Experimental Physics and
Industrial Control System

I completely agree with Andrew. I also did not want to start a new work here. We all have enough to do ;-)

It was/is a didactic problem which unfortunately makes it difficult for beginners to start with EPICS. And we 
should not disregard this problem.

I have a lecture and tutorial "EPICS for beginners" this semester. And we have shown the advantages of 
DRVL and DRVH in the context of analogout and longout to e.g. limit output voltages of PowerSupplies
(in the  IOC so that the operator cannot make a mistake). 

In the homework the students then limited an offset (which is physically fixed at the moment)
by DRVL = DRVH = value. They were then disappointed that it did not have the desired effect.

The description in the documentation was overread/ignored.

We have now solved this so that with PINI=yes the fixed value is set at 
iocInit and with DISP = 1 no caput works anymore.

Heinz 


> On 29. Jan 2021, at 17:29, Johnson, Andrew N. via Core-talk <core-talk at aps.anl.gov> wrote:
> 
> I’m not convinced that any changes are really needed. I don’t want us to add another menu field for this, and I see Heinz’s somewhat unusual problem as a corner case that we have other ways to solve without changing the code.
> 
> Heinz, can you tell us whether my solution using IVOV or Ralph’s meets your needs please? I think we need to hear from you to stop alternative ideas from proliferating.
> 
> Thanks,
> 
> - Andrew
> 
> 
>> On Jan 29, 2021, at 10:19 AM, Ralph Lange via Core-talk <core-talk at aps.anl.gov> wrote:
>> 
>> On Fri, 29 Jan 2021 at 15:57, Ben Franksen via Core-talk <core-talk at aps.anl.gov> wrote:
>> Am 28.01.21 um 14:20 schrieb Zimoch Dirk (PSI) via Core-talk:
>> > One could of course change the check to if ((prec->drvl !=0 || 
>> > prec->drvh !=0) && prec->drvh > prec->drvl) But then in ao as well.
>> > 
>> > It would change traditional behavior (which used to be documented 
>> > wrongly in earlier EPICS releases) and potentially break existing 
>> > IOCs where people use DLRH = DRVL != 0 and expect the limits to be 
>> > ignored. It would also introduce a new "special case" because the 
>> > record can then be fixed to any value except 0. Another option is to 
>> > change the defaults from 0,0 to -verybig,+verybig.
>> > 
>> > I think a poll would be needed.
>> 
>> -1
>> 
>> This makes a bad situation worse by complicating the rules. The problem
>> is the use of in-band exceptional values which is bad design and will
>> invariably hit you back sooner or later. Better fix: add a new field to
>> select whether to use or ignore DRVH/L.
>> 
>> To keep compatibility, that would have to be a menu AUTO/NO/YES with a default of AUTO.
> 
> -- 
> Complexity comes for free, simplicity you have to work for.
> 

Attachment: smime.p7s
Description: S/MIME cryptographic signature

References:

A general question about convert e.g. in longout record drvh, drvl handling Heinz Junkes via Core-talk

Re: A general question about convert e.g. in longout record drvh, drvl handling Ralph Lange via Core-talk

Re: A general question about convert e.g. in longout record drvh, drvl handling Heinz Junkes via Core-talk

AW: A general question about convert e.g. in longout record drvh, drvl handling Zimoch Dirk (PSI) via Core-talk

Re: A general question about convert e.g. in longout record drvh, drvl handling Ben Franksen via Core-talk

Re: A general question about convert e.g. in longout record drvh, drvl handling Ralph Lange via Core-talk

Re: A general question about convert e.g. in longout record drvh, drvl handling Johnson, Andrew N. via Core-talk

Navigate by Date:

Prev: Build failed: epics-base base-iocsh-auto-paths-614 AppVeyor via Core-talk

Next: Build completed: epics-base base-7.0-615 AppVeyor via Core-talk

Index: 2002  2003  2004  2005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  <2021>  2022  2023  2024 

Navigate by Thread:

Prev: Re: A general question about convert e.g. in longout record drvh, drvl handling Johnson, Andrew N. via Core-talk

Next: Unit tests on Windows DLL internals Ralph Lange via Core-talk

Index: 2002  2003  2004  2005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  <2021>  2022  2023  2024