On Sun, 16 Nov 2014 11:24:05 -0800 (PST)
Sean Sill <
sms...@gmail.com> wrote:
> That's probably the difference then. Richard did you overclock yours?
Yes, I think it is (I followed the recommendation on the Ola Pi page, I
think). I'm surprised overclock has this effect, because the UART
driver is smart enough to work out the correct multiplier based on the
clock rate configured on the command line (this fix has been in
Raspbian since before I started), which is supposed to make the speed
selection clock-independent (the clock only restricts the range which
can be selected).
The ola code does read what the UART reports as the achieved in/out
speed, but all it does is log it - there is no check / error if it is
out of range.
Christian, I'd really like to know what speeds the log on the Pi
reported when you were getting the wrong speed. I know without the
parameter at all, you end up set to 9600 baud (I have an ongoing todo
item to catch this case and print something informative), I would like
to know what happens in your case as well.
Can you confirm that you are using a reasonably up-to-date Raspbian
image on your Pi - it sounds like I need to do some testing!
Back in this list archive somewhere (22/02/2014 'Raspberry Pi built-in
UART') is a simple C program I wrote for checking if the correct baud
rate is achieved, which sends a magic byte value to produce a continuous
square wave (at the bit period), which is what I used for most of my
testing. This is easier than full DMX waveform to inspect with an
analogue scope!
Richard