Upgrade of the ESRF Vacuum Control System

Abstract

The whole vacuum control system of the electron storage ring (SR) of the ESRF is in operation since more than ten years now. Apart from difficulties to have appropriate support for the old system, we start facing problems of aging and obsolescence. We have been reviewing our philosophy of data acquisition and remote control in order to upgrade our systems with state of the art technology, taking into account our operational experience. We have started installing shielded “intelligent” devices inside the SR and taking advantage of the latest developments linked to new technologies, such as OPC Server (Openness, Productivity, Connectivity), Webpage instrument control and more. This paper outlines our actual work dedicated for Programmable Logical Controller (PLC) applications.

Full text

Page 1

UPGRADE OF THE ESRF VACUUM CONTROL SYSTEM
D. Schmied, E. Burtin, P. Guerin, M. Hahn, R. Kersevan*, ESRF, France
Abstract
The whole vacuum control system of the electron
storage ring (SR) of the ESRF is in operation since more
than ten years now. Apart from difficulties to have
appropriate support for the old system, we start facing
problems of aging and obsolescence.
We have been reviewing our philosophy of data
acquisition and remote control in order to upgrade our
systems with state of the art technology, taking into
account our operational experience. We have started
installing shielded “intelligent” devices inside the SR and
taking advantage of the latest developments linked to new
technologies, such as OPC
Productivity, Connectivity), Webpage instrument control
and more.
This paper outlines our actual work dedicated for
Programmable Logical Controller (PLC) applications.
Server (Openness,
INTRODUCTION
Our initial aim was the exclusive use of off-the-shelf
vacuum equipment being totally integrated in our general
control system [1, 2]. This appeared to have many
advantages such as:
• share of human resources for development, failures
and maintenance tasks is possible
• flexibility in the case of upgrades
• common historical database for storage ring (SR)
and beam lines is possible
But we also faced control problems during shutdown
periods, due to modification and maintenance work of
different sub-systems. The control of vacuum equipment
and monitoring of bake out was sometimes impossible via
the standard control system.
Parts of the control system and vacuum instrumentation
start to show aging problems or getting obsolete with
more frequent failures or maintenance problems.
Triggered by the replacement of our entire network of FE-
computers from a diskless VME based real-time operation
system to industrial PC’s under LINUX, we easily
migrated most of our vacuum equipment while redefining
parts of our remote control.
The remote operation of all permanently installed SR
vacuum equipment concerns 200 inverted-magnetron
gauges, 70 Pirani gauges, 400 sputter ion pumps (SIPs),
60 residual-gas analysers (RGAs), 360 non-evaporable
getter (NEG) pumps and 2000 thermocouples (TCs).
Due to the hazardous environment, no intelligent board
or electronics has been installed in the SR tunnel, which
implied long cables and numerous electrical connections.
This turned out to be a bad choice in the case of TCs, as
well as the fact that our data acquisition system was
simultaneously dedicated for bake-out and vacuum
chamber monitoring.
Temperature measurements to monitor the vacuum
chambers’ temperature during operation requires a high
flexibility in terms of number of used channels and their
identification, and seems to be much more dynamic
compared to the requirements for our bake out
monitoring.
EVOLUTION OF THE ACTUAL SYSTEM
Initiated by the described weaknesses of our system,
but also because of new, future machine requirements, our
temperature monitoring system needs to be adapted.
The installation of in-vacuum undulators and NEG
coated vacuum chambers requires a much higher level of
control, monitoring and data storage, in order to meet the
ultra-high vacuum specifications of the installed
chambers, and the correct operation of the storage ring,
front-ends, and beam lines.
To contribute to the smooth operation of our SR, there
is a need for a reliable vacuum alarm system. Actually, we
have a “two-state” vacuum alarm which consists of the
state “ok” or “tripped” based on our interlock system.
Even if the status for each vacuum equipment is reported
to the general vacuum application, an overall evaluation is
missing which allows the operation crew to recognize
critical situations, or which helps the vacuum group to
anticipate failures. For that reason, we have started with
the implementation of more sophisticated alarms for our
pressure measurements and SIPs operation.
Our aim is to include also temperature measurements
and in some locations partial pressure measurements.
TEMPERATURE MONITORING SYSTEM
The actual temperature acquisition system consists of
~2000 thermocouples placed along the vacuum chamber
walls or inside the vacuum system for bake out and
machine diagnostics. Each of the 32 vacuum sections is
equipped with two connection boxes to transfer from TCs
to Cu-extension cables in order to connect them to the
temperature acquisition unit located outside the SR
tunnel.
The thermocouples we originally chose were too thin, and
due to radiation and bake out their external liner
(fiberglass material) became extremely brittle. The long
cable length and multiple electrical connections showed
to be a frequent source of faulty temperature indications
leading to troubled interpretations of their meaning.
Lack of flexibility and frequent changes, made it
extremely difficult to keep track and maintain a reliable
system since any in-situ changes implied the update of
resources and restart of server/client programs.

______________
*On sabbatical leave at the SNS/ORNL, Oak Ridge, USA
Proceedings of 2005 Particle Accelerator Conference, Knoxville, Tennessee
28570-7803-8859-3/05/$20.00 c ?2005 IEEE

Page 2

Figure 1: Temperature acquisition layout.
For
thermocouples,
maintained for bake out monitoring, but will no longer be
accessible via the general control system.
New thermocouples dedicated for machine diagnostics
will be installed. We choose twisted wires of kapton®
isolated thermocouples, which will be directly connected
to a small modular PLC unit located inside the SR tunnel.
We installed one of these PLC’s without shielding in
order to check its radiation hardness relative to the chosen
location. Measurements showed a total dose of 3.78Gy
before its failure. Comparative temperature measurements
with our standard measurement system (Figure 2) indicate
a slow degradation of the actual measurement signal but
no erroneous measurements or other problems related to
communication or failures.

these reasons,
including
the
Cu-extensions,
presently installed
will be

Figure 2: Temperature profile of an unshielded PLC.
With 3 mm of Pb-shielding, the dose rate could be
reduced to 20mGy/month without any degradation of the
measurement signals. This could be confirmed with a test
unit, installed since more than 2 years and without any
remarkable degradation.
In order to improve the experienced control problems
due to frequent changes, we transfer most of the
configuration and identification resources on the PLC
level. Simple server programs of our general control
system access the PLC Ethernet interface, via the TCP/IP
protocol for data acquisition. With the help of OPC
servers and a dedicated PC server, we are able to modify
and configure the actual system by ourselves, such as add,
remove or change names of thermocouples, without
support from the computing group. Once the modification
has been completed, the PLC dynamically updates the
user applications and historical database. This dynamic
upload of configuration changes should significantly
improve our reliability, since the mismatch of names,
channels within different applications was often a source
of problems in the interpretation of equipment and
machine operation and performance.
BAKEOUT MONITORING SYSTEM
With the aim to keep our vacuum control system in its
actual state, we prepared a fast plug-in layout for a mobile
bake out monitoring PLC. This concerns the survey of
total and partial pressure measurement, temperature
measurement, NEG activation, interlock and alarm
handling. Refer to the pressure monitoring; we will keep
the serial interface for the standard remote operation and
use in parallel the analogue signals. All analogue and
digital signals are pre-cabled and accessible in the
electronic racks of each vacuum section.


Figure 3: Layout of bake-out monitoring.
The monitoring of the bake out and control of the NEG
activation is available through dedicated graphical
applications via the embedded PLC Web server program.
The archiving of the various signals is done on a
dedicated PC via the TPC-IP interface and OPC servers.
CONCLUSIONS
The fact that we are faced to a large number of
analogue (2600) and digital I/O (2000) signals dictates the
use of high level PLC, with massive calculation power,
which presents a clear overkill considering our
requirements. This reflects not only an economical point
of view but also the fact that these systems, because of
their complex architecture and multiple interfaces, are
often subject to modifications. Their maintenance appears
similar to those of PCs regarding batches or software/
hardware changes.
Proceedings of 2005 Particle Accelerator Conference, Knoxville, Tennessee
0-7803-8859-3/05/$20.00 c ?2005 IEEE2858

Page 3

Figure 4: Monitoring and archiving of bake out data.

Since we need to split these signals due to cable length
and space problems, but also from an operational point of
view, our choice was to reduce the permanently installed
signals to a minimum, and therefore allow a structure
based on simple PLC units. More sophisticated PLCs are
used for our mobile PLC systems, which are considered
as stand-alone-units, independent of the ongoing activities
during shutdown periods. We take advantage of the Web
server technology, which we found easy to use and to
modify since there is no code to be written. The fact that
it is portable to various platforms appears to be of great
advantage, without the necessity for any control system.
The need for these improvements became evident during
commissioning phases where the control system was not
yet ready or different systems were used (SR, beam lines)
or systems were subjected to frequent changes.
The shielded PLC units installed in the SR tunnel for
temperature acquisition perform very well. Some of the
units are installed since 2 years without any degradation
or failures. The dynamical update of changes made to the
system and its mirroring into the historical database or
user interfaces works very well.


ACKNOWLEDGMENT
Entire Vacuum group crew, but also Computer Services
and Radioprotection in particular J.M. Chaize, F. Poncet,
P. Verdier, B. Vedder and P. Colomp.
REFERENCES
[1] D. Schmied, “ESRF Vacuum Control System”, 41st
IUVSTA Workshop, Brdo Pri Kranju, Slovenia, June
2004.
[2] D. Schmied, “Vacuum Control and Operation”, CAS
Vacuum Technology, Snekersten, Denmark, June
1999.

Figure 5: Web server based user interface.
Proceedings of 2005 Particle Accelerator Conference, Knoxville, Tennessee
2859 0-7803-8859-3/05/$20.00 c ?2005 IEEE