Re: async UART errors due to frequency

--- In jallist@yahoogroups.com, "Wouter van Ooijen" 
wrote:
> 
> For everyone who wants to apply the 5% rule to a *software* UART:
make
> sure to take into account that the start bit might be detected some
> (arbitrary) time after its starting edge when you use polling.
> 

Good point! Thanks for bringing us back down to earth!

I guess polling the start bit at 16X would be hard to achieve, plus we
might "ignore" the "uart" for awhile, assuming we have "real" work to
do.

If you could afford to waste some bandwidth, perhaps you could require
a "break" at the front of each message, or even 1 or 2 characters of
0xAA or 0x55.

Anyway, like I said at the very beginning of this-- I plan to use a
crystal most of the time for my UART work.

William