Does somebody know what is 'usart_ker_ckpres' in USART BRR description?

[remixoff_cr]

Hi all,

Just read RM0440, the section is USART Baudrate generator. There is the formula contatins input never met in all other document. (found 3 times, all of them are on the screenshot).

 

What happened with ST technical writers? I never met cases like that 10 years before.

 

[Imen.D]

Hello @remixoff_cr,

Thank you for pointing out this issue.

usart_ker_ckpres should be usart_ker_ck_pres, which is the peripheral input clock divided by a programmed prescaler.

I highlighted your request to the appropriate team in internal ticket number (ID 168184) for correction and improve this part.

(PS: internal ticket number (ID 168184) is not accessible or usable by customers).

Thanks for your contribution.

Imen

[Tesla DeLorean]

The F1 was simpler.

It is the instance clock, so nationally APB1, APB2, or the peripheral clock source routing depending on which the specific U(S)ART is plumbed too.

[waclawek.jan]

There's also an additional prescaler within UART itself, so that's why usart_ker_ck (which comes from RCC) is not the same as usart_ker_ck_pres .

What's still wrong - and with increasing complexity gets worse - with the STM32 RMs is, that the individual peripherals chapters are written strictly from the perspective of separate IPs, and the Interconnect chapter - which should connect them all - is severely lacking. In this case, it's either RCC chapter, or the Interconnect chapter, or both, which is to explain, that uart_ker_ck is output from primary and derived clock sources of RCC through muxes controlled by RCC_CCIPR.[LP]U[S]ARTxSEL bitfields; and, ideally, also description of those bitfields in RCC chapter could contain references to usart_ker_ck.

JW