Open Abstracts

ID121: Enhanced Data Logging Capabilities in Vsystem

J. P. Girard, R. T. Westervelt, E. Zharkov and P. N.

Clout Vista Control Systems Inc.

The Vsystem logging engine and related utilities has been greatly enhanced to support a range of advanced features and also to support a wide variety of data archiving needs. The new data storage format of the Vsystem Vlogger will be described. How the new storage format provides support for new features such as circular logging, flexible data timestamps, and SQL access to the data without sacrificing archiving performance will be discussed. The new Vlog storage format supports many different ways to analyze the archived data. The flexible analysis tools built onto this structure that allows for structured queries of the data will be described in the context of several typical applications using Vlog. These applications include "black box" recording for post failure analysis, very high resolution recording, training systems and alarm recording. The new Vlog system has been integrated with the rest of Vsystem and ported to a wide variety of platforms, including Windows NT. Implementing Vlog on these platforms illustrates some of the innovative techniques used in developing Vlog. In particular, the integration between Vlog and Valarm, the alarm monitoring and recording portion of Vsystem, shows the flexibility of the new Vlog architecture. The new Vlog architecture provided the ability to vastly increase the functionality of Valarm to meet new requirements.

Submitted by: Dr. Robert T Westervelt
Full address: 134B Eastgate Dr. Los Alamos, NM 87544 USA
E-mail address: bob@vistanm.com
Fax number: (505) 662-3956
Keywords: Archiving, Vsystem; Alarm Recording

ID122: A Comparative Reliability Study of Computer Network Configurations for Large-Scale Accelerator Systems

E. W. Kamen, and S. S. Singh

School of Electrical and Computer Engineering Georgia Institute of Technology Atlanta, GA USA 30332-0250

The accelerator production of tritium (APT) facility which has been proposed is an extremely complex system consisting of four major subsystems: accelerator, target/blanket, tritium separation, and balance of plant. A major goal is to have all control modules and important system components in these four subsystems connected to a network through which information can be sent to a master control room for monitoring and controlling the APT facility. A key challenge is to achieve a very high degree of network reliability and network availability at a reasonable cost. In this paper we propose an ATM switching hub architecture for the APT control and information network which "sits" above the input- output controller (IOC) level. In addition to specifying a switching hub layout for the APT facility, we investigate the following two specific aspects of the proposed ATM framework: (1) The use of "LAN emulation" versus "native ATM" in realizing the interface between the ATM network level and the IOC level. The interface will be nontrivial if it is required that the IOC level be based on a broadcast shared-medium connectionless framework, as this is not compatible with ATM which is a connection-based fixed cell length framework. An interesting issue that will be pursued is the possibility of using a ATM network interface card for a VME bus architecture which is the current format for EPICS. (2) The development of a model-based approach for predicting the reliability of the proposed ATM switching hub structure. By reliability, we mean the probability that network communications can be maintained (at some specific bandwidth) in the presence of transmission link failures in the possible virtual paths between terminals in the ATM switches. In the last part of the paper, we will make some comparisons between the proposed ATM switching hub architecture, a standard Ethernet bus architecture, and FDDI. The comparisons will be based on bandwidth availability and reliability.

*This work was supported in part by ERDA Contract No. 96082 and managed by Westinghouse Savannah River Company.

Submitted by: E. W. Kamen
Full address: School of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA USA 30332-0250
E-mail address: kamen@ee.gatech.edu
Fax number (404) 894-4641
Key words: ATM switching hubs, APT information network, LAN emulation

ID123: A High Reliability Romping Control Sub-System

J. Caligm, W. Hunt

Indiana University Cyclotron Facility OUCF

This paper describes a portion of the control system developed at IUCF for the Cooler Injector Synchrotron (CIS). We describe the precision ramping sub-system. The sub-system is VME based and employs fiber optic data transmission for High noise rejection. The hardware includes several modules designed and manufactured at IUCF to meet specifications not attainable with commercial hardware. These modules feature active redundancy with automatic switch over for high reliability. 'Ihey also include built in test and self diagnosis with centralized failure monitoring as well as system health monitoring for rapid maintenance. The modules also exhibit very low drift and self calibration for maximum rentability. we describe the Precision ramping modules and a deep memory timing squencer.

Submitted by. J. Callahan
Postal Address:Physics,Dept. IUCF 2401 Milo B.Sampson Ln Bloomington Indiana/USA 47,159
E-mail Address: callahan@lucf.indiana.edu
Keywords: Ramping, DAC, ADC, High Reliability, Timing

ID124: A Control System for Accelerator Tuning Combining Adaptive Plan Executionwith Online Learning

W. Klein, M. Kroupa, C. Stern, and R. Westervelt G. Luger and E. Olsson

Vista Control Systems University of New Mexico

The Intelligent Controls Group at Vista Control Systems Inc. is developing a portable system for intelligent accelerator control. Our design is general in scope and is intended to be configurable to a wide range of accelerator facilities and control problems. We have so far tested this system at two sites: the Brookhaven National Laboratory ATF and the ATLAS facility at Argonne. The control system employs a multi-layer organization in which knowledge based decision making is used to dynamically configure a variety of lower level optimization and control algorithms. An object-oriented physical access layer (PAL) supported by the Vsystem control database allows intelligent reasoning and decision making at the higher levels to abstract from the lower level details of hardware manipulation, signal processing, and synchronization. A teleo-reactive (TR) interpreter (Nilsson 1994) is used to select and execute sequences of control actions necessary to achieve tuning goals. This TR architecture combines the sense-react looping of conventional control with intelligent planning, allowing adaptive opportunistic execution of a flexible control plan. To this framework we have added some initial learning capabalities as well as a capacity to dynamically replan in cases where the current control plan is ineffective. We describe recent experimental results at Brookhaven and Argonne, including the results of teleo-reactive control as well as experiments with online learning and replanning.

Submitted by: Carl Stern & Bill Klein Vista Control Systems Los Alamos, New Mexico USA: (505) 662-2484
Email address: stern@vistanm.com klein@vistanm.com
FAX Number: USA: (505) 662-3956
Keywords: automation, control, teleo-reactive, object-oriented, Vsystem.

ID125: The Control System for a New Proton Extraction Probe at TRIUMF

D. B. Morris, K. S. Lee

TRIUMF

A motor control system was designed for the new Beamline 2A Proton Extraction Probe of the 500 MeV Cyclotron at TRIUMF. This beamline is being built for injecting into TRIUMF's new ISAC facility. The probe uses four motors for positioning the extraction foil within the vacuum tank. The mechanics of the probe use a pantograph for lowering the extraction foil into the beam plane, which affects the radial position of the foil. An automatic correction of this shift was crucial for proper foil positioning. Requirements of the system included accurate positioning, integration into the existing central control system, an X-Windows Motif operator interface and the potential to replace existing systems in use at a future date.

Submitted by: David B. Morris
Full Address: 4004 Wesbrook Mall Vancouver B. C. Canada V6T 2A3
Email Address: cadfael@triumf.ca
Fax Number: 604-222-7307
Keywords: motor, extraction

ID126: LUMINOSITY OPTIMIZATION FEEDBACK IN THE SLC

L. Hendrickson, S. Bes, P. Grossberg, D. McCormick, N. Phinney, P. Raimondi, M. Ross

SLAC Stanford, California 94309 USA *

The luminosity optimization at the SLC has been limited by the precision with which one can measure the micron size beams at the Interaction Point. Ten independent tuning parameters must be adjusted. An automated application has been used to scan each parameter over a significant range and set the minimum beam size as measured with a beam-beam deflection scan. Measurement errors limited the accuracy of this procedure and degraded the resulting luminosity. A new luminosity optimization feedback system has been developed using novel dithering techniques to maximize the luminosity with respect to the 10 parameters, which are adjusted one at a time. Control devices are perturbed around nominal setpoints, while the averaged readout of a digitized luminosity monitor measurement is accumulated for each setting. Results are averaged over many pulses to achieve high precision and then fitted to determine the optimal setting. The dithering itself causes a small loss in luminosity, but the improved optimization is expected to significantly enhance the performance of the SLC. Commissioning results from summer of 1997 are reported.

* Work supported by the Department of Energy, contract DE-AC03-76SF00515

Author's name: Linda Hendrickson
Full address: Stanford Linear Accelerator Center 2575 Sand Hill Rd. MS 46 Menlo Park, CA 94025 USA
E-mail address: LJH@SLC.SLAC.STANFORD.EDU
Fax number: 1-415-926-3515
Telephone: 1-415-926-3913
Keywords: feedback

ID127: SLC FEEDBACK PERFORMANCE IMPROVEMENT AT HIGHER FREQUENCIES

L. Hendrickson,F.J. Decker,P. Grossberg,N.Phinney, P. Raimondi, M. Ross, H. Shoaee, J. Turner

SLAC, Stanford, California 94309 USA *

A beam-based fast feedback system has been used to stabilize orbits, energy and other beam parameters throughout the SLC since 1991. In early implementations, the measured frequency response in the Linac was less than optimal for 120 Hz operation, partly due to hardware limitations which prevented the feedbacks from operating at the full beam rate. In recent runs, hardware upgrades and a new operational strategy significantly improved the feedback system and reduced beam jitter. New performance results now achieve the ideal response required for a future linear collider. Additionally, two of the feedbacks were upgraded to damp oscillations at and near the Nyquist frequency of 60 Hz. Specialized correctors support separate control of the beam position on alternate pulses of the accelerator, and customized software implements independent control for even and odd tagged pulses. Performance measurements indicate that the feedback is able to damp beam noise at and around 60 Hz in addition to controlling low frequency oscillations.

* Work supported by the Department of Energy, contract DE-AC03-76SF00515

ID128: DSP Application to Parallel Processing in JT-60 Plasma Control

T. Kimura, K. Kurihara, Y. Kawamata and K. Akiba

Japan Atomic Energy Research Institute, Naka Fusion Research Establishment

Plasma shape control in a tokamak-type fusion device is one of the important issues for improving plasma confinement performance, increasing the efficiency of plasma-RF wave coupling and protecting in-vessel components like divertor tiles. Real-time reproduction of the full shape is indispensable for the shape control. A real-time plasma shape visualization system using a single RISC processor (MC88000, Motorola) was developed and applied to experiments in the JT-60 tokamak [1]. In this system, it took about 90 ms to obtain the plasma shape. In spite of such a slow reproduction, the behavior of plasma was well observed. In order to observe the plasma behavior more naturally using this system and apply the system to real-time plasma control, we have to reduce the time required for the execution time of the reproduction. Its main computational procedure is to calculate the outermost flux surface using the flux function and magnetic signals. Parallel processing is one of the measures to reduce it in this procedure. Then, a TMS320C40 DSP-based multiprocessing system has been developed for reducing the computational time. The DSP, which is housed in a VME module, has six so called "commports" for inter-processor communication and one VMEbus port. The optimum number of the DSPs was estimated according to the degree of parallelism in the above computational procedure. The compiler "Parallel C" (3L) is used for program development in this multi-processing platform, where parallel and pipe-lined processing of signal input, calculation of the plasma shape, and command output is executed. VMEbus-based reflective memory modules (VMIC) are used for high speed data transfer of the magnetic signals from the plasma feedback control system to this DSP system. This real-time shape visualization system will be applied to the JT-60 experiments from the coming June. The performance on the parallel processing will be reported at the conference. [1] Kurihara, K., Fusion Technology 22 (1992) 334.

Submitted by: Toyoaki KIMURA
Full address: JT-60 Facility Division I, Department of Fusion Facilities, Naka Fusion Research Establishment, Japan Atomic Energy Research Institute 801-1 Mukoyama, Naka-machi, Naka-gun, Ibaraki-ken 311-01, Japan
E-mail address: toyoaki@naka.jaeri.go.jp
Fax number: 81-29-270-7459
Keywords: JT-60, plasma control, DSP, parallel processing, visualization

ID129: Operation Experiences and Development of the Control System
in the SRRC

K. T. Hsu, K. K. Lin, Jenny Chen, C. J. Wang, J. S. Chen, C. H. Kuo, C. S. Chen, K. T. Pan, K. H. Hu, C. C. Kuo,

Richard Sah Synchrotron Radiation Research Center No. 1 R&D Road VI, Hsinchu Science-Based Industrial Park, Hsinchu 300, Taiwan

The control system of the SRRC can be divided into console level and device level. Console level computer consists of workstations and PCs. Device level are VME crates based embedded system. Several PC based system are also supported on devices level for special applications. The operating system on console level are VMS currently, DEC UNIX and WindowsNT support will coming soon. Devices level support pSOS+, LynxOS, DOS/Windows/WindowsNT operating system. Most of the software components are in-house development in the past. Integrated commercial software package with control system are current effort. Sharing software with accelerator control community are in progress. Applied web for control related application is under way. On the development on VME crate system, PowerPC based VME CPU module runing LynxOS have been ported and compatible with existed system. Topology of the control network have been changed to switched ethernet and support fast as well as standard ethernet. Sophisticated electronic instruments are connected to the control system via IEEE-488 to ethernet adapter. Project to integrate the control of turn-key injector with main control system have been initiated recently. Achieved a seamless integration with limited resources are highly desirable. The operation experience and development of the SRRC control system will present at the conference.

Submitted by : Kuo-Tung Hsu
Full address : No. 1 R&D Road VI Hsinchu Science-Based Industrial Park Hsinchu 300, Taiwan
E-mail address : kuotung@srrc09.srrc.gov.tw
Fax number : + 886 3 5783892
Keywords : Control, Networking

ID130: Timing System and Injection Control in the SRRC

K. T. Hsu, K. K. Lin, K. W. Hu, Jenny Chen J. S. Chen, C. H. Kuo, C. J. Wang, G. J. Jan*

Synchrotron Radiation Research Center No. 1 R&D Road VI, Hsinchu Science-Based Industrial Park, Hsinchu 300, Taiwan * Department of the Electrical Engineering and Institute of Electro-Optics National Taiwan University, Taipei 107, Taiwan

The accelerator system of SRRC have been operated nearly three years. The are good as well as bad in present timing system. The timing system are updated recently to improve its performance and satisfied newly requirements. The SRRC consists of three accelerators. The timing system provide trigger signal to synchronise the injection sequence of a 50 MeV linear accelerator, 1.3 GeV booster synchrotron and the pulse magnet of the 1.3-1.5 GeV storage ring. The timing system also provide trigger signal for diagnostics devices of the accelerator system. Trigger signal with picosecond jitter are also provide for streak camera trigger. Highly low jitter timing are also provide for mode-locked laser system synchronization the system will be used for the synchrotron-laser hybrid experiment in near future. Bunch synchronization signal for time-resolved experiments are also provided. Several injection control applications are available for the machine operation. Details of the timing system will present at the conference.

ID131: An Accelerator Control System Archive using Windows NT and the Web

Winfried Schütte

Deutsches Elektronen-Synchrotron DESY

The data archive for the PETRA type Accelerator Control Systems holds about ten million numeric and a hundred thousand string data each. It is defined by a Microsoft Access 2.0® database. (See previous conference "The PETRA Archive System") The data are stored via a Visual Basic 4 program in a local Microsoft SQL Server 6.5® database. These contain only the most recent (many weeks) data. Within typically 30 seconds the data of all those primary archive databases are collected at the central archive server. Here the old data are deleted selectively. So beam currents with a second interval are only kept for ten days, other data is kept indefinitely. All the archived data can be accessed from within DESY by a web browser. The basic architecture, service programs and data table layout will be described. Going from Windows 3.1 to NT4.0 was a surprisingly easy exercise.

Submitted by: Winfried Schütte
Full address: Deutsches Elektronen-Synchrotron DESY MKI Notkestr. 85 D 22603 Hambur Germany
E-mail address: schuette@desy.de
Fax number: +49/40/8998-4303
Keywords: database, archive, distributed system, integrating commercial products

ID132: Towards a Data Quality for Supervision Data in the CERN's Technical Control Room

R. Martini, P. Ninin

CERN, Geneva, Switzerland

The implementation of the Technical Data Server (TDS), has highlighted the relianceof many of the software products used for the supervision of the technical infrastructure at CERN, on the quality of the data used to represent the different processes being monitored. This data, which we call the static data, describes the equipment used at all levels of the control system to process the real-time information (dynamic data). Static data covers information such as equipment address, alarm related information, monitoring processes, responsible persons etc. All the modules that make up the TDS, as well as many of its client applications, are driven by this static data which is managed by a relational database composed of around 80 tables holding over 500 columns. It is essential for the proper functioning of the supervision software, and hence for the quality of service provided to those involved in the monitoring of the technical infrastructure, that the data used is up to date, correct, meaningful, accessible, and easily maintainable. The life cycle of the key data elements used in these applications is examined with a view to the formulation of a data quality policy. This policy covers the assignment of responsibility for maintenance of the data, the definition of naming standards and the definition and description of processes that are necessary to ensure that the information available in the Technical Control Room is of the highest quality.

Submitted by: R. Martini, P. Ninin
Full address: CERN, Geneva, Switzerland
E-mail:
Keywords: Supervision, Software Engineering, Database, Data Quality

ID133: Industrial Control Systems for the Technical Infrastructure at CERN

P. Ninin, E. Cennini, P. Sollander

CERN, Geneva, Switzerland

Industrial control has been used for several years for the control of CERN's technical infrastructure, particularly for fire and gas detection, access control and cooling water systems.Until recently, industrial equipment has been connected to the control network at the lowest level via CERN specific front ends. Since 1993 they have been integrated using drivers of industrial supervision systems. Today the integration is performed using TCP/IP based protocols and the distributed Technical Data Server (TDS) supervision system.This paper summarises the evolution of the different techniques, the network architectures and our integration philosophy for the safety system of the Chorus and Nomad experiment, the West Zone and the PS cooling water systems. These systems are based on Siemens Programmable Logic Controllers (PLC) using for data integration into the control room layer the Siemens proprietary fieldbus Sinec L2, the protocol Sinec H1, in combination with the driver from the supervisory system Factory-Link. Also the Access Control and Machine Interlock Safety systems of the SPS accelerator has been implemented using PLCs connected to Sinec L2 fieldbuses and the H1 protocol transported on a specific Ethernet network. However this system is totally supervised by a Factory-Link software application.The re-engineering of the systems for SPS air conditioning, gas and fire detection will use the multi-protocol technology and the new concept of industrial control system integration of the TDS. Rather than using a specific Ethernet segment, the Sinec H1 protocol will be transported on the general control network up to an equipment controller. This process converts and integrates PLC specific data frame to the TDS, using an asynchronous data exchange protocol based on TCP/IP.

Submitted by: P. Ninin, E. Cennini, P. Sollander
Full address: CERN, Geneva, Switzerland
E-mail address:

ID134: A DSP-based Control System for the ISAC Pre-Buncher

M. Laverty, K. Fong, and S. Fang

TRIUMF 4004 Wesbrook Mall, Vancouver, B.C., Canada V6T 2A3

A DSP-based control system has been designed for use in the ISAC pre-buncher. This system uses three harmonics of the 11.667 Mhz fundamental to synthesize a sawtooth waveform for driving the pre-buncher cavity. A total of three DSPs are used, with a single 24-bit DSP providing both in-phase and quadrature control of each harmonic. The complete system including low-level RF components is housed in a VXI rack. The supervisory control consists of a Windows-based server, which broadcasts system status using datagrams, and listens on control commands via TCP. JAVA-based control clients, running either as standalone applications or as WWW applets, interpret these messages to provide control and display of the system operating parameters.

Submitted by: M. Laverty
Address: TRIUMF 4004 Wesbrook Mall Vancouver, B.C., Canada V6T 2A3
E-mail: mapl@triumf.ca
Fax: 604-222-1074
Keywords: DSP control VXI TCP Java

ID135: Orbit Feedback Development In SRRC

C. H. Kuo, K. T. Hsu, Jenny Chen, K. K. Lin , C. S. Chen, R. C. Sah

Synchrotron Radiation Research Center No.1 R&D Road VI, Hsinchu Science-Based Industrial Park, Hsinchu, Taiwan,

A global and local orbit feedback system project have been launched in the Synchrotron Radiation Research Center (SRRC). In this paper, we will discuss the suppressant results of orbit feedback experiments with any vibrations and orbit drift that lead to distortions in the closed orbit and result in a larger effective emittance. Together with the brightness reduction, beam motion induced incident light position and angle varying can degrade the advantages of using synchrotron light. Insertion devices are essential to produce high brilliance synchrotron radiation, however it influences the electron orbit and the lattice of storage ring. Feedback system is used to eliminate these undesirable effects. From control points of view, orbit feedback is an typical multiple input multiple output (MIMO) problems. Technical, it is difficult to implement an analog matrix operation consisting of large amount of beam position monitors ( BPMs ) and correctors. On the other hand, the flexibility of system development is considered. Consequently, digital processing was used here to implement feedback system.

Submitted by: C. H. Kuo
Full address: No.1, R&D Road VI, Hsinchu Science-Based Industrial Park 30077, Taiwan,
E-mail address: ako@alphID1.srrc.gov.tw
Fax number: 886-3-5783892
Keywords: orbit, feedback, control

ID136: Downsizing Databases to Commodity Hardware

D. Finstrom, T. Zingelman

FNAL

Commercial relational databases offer many desirable features to a control system, but their use on traditional hardware (mainframes or large UNIX servers) can be cost prohibitive. Fermilab is taking advantage of this technology while utilizing commodity hardware and shrink-wrapped software. This allows exploitation of relational databases with significant cost savings and an increase in performance over our current configuration. This paper will discuss the capability of these systems, the associated cost-savings, and the application advantages they offer.

Submitted by: D. Finstrom
Full address: Fermilab P. O. Box 500 Batavia, IL 60510 U. S. A.
E-mail address: dfinstrom@fnal.gov
Fax number: (630) 840-4688
Keywords: Commodity Computing, PC, Database

ID137: Sharable GUI Objects for the Operators' Consol

S.Dasgupta* and Isamu Abe

*Variable Energy Cyclotron Centre, 1/AF Bidhan Nagar, Calcutta 700 064, India KEK, 1-1 OHO, Tsukuba, Ibaraki 305, Japan

There are a reasonable number of components in different kinds of accelerators that have very similar operator interfaces; examples being magnets, RF generators, vacuum pumps, beam handling equipments etc. Sofar as testings, maintenance, and some parts of operations of such equipments are concerned, the look and feel of these can be uniformised. In other words the Man Machine Interface of any accelerator can be developed in a way to be a globally usabledevelopment.The obvious solution in such cases have to be based on object- oriented development approach. The above devices can each be made an individual class. The variety and diversity of such an equipment across accelerator families must be accommodated in its class- structure by the way of properties of the class. That can ensure that the developments produce codes usable elsewhere. Any new user should be able to set the properties with the quantities unique to the particular installation and generate an unique MMI for that accelerator. An attempt is made here to prepare such an object-oriented software meant to be used in the control consoles using PC's on NT / Windows95. Visual J++ language is used for object-oriented programming. MMI object generations are done by Java tools. Also the communication softwares meant for monitoring and controlling of lower- level devices will be attempted to be linked to this software.

Submitted by : Isamu Abe
Full address : KEK Accelerator Lab. 1-1 OHO, Tsukuba,Ibaraki 305, Japan
E-mail address : isamu.ABE@kek.jp
Fax number: (81)-298-64-7438
Keywords : GUI objects, software sharing, MMI, NT console

ID138: A Distributed Object-Oriented Telescope Control System based on RT-CORBA and ATM

J. M. Filgueira

Control Group -GTC Project Instituto de Astrofísica de Canarias

The GTC project is in charge of the construction of an optical-infrared 10 meter class telescope at the ORM observatory in Canary Islands. The control system of the GTC will be responsible for the management and operation of the telescope, including its instrumentation.The continuous and rapid development of the technologies related to hardware, software and communications has permitted a greater complexity in control systems. In their turn, the new techniques of active and adaptive optics, the new methods of optimisation of useful observing time and the programmes of continuous evolution to ensure competitive telescopes capable of assimilating future technological advances, present new challenges in the design of control systems.The life-cycle of the control system of the GTC will be subject to a continuous flux of changes brought about by different factors (evolution of the requirements, continuous development of new instruments, correction of faults, etc.). These factors must be taken on board with the minimum impact possible on the availability of the GTC once it enters into operation. This will only be possible through the selection and planning of an adequate technological framework that enables these changes to be assimilated throughout the entire life-cycle of the telescope. In order to accomplish that, the adoption of open 'de facto' standards like ATM, CORBA and POSIX becomes a key point. In the past the technologies needed in order to achieve the required network bandwidth pushed the adoption of heterogeneous network architectures based in different technologies: FDDI, Ethernet, etc. This resulted in complex and non scalable solutions with fixed QoS (Quality of Service). Recently ATM allows the unification of all network communication requirements into one network technology, this results in more flexible, scalable, and easy to maintain networks while. Additionally, ATM provides important advantages like high bandwidth, end-to-end QoS and virtual connections. This makes ATM an appealing solution for real-time LAN based control systems. CORBA is a distributed object computing middleware standard. It supports the development of flexible and reusable distributed services and applications providing independence of hardware platform, network technology, operating systems and programming languages. Current researches on RT CORBA (e.g. real-time event services and end-to-end QoS guarantees) can provide an effective architecture capable of supporting current control systems while maintaining their real-time behaviour. Recently several vendors have recognised this issue porting their ORB to real-time POSIX compliant OS platforms. The combination of RT-CORBA, ATM and POSIX OS can provide a technology framework capable of supporting the functionality required and guarantying the evolution of control systems while preserving investments.

Submitted by: José M. Filgueira
Full address: GTC Project - Instituto de Astrofísica de Canarias 38200 - La Laguna Canary Islands - SPAIN
E-mail Address: jmfilgue@iac.es
Fax number: +34-22-315 032
Keywords: CORBA, ATM, distributed object-oriented systems

ID139: Control system of HIMAC secondary beam course for medical use

S. Kouda, A. Kitagawa, J. Yoshizawa, M. Torikoshi, M. Kanazawa, T. Murakami, K. Noda and M. Suda

National Institute of Radiological Sciences (NIRS) JAPAN

Y. Ishikawa

Accelerator Engineering Corporation JAPAN

J. Matsuura, H. Okamura and T. Usami

Mitsubishi Electric Corporation

A new secondary beam course is under construction at HIMAC (Heavy Ion Medical Accelerator in Chiba) in NIRS. A primary purpose of the beam course is to provide radioactive beams, such as 11C, 19Ne, etc., for a study of heavy ion therapy. The secondary beams are generated by the primary beam through the projectile fragmentation at the production target, then the wanted radioactive beam is separated by an energy degrader and two analyzing targets. The control system calculates the beam transport parameters for the primary and secondary beams, and set the parameters to all device. In addition, in order to identify the secondary particle, the control system acquires and analyzes the data of the particle energy loss (delta E) and time-of-flight (TOF). For medical application, the following conditions are required; a good reproducibility of beam qualities, reduced beam tuning time, and easy operation for routine duties. To satisfy those requirements, the control operations described above sentences are integrated in one framework. The control system consists of three computers of a data acquisition (SBC-DAU), a device control (HEBT-SCU) and a analysis and sequence control (SBC-GCU) because of distributing load of the system. The SBC-DAU is a DSP module on VMEbus and takes data from monitoring detectors of the delta E and the TOF. The HEBT-SCU is a minicomputer, which controls magnets, slits, profile monitors in the beam course via a remote I/O (RIO) system in HIMAC. According to a programed sequence, the SBC-GCU (UNIX work station) takes and analyses data from the SBC-DAU via Ethernet and calculate control parameter of the devices such as magnet currents and controls the devices using the HEBT-SCU via Ethernet.

Submitted by: S. Kouda
Full address: Division of Accelerator Physics and Engineering, National Institute of Radiological Sciences, 9-1, Anagawa 4-chome, Inage-ku, Chiba-shi 263 JAPAN
E-mail address: kouda@nirs.go.jp
Fax number: +81-43-251-1840
Keywords: secondary beam, medical, integrated

ID140: The GNU control system at CAMD

B. Craft and P. Jines

Center for Advanced Microstructures and Devices, Louisiana State University

The Center for Advanced Microstructures and Devices at Louisiana State University operates a 1.5 GeV, synchrotron radiation dedicated, electron storage ring. The injector is a 180 MeV linear accelerator. A new control system utilizing Intel Pentium processors and the Linux operating system is being developed. The software being developed will be released under the GNU license. The original procurement of the entire accelerator system, including the control systems, utilized a single, fixed-price contract. The control system for the storage ring was based upon a commercial control system software package running on a DEC MicroVax and a DEC VaxStation. The interface hardware included CAMAC, GPIB, and RS-232. The control system for the injector used third-party software running on top of OS9 in a VME crate controller. The interface hardware was home-brew TTL driven by a VME bus-adapter card. This system has been in operation since the fall of 1992. Efforts to interface the two control systems to each other were unsuccessful. Though OS9 provided basic TCP/IP capabilities, the third-party control software consumed essentially 100% of the available cpu. Pinging the linac control system resulted in time-outs of watch-dogs which in turn lead to shutdown of the linac. The control system for the storage ring used an ASCII interface and its performance was limitted by the MicroVax. Taking advantage of hardware and software development over the last five years, the new control system will provide significantly increased performance and a far superior man-machine interface. Integration of the control functions necessary for the operation of the storage ring and the linac will allow automation of many routine tasks. Significant budgetary savings will be realized since neither hardware nor software maintenance will be necessary. The development effort will take less than two man-years to replace the functionality of the original control systems.

submitted by: Benjamin Craft, III
mail address: LSU-CAMD 3990 W. Lakeshore Dr. Baton Rouge, LA 70803 USA
E-mail: bcraft@unix1.sncc.lsu.edu
fax: 011-504-925-7078
suggested session: Inexpensive (PC based) Systems