Considerations for Hall D Slow Controls

[Elton Smith <elton@jlab.org>]

Meeting: Mon Oct 22 1:30
------------------------
Notes and comments by Elton. Feel free to send additional
comments/corrections/suggestions.

Attending: Chris Cuevas (Fast Electronics) , Elton Smith (Hall B,D), Arne
Freyberger (Hall B), Karen White (Accelerator), Mark Ito(Hall B), Dave
Doughty, Steve Wood (Hall C), Matt Bickley (Accelerator), Elliott Wolin
(DAQ, Hall D), Paul Brindza (Halls A,C,D)

Paul reviewed the status of the magnet. The original control system is
completely manual. The plan is the replace it with a modern set of
controls. Paul's preference is to have a single computer that controls all
aspects of the magnet (cryogenics, power supply, etc)

Note added by Steve: Paul suggested that it is nice to have the local
controller for a magnet be somewhat isolated so that the magnet controls
are not easily modifyable by someone with a little bit of EPICS knowledge.
A PLC naturally provides this isolation.  But we also have to make sure
that the PLC is modern and well supported so that changes can be made by
the qualified magnet experts.

Karen reviewed requirements by the accelerator
* the accelerator plans to use EPICS for the upgrade
- Future development of EPICS is to go to a system which is more a)
  hardware independent and b) uses cheaper hardware
- currently there are 3.5 flavors of EPICS at Jlab (one per
  hall+accelerator)
- The accelerator strongly suggests that EPICS be chosen for control.
- Minimally, we should have a system that can use EPICS as a data logger.
- The accelerator group is willing to support this effort at any required
  level, from taking on the entire software responsibility to giving
  various levels of support to others.

* Any system chosen should have a completely open exchange protocol
* One possible option is to have a PLC-based control system. The PLC can
  be addressed by EPICS or by another program such as LabView.

Example: the Control group installed EPICS at Michigan (which
accelerator?), gave training and support to their personnel. Total effort
was approximately two weeks and this allowed them to implement the system
successfully.

Main question about control systems other than EPICS is the one of
support.

EPICS is also being used extensively in any places including ANL, LANL,
BESSY, DESY, JLAB, PSI, Keck, SLAC.  JLAB might even be the largest
installation.

Note by Elliott: Many accelerators are NOT using EPICS on purpose, most
notably LHC...what does this tell us?  It would be good to get a list of
the controls systems used by all the major accelerator labs (LHC,
Fermilab, RHIC, BNL, SLAC, DESY, KEK, BEPC, etc.).

Are there any systems which EPICS cannot satisfy? (Probably not, but Mark
noted that an equally important question is how difficult is it to develop
an EPICS implementation).

For the Hall D slow control, do we need a common control system, or simply
a common interface? For example, can CDEV serve as the interface?

EPICS: there are 3 well-developed synoptic viewers for EPICS, none of
which run under CDEV.

For Hall D (also for current experiments), the main systems which need
algorithmic control are the magnet and target systems. Other components
are typically "set and forget". (Question: what about harp scans,
perhaps voltage scans?)

Options for PLC-based control systems. With a well-supported PLC
(accelerator is investigating various options, SNS is specifying its own
brand).


Possible strategy:

            PLC  -->  EPICS
             |
             |
             ^
          LabView for
          testing

Cheers, Elton.





-- 
Elton Smith
Jefferson Lab
elton@jlab.org
(757) 269-7625