STM32H7 LSE Hardfault

[Eralp]

I'm using external crystal for RTC on STM32H750VB. Sometimes when i powered up to board, its hanging on LSE initialization and entering hard fault. If i remove rtc backup battery and plug it again, its working. This happens very rarely :S I couldn't find where the problem originated.

[waclawek.jan]

Isn't the hardfault consequence of badly handled timeout? LSE may take surprisingly long to start up, 10 seconds is not an unlikely time.

JW

[Eralp]

Yes, it's true. But it's not releated to crystal. I mean, when this situation happen i can power up/down board a lot of times and LSE always fail. But if i remove RTC Backup battery and plug it again LSE surprisingly decide to work. I couldn't find the relationship between RTC backup battery and LSE?

[waclawek.jan]

> I couldn't find the relationship between RTC backup battery and LSE?

LSE is powered from VBAT. Review supply voltage domains description in datasheet.

If LSE is prone to failure, you should try setting different LSE drive strength in RCC_BDTR, or review the whole LSE circuit design, read AN2867. You may want to start with a "known good" board such as Nucleo or Disco.

JW

[Eralp]

> LSE is powered from VBAT. Review supply voltage domains description in datasheet.

Yes it's true if you have not powered board. But i think i understand the problem now. When rtc backup battery attached, lse always powered on and if it stops working the only way to reset to lse is remove backup battery.

> If LSE is prone to failure, you should try setting different LSE drive strength in RCC_BDTR

I have already tried low/medium/high LSE drive capability.

Thanks for reply.