In Normal mode, your recovery log usage should reduce on commits rather
than on DB Backups, which means your recovery log usage should constantly go
up and back down throughout the day.  If you're maxing out your recovery
log, something is likely "pinning the tail" of the log.  I would suggest you
find out what is pinning log and resolve that issue.
   As an example, I had a user trying to back up a 20GB database flatfile
from an AIX client over Token Ring.  As it was a backup of a single file,
the transaction never committed, so the transaction was pinning the log and
causing me to run out of recovery log space.  My solution was to have them
to backup the file over the ATM backbone.  The backup then didn't pin the
log long enough to fill the log.

Good luck,

Alex Paschal
Storage Administrator
Freightliner, LLC
(503) 745-6850 phone/vmail