HSEBYP on Nucleo-L073RZ

[srontje]

I attempted to override SystemClock_Config() to use HSE (8 MHz provided by the ST-Link) instead of HSI. So I copied the existing code, modified a few bits and failed. Then I rewrote the whole function using direct register access, but that failed as well. Then I tried another minimal code which directly uses HSE without PLL, that failed as well.

Problem: In all three cases, the CPU and also the systick timer ran at half speed instead of the configured one.

I verified the external clock signal (coming from the ST-Link) with an oscilloscope. It looks stable and has really 8 MHz. After hours of investigation I found a solution: set the HSEBYP bit.

But I did not find any explanation in the reference manual. It only mentions that this bit disables the output of the HSE oscillator, so the pin can be used as a regular I/O. But that was not my problem. My problem was half speed.

I did never have this issue with other STM32 models (STM32F103C8, STM32F103RB, STM32F303CC, STM32F303RE, STM32G432CB).

I tried the same codes without framework in the Cube IDE, and got the same issue. So I am aware that the question is a bit off-topic, but I hope that someone can help me anyway to understand the issue.

failed1.c
failed2.c
failed3.c

[cookesan]

For a Nucleo board where the 8 MHz clock is coming from the ST-LINK MCO into OSC_IN, RCC_HSE_BYPASS is the correct configuration. That signal is an external clock source, not a crystal/resonator connected across OSC_IN/OSC_OUT.

So the clock setup should use:

RCC_OscInitStruct.OscillatorType = RCC_OSCILLATORTYPE_HSE;
RCC_OscInitStruct.HSEState = RCC_HSE_BYPASS;

and HSE_VALUE should match the actual 8 MHz input. Running it as normal RCC_HSE_ON makes the MCU try to drive/use the HSE oscillator circuit as if a crystal were attached, so the half-speed behavior is a symptom of an invalid clock-source mode rather than a valid alternate setting.

Also check the Nucleo solder bridge/user manual state for the ST-LINK MCO path, but if you have already measured a stable 8 MHz on OSC_IN, bypass mode is the right fix.

[fpistm]

Hi,
this is defined in the User manual of the board:
https://www.st.com/resource/en/user_manual/um1724-stm32-nucleo64-boards-mb1136-stmicroelectronics.pdf

Nucleo L073RZ I have is the MB1136 C-04 so HSE is provided by the STLink and so HSE have to be configured bypass.
You can also have a look to the hardware configuration (SBx and Rx).

[srontje]

Thank you. Reading the Reference manual, I thought that using HSEBYP is optional.

I measured the IN and OUT signals of the wrongly configured oscillator. Looks interesting and seems to confirm that the wrong configuration cannot provide a stable clock.

MAP001.BMP
MAP002.BMP