Timers and Macros Quitting After "X" Days

[andre]

Hi Roger, The caps I used were ceramic type
X7R.  These are good high frequency
decoupling caps.  The electrolytic caps C11
and C12 have too high a series inductance
and are ineffective at high frequencies.
The linear regulator U5 has a really poor
filter characteristic so without the
ceramic caps around it, it's garbage in,
garbage out. Unfortunately the
microcontroller's reset circuit can't
respond to noise and doesn't react.  I
suspect that the watchdog is iether
disabled in the firmware, or wasn't given a
restart vector so it's ineffective as well.
My unit chalked up yet another night of
sane operation:)
Cheers.

[roger1818]

Andre:  What you are saying about the
capacitor makes sense.  This could explain
why some CM15As seem to be more reliable
than others.  The ones that work better
could be using capacitors that happen to
have a series inductance that is low enough
to provide adequate filtering.  The
“diagnostic” units probably have capacitors
from an especially good batch.

As far as the watchdog is concerned though,
I believe it is enabled, but the problem is
that since it uses a software time of day
clock, a reset causes the clock to stop
until it has been set again.  As a result,
anything with time dependence (timers and
macros with time delays) won’t work.