RE: Subroutine invoked only when PV limit exceeded?

["Poff, Mark A" <mark.a.poff@boeing.com>]

Thanks, Andrew, I'll try that.

-Mark

-----Original Message-----
From: Andrew Johnson [mailto:anj@aps.anl.gov] 
Sent: Monday, July 28, 2014 2:26 PM
To: Poff, Mark A; tech-talk@aps.anl.gov
Subject: Re: Subroutine invoked only when PV limit exceeded?

Hi Mark,

On 07/28/2014 01:01 PM, Poff, Mark A wrote:
> Andrew, thanks very much for that suggestion.  I modified the sub 
> record as follows and it seems to work:
> record("MY_PV_sub")
> {
> field(SDIS, "MY_PV.SEVR")
> field(DISV, "0")
> field(INPA, "MY_PV")
> field(SNAM, "MY_PV_subroutine")
> }
>  
> This prevents the subroutine from being called when the Severity is 
> zero
> - which is a significant improvement - but if the PV value bounces 
> around while in the HIHI range (for example), the subroutine is called 
> each time.  What I was hoping for is a setup which calls the 
> subroutine only once when an alarm state is entered, and isn't called 
> again until the alarm state is exited/re-entered.  For example, 
> perhaps the subroutine, when called, could examine the current 
> severity level and change the DISV value on the fly to a different 
> severity level...but that might be a pretty ugly way to do it.  Do you have any thoughts?

You should look at the HYST field of the longin record, which was added to prevent alarm chattering (it specifies a dead-band that the value must exceed before the record will exit back out of the current alarm level). Then try removing the foward link from MY_PV and make the SDIS link in MY_PV_sub a CP link, so it will process the MY_PV_sub record only when a monitor event occurs because the value of MY_PV.SEVR has changed.

I think a combination of those two should get you what you want.

- Andrew

--
Advertising may be described as the science of arresting the human intelligence long enough to get money from it. -- Stephen Leacock