STM32 U575 LSE Clock doesn't start.

[sneal]

LSE clock is set in the Cube and initialised

After intitialisation the following registers are not set correctly and the LSE clock isn't called

RCC->BDCR->LSEDRIVE not set to hight event though macro called

RCC->BDCR->LSCOSEL set to LSI

RCC->BDCR->LSCOEN not set

This looks to be a bug in the HAL library

[Mike_ST]

Hello,

If you can, please share the.ioc file as you are using CubeMX or IDE, or share some code, so that the community can check/reproduce/help.

Does this happen on a nucleo board ?

>> This looks to be a bug in the HAL library

That can happen, but we need more inputs.

Thank you.

[sneal]

Hello,
Yes this happens on a nucleo board as well
Have attached project
Sue

[sneal]

HI Mike,
I managed to get the LSE clock running by adding this code
/* USER CODE BEGIN SysInit */
// re-jig the LSE clock
// STOP LSE clock
CLEAR_BIT(RCC->BDCR, RCC_BDCR_LSEON);
__HAL_RCC_LSEDRIVE_CONFIG(RCC_LSEDRIVE_HIGH);
SET_BIT(RCC->BDCR, RCC_BDCR_LSCOSEL);
SET_BIT(RCC->BDCR, RCC_BDCR_LSCOEN);
SET_BIT(RCC->BDCR, RCC_BDCR_LSEON);
/* USER CODE END SysInit */
Sue

[sneal · Accepted solution]

Hi Mike,
I found the problem
I was calling __HAL_RCC_PWR_CLK_ENABLE();
Before SystemClock_Config which was why the __HAL_RCC_LSEDRIVE_CONFIG(RCC_LSEDRIVE_HIGH)
Didn’t work.
All fixed now
Kind regards
Sue

[waclawek.jan]

> I was calling __HAL_RCC_PWR_CLK_ENABLE();

> Before SystemClock_Config which was why the

> __HAL_RCC_LSEDRIVE_CONFIG(RCC_LSEDRIVE_HIGH)

> Didn’t work.

I fail to see the causality here.

JW

[sneal]

__HAL_RCC_PWR_CLK_ENABLE()
Enabled the LSE clock so that the call to __HAL_RCC_LSEDRIVE_CONFIG(RCC_LSEDRIVE_HIGH)
Didn’t work as the clock was enabled and the calls can only work if disabled.

[Piranha]

Enabling the clock for PWR peripheral has nothing to do with RCC_BDCR register, in which the LSE is configured.