Experimental Physics and
Industrial Control System

Ned Arnold <nda> · Wed, 22 May 2002 12:42:57 -0500 (CDT)

>   I have a monitoring problem in our EPICS system. Someone (probably)
> is turning on the pulser calibration mode for part of the muon system
> in the D0 detector. I want to find out if this is happening through
> the EPICS system or via one of the several other paths of destruction
> that the hardware types leave for themselves.
> 
>   So, we set up a channel archiver to watch for changes in the value
> field of the EPICS longout record that is connected to the hardware
> control register.  This record has the SCAN field set to "Passive";
> so, it will only process when the VAL or PROC fields are written.
> Unfortunately, the archiver will only see changes of value and, if the
> same bit pattern is written to the record's VAL field, the archiver,
> which is using a monitor, will not record the transaction because the
> value does not change.
> 
>   Well, when the record processes, for whatever reason, the value of
> the timestamp changes and I can just archive changes in the TIME
> field. Well no! Someone decided that the TIME field should have the
> DBF_NOACCESS property (it seems to me that SPC_NOMOD would have been
> adequate to protect the timestamp, but perhaps there are other
> considerations) and, so, the archiver cannot connect to it.
> 
>   This leads to the question: How can I archive an entry every time
> that an output record processes?
> 
> 						Fritz

Fritz,

By setting the .ADEL and .MDEL to -1, you will force the record to 
"post monitors" every time it processes.

	Ned

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System