Default HSE set to 25 MHz for NUCLEO-F722ZE

[Anton Martinsen]

Is it intented for STM32CubeMX drivers for NUCLEO-F722ZE to have 25MHz as HSE input by default? The hardware I bought from ST had an 8MHz oscillator as input for HSE, not 25 MHz.

[Amel NASRI]

Hi @Anton Martinsen​ ,

This is reported to our STM32CubeMX development team who will take care to bring required corrections.

Thanks,

-Amel

[Tesla DeLorean]

Seen the stm32xxxx_hal_conf.h​ contain the wrong HSE_VALUE on occasion as examples moved from EVAL or DISCO series boards with alternate clocking for Ethernet, etc.

As a general rule I tend to print CPU and bus speeds to a debug console so I can quickly check they are as expected, helps when dealing with dozens of boards.​

[Amel NASRI]

Hi @Community member​ ,

Which example please? What I noted is the difference between stm32f7xx_hal_conf.h and system_stm32f7xx.c regarding HSE_VALUE for same project. This has to be highlighted as well.

-Amel

[Tesla DeLorean]

It was more of a historic observation, a quick grep yielded this non-comprehensive list

STM32Cube_FW_F7_V1.16.0\Projects\STM32F767ZI-Nucleo\Examples\FLASH\FLASH_SwapBank\Inc\stm32f7xx_hal_conf.h

STM32Cube_FW_F7_V1.16.0\Projects\STM32F767ZI-Nucleo\Examples\SPI\SPI_FullDuplex_AdvComIT\Inc\stm32f7xx_hal_conf.h

STM32Cube_FW_F7_V1.16.0\Projects\STM32F767ZI-Nucleo\Examples\SPI\SPI_FullDuplex_AdvComPolling\Inc\stm32f7xx_hal_conf.h

[Anton Martinsen]

Hello @Community member​ ,

I am pretty much a beginner at embedded development so if I misunderstand something here, I apologize.

I do not know how I should have printed anything, the main reason I came across this was because I could not print anything over UART or ITM_SendChar.

I tried looking at SystemCoreClock but this was showing the "correct value" because it was calculated from the defined HSE value.

Eventually I noticed there was some activity on UART and I started guessing baudrates until I found there was stuff being printed but with a baudrate of a factor 3 lower than I thought I had setup the UART.

Then I noticed SPI and ITM had the same behaviour which then lead me to believe HSE was wrong. Then I noticed there was no 25MHz oscillator on the board, only an 8MHz.

Best regards,

Anton

[Amel NASRI]

Thanks @Community member​ .

This is reported to our STM32CubeF7 development team in order to review examples and set the correct HSE_VALUE.

-Amel

[Tesla DeLorean]

ST just has history here with EVAL boards using Ethernet and a single 25 MHz crystal, this includes a handful of DISCO boards.

The bulk of the DISCO/NUCLEO boards use the 8 MHz clock source output by the ST-LINK/V2. The ST-LINK/V3 clocks at 25 MHz, and outputs an awful 8 MHz

Occasionally when examples have been cross-ported there is a mismatch between HSE_VALUE and reality, this might also pollute the PLL settings, sometimes not. It usually shows up as a disparity in baud rates or core speed computations. I occurs sufficiently often, either with stock examples, templates, or user porting efforts, that it gets called out high on the check-list, and especially when the user complains about a 3X off baud rate.

The parts also start off the HSI (or MSI), so the HSE is usually not required, and earlier SPL code had fall-thru behaviour for HSE or PLL not starting rather than vectoring off to Error_Handler()

[Nesrine.JLASSI]

Hello @Anton Martinsen​ 

Thanks for the feedback, issue reported to the tool development team that will plan the correction.

Best regards,

Nesrine