Experimental Physics and
Industrial Control System

"Bob Dalesio" <[email protected]> · Wed, 10 Apr 2002 12:44:12 -0600

> What I finally did, and seems to be running fine, was a
> solution using 
> the IVOA in the calcout, sending a "Reject-alarm" to the
> user and reject 
> the value - the invalid value is not sent out.
> In fact, with this combination I could implement both the
> options in
> Kay's original posting (although I had to play with the
> alarm states
> a bit - "Invalid" versus limit checks, like in what
> Andrew wrote.)
I want to recall that the IVOA - invalid action - was
intented to be used wih Maximize severity. If you had a
control logic where any of the records in the chain were
INVALID, then the output needed to decide how to handle it.

In summary, I favor catches both Timo's and Kay's situations
in the AO record clamping code. This is before alarms are
set, IVOA is taken and monitors are posted. The discussion
for this preference follows.

If this functionality is changed as suggested above - what
occurs in these two cases:

The value was outside the limit, the state is set to
invalid, the IVOA catches it after alarms are checked but
before monitors are posted. Does it change the value to the
in-range value and leave the INVALID alarm? Or does it try
to remove the INVALID alarm? If it tries to remove the
INVALID alarm, it needs to know if the INVALID was the
result of the MAXIMIZE severity and an INVALID Input as well
as the limit being exceeded. I think it would be more
straight forward to make the value in range and leave the
INVALID alarm. It may be confusing, however.

When the value coming in throgh the DOL link / Maximize
Severity link is the reason for the INVALID alarm, the IVOA
that says keep the old may now get an older value. The code
to check this behavoir would also need to look and see what
caused the INVALID alarm.

On the other hand, the suggestion to include a field for
what to do when a value is out of range seems a seperate
issue to me. A value can be any combination of in range/out
of range and VALID/INVALID.

The value clamping occurs before alarms are checked. NaN and
DRVH/DRVL range should be caught at this time. Perhaps we
should add the ability to set the alarm condition and
severity for this error. If we set it to INVALID, then we
will also get the behavior of IVOA. If we set it to MAJOR,
we have the value either clamped or not changed at all, a
MAJOR alarm severity and a RANGE ERROR on the alarm status.
If we set it to INVALID, the clamped/reset value will also
be subjected to the possible IVOA choices - no action, go to
know value, do not drive output.

Bob

Bob Dalesio
[email protected]
410 557 0297 Maryland
505 667 3414 Los Alamos
505 699 1632 Cell Phone
505 667 6087 Our Secretary (Judy)

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System