Experimental Physics and
Industrial Control System

"Liyu, Andrei" <[email protected]> · Thu, 03 Nov 2005 12:29:44 -0500

Hi, All

I wouldn't refuse to see Your comments :)

1. I looked on alarm code in ai record. After real actions and before
monitor process() calls

            checkAlarms(pai); 

checkAlarms() function belongs to ai. But other records types have their
checkAlarms() functions.

Then checkAlarms() function calls recGblSetSevr( precord, ...) function.

It is macros. So we can EASY change macros and new function will write
message to errlog server (I don't know another suitable central engine
in Epics).

	After that we will have all alarms in one place. I believe that
advantages are clear.

1a. Just for information. D0 in Fermilab has done something similar.
They attached "hook to alarm". I don't see this way. I explain way to
substitute code. It could be way with "memory substitution". But this is
real hack and I am not sure it will work on any platforms. 

2. I also have another misunderstanding.
	As I understood ALARM situation is when parameters go out of
range but equipment works perfectly. If equipment doesn't work correctly
it is ERROR situation. 
	In simple (within device) ALARM situation operator tries to tune
parameters that go out from ALARM situation. If operator couldn't do
that he stops beam because it will become ERROR situation in machine
level. 
	In simple (within device) ERROR situation operator stops beam if
it is technology or disable diagnostic detector. 
	So ALARM and ERROR are basically different. 

	Now Epics alarm severity has 4 definitions: NO_ALARM,
MINOR_ALARM, MAJOR_ALARM and INVALID. There are some sentences in Alarm
specification of Records book. It seems INVALID doesn't belong to ALARM.
It belongs to ERROR.

	Can anyone share another base idea about alarm/error in control
system?

Thanks, Andrei.

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System