Motor Control Issue with MCSDK 6.4.0

[kumarairia]

Hello,

I am working with the STM32F303E-EVAL control board and the STEVAL-IPMNM2S power board, using IAR Embedded Workbench 9.40.

I have struggled to get the motor running. I initially tried both MCSDK versions 9.3.2 and 9.4.0 separately, but in most cases, the motor either does not rotate or runs with excessive noise — overall, it fails to operate with error.

However, on one occasion, I accidentally recreated a project using MCSDK 9.3.2 and then rebuilt it using 9.4.0. Surprisingly, that build worked — the motor rotated smoothly, and I was able to Motor Pilot control speed and torque successfully.

Since then, I have tried many times (over 100 attempts) to reproduce the same working setup using both versions, separately and in combination, but without success. The motor fails to start or runs poorly.

Could someone please help me understand what might have gone right in that one successful attempt and how I can reliably replicate it?

Thank you in advance for your support.

[GMA]

Hello @kumarairia,

Did disabling both auxiliary sensors and tuning the pulse-width modulation (PWM) frequency improve performance?

[kumarairia]

Hello Sir,

Sorry for the delayed response.

After spending significant time and money working with STM and the MCSDK 6.3.2, 6.4.0, and recent 6.4.1, I have to share that the experience has been very disappointing. The framework has caused unnecessary delays and stress, and in my opinion, it is not suitable for mass production in its current form.

I have bought so many boards both power and control boards.

I have uploaded everything for your reference. The SDK has multiple issues and is not user-friendly, which makes the development process much more difficult than expected.

Going forward, I would suggest avoiding presenting this as a production-ready solution, as it sets false expectations.

Regards,

[GMA]

Hello @kumarairia,

Thank you for your feedback.
We put significant effort into developing the best solutions we can provide, and we are sorry that it does not meet your expectations.

To improve our product, referring to your video, once stepping through the NVIC initialization function, the debugger continues running.
This issue could be caused by an incorrectly mapped interrupt that is enabled by the NVIC initialization. If you pause the debugger at the infinite loop point, does it stop at a specific IRQHandler position in the startup_stm32f303xe.s file?

[kumarairia]

Hello Expert,

I do not believe this issue is on my side. I have tested multiple times using MCSDK versions 6.3.2 and 6.4.0, and the problem still persists.

When I tried version 6.4.0 with IPMNM2S, I noticed that with driver protection enabled, the issue disappeared. However, as soon as I enabled OCP, the same issue occurred again.

Based on these tests, I think this may be a bug in the MCSDK related to the internal op-amp and comparators.

I have attached here picture what happens when i pause debugger. 

I need help to solve the issue please.