The evolution of Jefferson Lab's control system Page: 3 of 5
This report is part of the collection entitled: Office of Scientific & Technical Information Technical Reports and was provided to Digital Library by the UNT Libraries Government Documents Department.
The following text was automatically extracted from the image on this page using optical character recognition software:
their commitment to collaborate and continue
development. The Unix based system utilized a toolkit
approach, but had no proven success running a large
control system. This option involved changing a limited
set of hardware and redeveloping device control
applications using EPICS tools. The system had many
existing device drivers, but no CAMAC drivers and no
existing high level applications, so these would also need
to be developed.
In the end, EPICS was selected for Jefferson Lab's
control system. Believed to provide the most modern
system of the three, sharing the TACL toolkit philosophy,
taking advantage of the best existing technology and
supporting multiple hardware interfaces all factored in
EPICS' favor. Additionally, the growing strength of the
fledgling collaboration was seen as an effective way to
get more and better tested software with fewer
programmers, reducing duplication of effort, and
benefiting all participating labs.
4 CONVERSION TO EPICS
The first step in the conversion process was to select a
portion of the existing control system, and attempt to
provide similar controls functionality for that portion .
This would serve as a prototype, providing the accelerator
controls staff with training and experience in EPICS, and
demonstrating the feasibility of converting the remainder
of the controls. The plan was to perform the control
system conversion while accelerator commissioning was
under way. At this point in time, roughly one third of the
beam line was already in use, and there was a strong
disincentive to halt the accelerator development schedule
for a prolonged period of control system rework.
The part of the accelerator selected for the prototype
was the superconducting RF system. One reason for
beginning with the RF system was its relative
independence from the rest of the control system, running
on dedicated front-end computers. The RF is a significant
component of the entire control system, representing
roughly one-half of all I/O points. Additionally, it is
highly repetitive, so once the complex controls were
developed for one cryomodule, the logic could be
replicated, substituting proper names, for the other 21
cryomodules. Demonstrating the replication capability
was key to showing TACL systems could be quickly
converted for EPICS. Success in this prototyping effort
would provide convincing evidence that the entire project
was feasible and help us develop a schedule for
A small group of software developers were selected to
participate in the RF prototype. They served as a
vanguard in learning about EPICS, immersing themselves
in the new control system with the idea that they would
mentor other developers as work continued to convert
other systems. In order to provide on the job training and
speed the RF prototype development work, Jefferson Lab
established a close relationship with the controls group at
LANL, where EPICS was first developed. LANL
provided on site support during the prototyping project.
Their EPICS expertise, immediate availability and insight
into controls development kept the project moving
forward. This proved to be very valuable assistance,
preventing potential roadblocks from impeding progress.
After 3 months of work, the prototype was complete, and
we were able to control the RF system using EPICS.
Having shown that EPICS was a reasonable option for
Jefferson Lab, we chose to make the conversion as
smooth and non-invasive as possible by providing a link
between the existing TACL controls and new EPICS
controls. By providing a live interface between the two
control systems, accelerator operators were able to
control hardware connected to EPICS through TACL
interfaces, and vice versa. This provided users the
opportunity to examine TACL and EPICS screens, side
by side, and verify proper operation. The integration
process was made straightforward by the fact that both
control systems use name-based management of process
variables. A single Unix process served as the locus of
communication between the two control systems, so that
all name translation and conversions could be made in
Once the RF system and the EPICS/TACL interface
were working well, we used the information we had
learned to plan for the conversion of the remaining
accelerator systems. The CHL was left for last because of
its more stringent requirements for high availability. A
failure of CHL control usually results in the loss of liquid
helium, so a few hours of down time can be quite costly.
The cryogenic system was small enough that TACL
performed well and reliability was very good. Given
these factors, the cryogenic personnel wished to see at
least a year of operational success with EPICS in the
accelerator before they would consider their own
conversion. EPICS continued to perform well during the
remainder of the accelerator conversion, and afterwards
the conversion of the cryogenics systems also went
smoothly. This completed the implementation of EPICS
for Accelerator Division systems. In the Physics Division,
the existing control applications were also ported to
EPICS, and new developments began using EPICS.
Using the same control system labwide has proven very
beneficial especially as we have been able to share
control system programmers across divisions as needed.
5 IMPROVEMENTS AND UPGRADES
As we began to work with EPICS, we added some IOC
features to support conversion of our existing systems.
After introducing EPICS for machine operations, we
began improvements on the console side of the system to
meet the needs of our customers.
In order to make EPICS work with our existing
CAMAC hardware; a driver was written according to the
Here’s what’s next.
This report can be searched. Note: Results may vary based on the legibility of text within the document.
Tools / Downloads
Get a copy of this page or view the extracted text.
Citing and Sharing
Basic information for referencing this web page. We also provide extended guidance on usage rights, references, copying or embedding.
Reference the current page of this Report.
White, K. S.; Bickley, M. & Watson, W. The evolution of Jefferson Lab's control system, report, October 1, 1999; Newport News, Virginia. (digital.library.unt.edu/ark:/67531/metadc702983/m1/3/: accessed November 20, 2018), University of North Texas Libraries, Digital Library, digital.library.unt.edu; crediting UNT Libraries Government Documents Department.