Re: [libopenstm32-devel] usart_set_baudrate(...)

[Piotr Esden-T. <pi...@es...>]

Hi!

On Mar 12, 2011, at 4:52 PM, Gareth McMullin wrote:

> On Thu, Mar 10, 2011 at 2:34 AM, Nico von Geyso <Nic...@fu...> wrote:
>> I had some problems with the usart_set_baudrate() function and so I took
>> a deeper look how it works.
>> The baud rade generation for stm32 is described in the section
>> "Fractional baud rate generation" in RM0008 Reference manual, Page 747.
> 
> If I read it correctly, all the talk in the manual makes it sound fancy and
> complicated, but the simple division calculates the correct value.

Yes that is what I figured myself. This sounds like marketting talk to make the divider sound more awesome than it is. :)

I updated the rcc code to store the clock speeds of the apb1 and apb2 into variables and use them to calculate the baud divider, using just a simple division. Tested it a little and it seems to work like a charm. just pull from my fork repository and give it a try. :)

> 
>> For the stm-h107 development the current implementation works fine for
>> usart 2,3 but not for usart1.
>> Usart1 has a different clock source (PCLK2, 72Mhz) and so you have to
>> divide the baud parameter (for usart_set_baudrate(...)) by two to get
>> the correct values.
>> As it is hard to get the current pclk1 or pclk2 values, wouldn't it be
>> better if there is a function like the following:
>> 
>>        void usart_set_baudrate(u32 clock, u32 usart, u32 baud); //set the
>> clock manually
> 
> I think this is the most obvious way to do it.  I would swap the clock and usart
> parameters though.

I think it is better to track the clock speed of the system and have the baud rate still be calculated automatically...

Cheers Esden

--
Blog: http://www.esden.net
Projects:
http://open-bldc.org
http://paparazziuav.org
http://github.org/esden/floss-jtag
http://github.org/esden/summon-arm-toolchain