Experimental Physics and
Industrial Control System

Steve Lewis <[email protected]> · Wed, 5 Apr 2006 12:31:22 -0800

Try this (one of the best features of ALH): use
$ALARMCOUNTFILTER count seconds
on each PV. Basically, it suppresses the triggering of the alarm
until at least 'count' transitions have occurred within a window
'seconds' or the alarm level must hold for 'seconds' long. So if you
get a 'nuisance' MINOR once every 30 seconds that lasts about 10
seconds, try count=2, seconds=60. Great for water bubbles in your
cooling system...
Another techniques is to set up a big chain of dfanout records to
push alternately MINOR or NOALARM to the SEVR field (HIGH, HIHI,
whatever you are using) based on a top-level bo record with labels
like "TUNE" and "RUN".

At 8:08 AM -0400 2006/04/05, Terry Carlino wrote:

As an AlarmHandler user it has always struck me that there should be
a better way to manage alarms. In the real world it is quite common
to have a device sitting just on the edge of the MINOR alarm point,
creating nuisance alarms. We do not generally manage alarms by
actively changing HIGH or LOW, since these are set by system
experts. This leaves the user with the choice of disabling the
alarm, which means that if the device reaches the MAJOR parameter no
alarm will be heard or dealing with the nuisance alarms. Bad enough
when a single channel behave such. When you have thousands of
channels, not uncommon for a few to have this problem, especially
during times of general machine instability, such as start up. It
would be nice to be able to separately acknowledge or bypass MINOR
alarms, a kind of "yeah we know we need to look at you, you're
alright for now call us when you get to the MAJOR setpoint.
Terry

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System