegister(0x00) after the issue, but it was 0x0000.

Best Regards,

Ernest

0 Isaac Lara over 2 years ago in reply to Ernest Cho

TI__Mastermind 19640 points

Hey Ernest,

No worries glad to help out. I am currently out of the office today and will not be back until Tuesday due to US holiday, but I will work with Vishnu to look into this issue for you. Thank you for your patience!

Best,

Isaac

0 Isaac Lara over 2 years ago in reply to Isaac Lara

TI__Mastermind 19640 points

Hello Ernest,

Thank you for your patience. The fault I believe could be an issue would be the lock 4 or 5 fault, where the device is stuck in open loop. You have to enable the faults in order for them to work this is done in register 0x94, did you ensure that these were enabled to demonstrate the fault?

Best,

Isaac

0 Ernest Cho over 2 years ago in reply to Isaac Lara

TI__Intellectual 2961 points

Hi Isaac,

Thanks for your feedback.

I did test with all Lock flags enabled(prev. setting:Lock1,2 disabled). Test was not flagged on the GUI as well as Fault reg(0x00) by using customer parameter. But once I change the Op2ClsThr from 25.6Hz to 12.8Hz or lower, then Abnormal Kt flag was set when motor was stopped. So I assume the motor stopped by abnormal KT in customer case. Let me know if I was wrong. (But how was it stopped w/o LockEn2 =1)

Sorry but one more question : Abnormal KT fault during Open to closed loop transition can predict by external noise or measure by something else? Customer want to provide evidence that this issue was not due to the device deviation.