[sdiy] In praise of the ATM STM32F303

[Neil Johnson]

Hi Simon,

> So maybe they improved it (the application note that the errata links to does
> not say whether the F446 is among the types in which they did).

I'm not surprised AN4073 doesn't mention the 446 parts - the app note
is dated July 2013 whereas the 446 errata is dated March 2015.  So I
think it safe to assume that whatever is in the app note also applies
to the later devices, they just haven't updated Table 1.

> The application note also is unclear about how much performance is lost when the
> recommendations are followed (i.e. disabling accelerator functions). Maybe for
> reading out pot voltages, or even for audio sampling, it may be more efficient
> to keep the processor running at maximum speed and spending some of the
> conserved MIPs on averaging ADC values.

I would read that document as saying they recognise their chips have a
problem and then present some ideas of how to fix it, but quite
honestly you need to experiment for your particular application.  They
aren't/can't write an errata or app.note that solves the problem for
all possible applications.

Neil
--
http://www.njohnson.co.uk