From: Mike Holden <rockbox_at_mikeholden.uklinux.net>
Date: Sun, 20 Jul 2003 15:40:47 +0100 (BST)

I've been doing some more testing on this, and I think I have seen a
pattern emerge. It looks like the 'ifdef code I posted earlier isn't the
cilprit after all, it just looked like it was as that was what I was
working on at the time ;-)

I have my FM with the flashed Rombox that was originally distributed
(030714-2210). I have various Rockbox firmwares on the hard drive. The
testing I have been doing recently is with a fully up-to-date CVS version
from about 30 minutes ago.

Rombox boots up cleanly and shows the battery level correctly, and also
prompts for resume as Resume=Ask.

If I then ROLO the clean up-to-date firmware, sometimes the battery level
is displayed correctly, and sometimes it is really low. Correct value is
around 90%. When it is displayed incorreclty, it is usually around 7, but
sometimes can be around 70.

The pattern seems to be that if the disk is already spinning at the time
you "play" the firmware, then it ROLOs it correctly. However if the disk
has spun down, then that seems to be the factor that causes it to not
display the level correctly. It seems to also screw up the resume as well,
as the resume prompt usually (but not always) appears and then disappears
without waiting for a keypress.

So maybe this is a disk spinup issue in ROLO? It still does not seem to be
100% reproducible, but seems to happen most of the time.

Further, if I look in Debug->Battery info when the battery level is
incorrectly displayed, when I first go in and look at screen 1 (battery
graph), the battery level is shown very low (0.02 V), but if I page down
to screen 2 (status page) then the voltage is shown correctly (4.00 V)

-- 
Mike Holden
Rockbox page: http://www.mikeholden.org/~rockbox

Received on 2003-07-20