Control Systems for Accelerators, Experiments and their Subsystems

Hardware and Software Technologies

Development and Management Tools

Graphic User Interfaces

Integrating Different Platforms or Commercial Systems

Networking and Connectivity

Databases

Integrating Web Technologies

Future Trends, Emerging Technologies and New PC Paradigms

Control System for the New Linac LUE200 in JINR, Dubna

O. V. Strekalovsky, A.S. Kayukov, M.M. Korjovkina, V. A. Shvets, A.P. Sumbaev - JINR, Dubna, Russia

LUE200 - the 200 MeV electron linac for the pulsed neutron source "IREN" is under development in FrankLab of JINR, Dubna. The main systems of the linac, such as the general timing control, the high-power RF system, the control system of klystrons, the power supply system of magnets, the vacuum subsystem and the cooling system should be integrated into the control system of LUE200. The integration will be based upon the SCADA software. The personal computers will be applied as control devices and PLCs will be used at a lower level.

V.I. Zaitsev - State Research Center of Russia Troitsk Institute for Innovation and Fusion Research, Russia

The operation algorithm of inertial fusion facilities is presented. It is showed that the control of inertial fusion facilities (high current z-pinches, lasers, heavy ions accelerators) have two very different time stage: the technological preparation with typical time about hours and the energy transport to the load and the processes into the load (time less 10-7 sec). Usually this part of operation algorithm determines two different parts of control complex: the technological preparation systems and the fast date acquisition systems. The PC as an element of control system can be used in the both parts. According to facility features the technological systems can have differences. Because this stage is not strong time limited the cheap interfaces like CAMAC under PC control can be suit. For technological parameters support feedback is used widely. Technological systems are conservative part of control and not changed during facility life. The time of second operation stage is very short. Therefore during this stage the feedback operation is not real practically and the data acquisition is main function of control at this stage. PC doesn't take part in work during this process. Recorded information is stored in the memory of analog or digital converters. Configuration of this part of control must be changeable according to experimental requirements. The functions of PC are the preparation of data acquisition systems, the data transfer, processing and monitoring of information. Processing algorithms are realized in the time between operation cycles. The supervisor system is united the both parts in to control complex. The examples of inertial fusion facility control systems are given.

Gongfa Liu, Jingyi Li, Weimin Li, Chuan Li, Lei Shang - National Synchrotron Radiation Lab. (NSRL), P.R.China

The injection control system of Hefei Light Source (HLS) storage ring is a subsystem of the upgraded HLS control system, which is based upon EPICS. Three programmable logic controllers (PLCs) are used as device controllers, which control one septum modulator and four kicker modulators of HLS storage ring. An Industrial PC is used as Input/Output Controller (IOC) and it connects the 3 PLCs with serial communication (RS232 mode) over fiber. A PC with Linux is used as operator interface (OPI), operator application are runnung on it. The control system was completed in july 2000, the run results demostrate that the control system is reliable and easy oprational.

Weimin Li, Jingyi Li, Gongfa Liu, Chuan Li - National Synchrotron Radiation Lab. (NSRL), P.R.China

HLS (Hefei Light Source) at NSRL (National Synchrotron Radiation Lab) consists of three parts, 200MeV linac, transport line and 800MeV storage ring. The control system is based on the popular EPICS (Experimental Physics and Industrial Control system). In this system, PC-based hardware is widely used for front-end controllers, Input/Output Controllers (IOC) and Operator Interfaces (OPI). A number of Programmable Logic Controllers (PLC) and SUN workstations are also used in the control system. This paper will cover the experience of using PC-based hardware under EPICS, the field bus used in HLS control system, data archiving, communication and data sharing, and some analysis tools for physics and operation use.

Jinhui Chen - Institute of High Energy Physics, P.R.China

RF cavities cooling water temperature control system is a very important subsystem in BEPC. It is a typical control system based on PC, including temperature probes, signal processing unit, personal computer, electric valve. Classical PID control strategy is adopted. The user interface and control program are developed in Visual Basic 6.0. At the end of this paper, the problems of the present subsystem arising during upgrade debugging and operation are discussed.

Valerie Gautard and Olivier Cloue - CEA-Saclay, Gif-sur-Yvette, France

ATLAS is a particle detector which will be built in the CERN at Geneva. The muon detection system is made up among other things, of 600 chambers measuring 2 to 6 m2 and 30 cm of thick. The chambers'position must to be known with an accuracy of +/- 30 micrometer for translations and +/-100 microradian for rotations in a range of +/- 5mm for translations and +/-5mrad for rotations. To do that, we have designed an optical sensor based on image analysis of a coded mask. Because the accuracy is very high and the number of sensors is significant (over 1000), we had to develop a bench. This bench has to be able to acquire some physical signals like position, temperature and images, to drive a controller used for the displacements of the stepping motors, to analyse, calculate and store the calibration parameters in a data base. To do that, we have decided that the control system use a PC, Windows 98 and Visual C++. This paper will describe the software of the whole bench.

S.P. Karasyov, R.I. Pomatsalyuk, I.N. Shlyakhov, A.Eh. Tenishev, V.L. Uvarov, NSC Kharkov Institute of Physics and Technology, Ukraine

Possibility of easy development of the PC based control system with the utilization of the printer LPT port is demonstrated. Its standard is supported by any IBM compatible PC and does not demand additional interface creation. Upon utilization of digital and analogous hardware with consecutive data acquisition such channel provides data exchange rate up to 120 kbod, that is sufficient for a control of not too large systems (e.g., technological accelerator). At the same time a transition to bi-directional bus according to the modern LPT standard increases exchange rate by more than the order of magnitude. Proposed approach is applied for elaboration of subsystems for the radiation field monitoring and control. The drivers of subsystems are written with Turbo-Pascal language to simplify a program preparation and to keep the operation rate of the accelerator control system.

A. Bogomyagkov, S. Karnaev, I. Nikolaev, E. Simonov, S. Smirnov, V. Zorin - Budker Institute of Nuclear Physics, Novosibirsk, Russia

The control system for the VEPP-4 facility was designed more than fifteen years ago and based on the home-developed CAMAC-embedded minicomputers Odrenok. Five years ago all computers were connected via Ethernet network [2]. This step allowed us to integrate PC into the VEPP-4 control system. The paper reviews a new opportunity for data acquisition and handling in the PC computers. Data acquisition system allowed us to study carefully different features of the VEPP-4M collider [2]. Data visualization tools based on standard software and on CERN ROOT framework.

Xiangcheng Kong, Qi Le - Institute of High Energy Physics, P.R.China

Since the reconstruction of power supply control system for quadrupole and focus magnets of BEPC injector was finally accomplished, the new generation of control system has completely substituted the old injector control system (DOS-PC). The foundation of this new control system is based upon CAN Bus Remote I/O (partially upon RS-485 Remote I/O), and the control console is based on Windows-PC environment. This article will introduce the components and main functions of the new control system, and then focus on the communication interface between this system and physical beam-tuning program, revealing the open characteristics of this system and the convenience of hardware extensions.

Gong Guanghua - Dept. of Engineering Physics, Tsinghua Univ.; Li Yuxiong - National Synchrotron Radiation Lab, Univ. of Science and Technology, P.R.China

A prototype Beam Loss Monitor system for the NSRL (National Synchrotron Radiation Laboratory) beam transportation line has been successfully tested. By detecting the charged particles produced when the beam hits the vacuum chamber wall, the system can diagnose the beam position and direction in the strong-pulsed radiation field, whose duty cycle is nearly 10-7, around the transport line. This system is composed of tens of front-end nodes and a console. Each node uses 4 ion chambers originally designed for the Tsinghua Tongfang container inspection system. A high-sensitive I/F converter is used as an amplifier, which can measure the low level current in the scope of 10-12A and can reach a dynamic arrange to 5 decades. The pulse number is proportional to the dose rate of the beam loss radiation field. The nodes are installed along the 88-meter transportation line. The average pulse rate of different nodes shows the position of the beam, and the difference among the 4 ion chambers shows the direction of the beam. The influences from the running of injector linac, from the insertion of FLAG and from the adjustment of driving coils have been observed and will be described in the paper. A 32bits micro-controller 68332 from Motorola is used in the node to collect the pulse queue, store the count number, do some simple calculations, show the count information on local LCD, and communicate with the console. It also provides a RS232 port that can connect to a terminal, which is useful for debugging and updating. The nodes are connected via a CAN bus, and a Personal Computer with a CAN adapter card is also connected into this bus to serve as the console. The console requests all the nodes for the count number and illustrates them in diagrams or curves. The console analyzes the data to show the whole orbit of beam in the transport line and finds the position where the beam loss exceeds a pre-defined level. If desired the BLM information can be sent to the control room through the Ethernet link of the console. These results are very useful to help operators to adjust the machine parameters. After several months' running, some new ideas to improve this system have been proposed, which will also be presented and discussed in this paper.

R. Apiwatwaja, G. Isoyama*, S. Rugmai, S. Rujirawat, N. Sanguansak T. Ishii and W. Pairsuwan. - Siam Photon Laboratory, Thailand; * ISIR, Osaka University, 8-1 Mihogaoka, Ibaraki-shi, Osaka 567-0047, Japan

A PC-PLC control system is used in the Siam Photon Source, which is the national synchrotron light source in Thailand. The machine was successfully filled with the electron beam in December 2001. During the course of daily machine operation, we find some inconvenience or bugs in the originally built control system and improve the system. The magnet control is found to be greatly beneficial from this newly design system. Part of the timing system was replaced with standard NIM modules and has yet to be fully integrated with the control system. An extra PC was installed to act as interface between PLC and older systems such as Slit, Pattern Memory, Digital Multi-meter and Video Switcher. Software polling technique was used to initiate a request to the devices. Problem of the slit control was solved by the modification of the communication technique on this intermediary PC. Another PC is handling data acquisition. Vacuum pressure, magnet power supply current is recorded every 10 minutes. Information on the system status is provided to the outside network through firewall. Data are being delivered in XML format to an external web server for greater system security. The old vacuum control unit is not fully integrated with the new control system. Only one fault signal was coming out of the old vacuum control unit but we ignore it. The development to integrate to the control system is underway. At the moment the control of storage ring vacuum system has to be assisted manually.

Takeshi Nakatani, Naoyasu Hosoda, Takemasa Masuda, Toru Fukui, Hitoshi Tanaka and Ryotaro Tanaka - SPring-8/JASRI, Japan

We introduced an off-the-shelf system for a simultaneous correlation measurement between electron beam orbits and insertion device operations in SPring-8 storage ring. The PC based measurement system, WE7000, supports many modules such as digitizer, function generator, thermometer, digital I/O and so on. The PC is running with Linux OS (RedHat 6.2), and WE7000 API libraries are prepared to attach any UNIX based control system easily. We have developed application software to fit the SPring-8 standard control framework. Using this framework, it is possible to control the measurement system remotely, and we can collect measured data to database via Ethernet. We report here the synchronizing measurement between beam position monitors (BPM) and operation effect of the insertion device as an example.

P. Chevtsov, S. Higgins, S. Schaffner, D. Seidman - Jefferson Lab, Newport News, USA

Several DirectNet Programmable Logic Controllers (PLCs) have been integrated into the central control system at Jefferson Lab. The integration is based on new software that consists of three main parts: a PLC driver with a state machine control block, a device support module, and a common serial driver. The components of new software and experience gained with the use of this software for beam dump systems at Jefferson Lab are presented.

Yu.B. Semenov, O.V. Strekalovsky, V.E. Zhuchko - JINR, Russia

Ten CAMAC crates housing the digitizing front-end electronics of the gas-detector part and the trigger logics, and one FASTBUS crate used for the photomultiplier read-out of the scintillator shell and the forward array, earlier were connected with a VME EUROCOM-6 computer by means of the parallel VSB Differential Bus Extension. Now we have replaced EUROCOM-6 by personal computer. So a special VSB to PC board was created and is used at the spectrometer FOBOS.

Yulan Yu, Yonggang Li, Mingchun Zhan, Liping Yang, Shuming Tang - BEPC Control Group,Accelerator Central, IHEP, China

The old vacuum control system of the BEPC storage ring was built in 1987, a VAX4500 computer with the CAMAC hardware controls the vacuum equipment on the storage ring. The current vacuum control system of the BEPC storage ring was upgraded in 2001. It is based on PLC architecture. One PC is mounted as a host. The commercial software FIX32 is used as software development tool on local control console PC and used to acquire the data of the vacuum equipment. The datum local PC machine is sent to the BEPC server via control LAN.

Rethfeldt Christoph and Busse Winfried - Hahn-Meitner-Institute Berlin, Germany

The control system of the Ion-Beam-Laboratory (ISL) dates from the late 1970's and was originally designed for the HMI-VICKSI accelerator complex. It is based on a single server processor (PDP11), a CAMAC fieldbus and software which was developed in-house.The strategy of the control system upgrade had to take into account the limited man-power available and the fact that the accelerator facility should continue to run. Our choice is the use of commercially available components, software and hardware, wherever possible. The new control system will be based on Vsystem by VISTA Inc. Common transparent APIs in combination with real time CAMAC-servers, like rtVAX's and Linux-PCs, will allow application programming on multiple OS-platforms in a heterogeneous network. In the first upgrade phase GUIs for operators or machine physicists were implemented along with Linux Application Servers, collecting and archiving machine data to generate 24h trend-charts. These charts are accessible via the Internet or by WAP-phones. Alarm messages can be sent out when requested.

K. Abrahamyan, I. Chachkhunashvili - DESY Zeuthen, Germany

At many facilities in DESY, including PITZ (Photo Injector Test facility at Zeuthen) or TTF (Tesla Test Facility at Hamburg), lots of VME crates are used in control systems. Sometimes it is necessary, to switch crates 'on' or 'off', to generate system resets or to check the current status of crate like temperatures, states, power channels voltages and currents, etc.. Normally the operator has to do it directly from control panel of every VME crate, but it is not convenient or simply impossible because the crate is in restricted access area. Therefore it is necessary to have remote access to control all VME crates uninterruptible and simultaneously. The connection between VME crates and the control system is done via CAN-Bus (on one CAN-Bus it is possible to put up to 127 VME crates). As a PC platform was chosen Win32 working with a PCAN-PCI controller, because the solution was cheep and easy to implement and to use on this platform. As Data Transfer Protocol between the control server and clients 'TINE' was chosen, as the widely adopted protocol at DESY. A Win32 based client application provides all functions to control VME crates. The most important functions are the following: reading the status of a crate; switching power 'on' or 'off'; changing fans speed; monitoring temperatures, voltages, currents and errors.

Andreas Luedeke - PSI, Switzerland

The Swiss Light Source (SLS) started user operation in August 2001. The standard operation mode is a frequent reinjection into the ring while users are taking data, the so called top-up operation. This keeps the beamcurrent stable to about a 2 per mille level. A set of dedicated application was developed to meet the special requirements of this operation mode. They cover the control of the injection and operational aspects like the early detection of degradations in the performance.

H.Katagiri, I.Abe, K.Nigorikawa, J.Kishiro and S.Kurokawa - High Energy Accelerator Research Organization (KEK), Japan

The microwave monitor system, based on VXI/HP-UX, has been installed in the KEK electron/positron injector linac. HP-VEE was employed in order to develop application program of the monitor system. We are considering reinforcement of the monitor system to achieve upgrades of the linac for super KEKB project. WE7000 was nominated substitute of the VXI-based DAQ system. The ActiveX-control for WE7000, which is provided by YOKOGAWA, has made possible software development with HP-VEE. Introduction of COACK for the monitor system is being studied in order to combine both DAQ systems.

A.Shirakawa, T.Kosuge, K.Nigorikawa, J.Kishiro and I.Abe - High Energey Accelerator Research Organization (KEK), Japan

Slow-positron project is one of the applications of electron linear accelerator (Linac). It comprises primary electron Linac, metal target to create positron, and beam lines for user experiments. We have made the control system for the primary Linac using COACK, which has started its operation. Although the scale of the machine is rather small, it contains many types of the accelerator components, and needs various control programs for each component. The COACK provides the cost/time saving for coding the control programs. The virtual machine concept in the COACK makes it very efficient to develop some components for the remote control which is not currently implemented in the accelerator. Quite stable and high-grade operations are expected in our project.

Hayashi Hiroshi, Narihara Kazumichi - NIFS (National Institute for Fusion Science), Japan

A high-speed data acquisition system based on PCs was developed for the LHD Thomson scattering diagnostic. It consists of sixteen FASTBUS ADC modules (1881M ADC, 64-channel inputs, 13 bit), a FASTBUS segment manager (1821 SM/I), a fast data logger, and two PCs. One PC controls the segment manager, which in turn controls 16 ADC modules. AD-converted data are transferred to the other PC (Windows NT). But in the burst mode, which is used for high repetition scattering measurements, data transfer speed exceeds the writing speed of hard disk. In order to overcome this difficulty, we newly developed a data logger with a 4GB buffer RAM in which a large amount of data is temporarily stored. Thus, the data logging with the speed up to 20MB/sec became possible

D. Bolkhovityanov, I. Pivovarov, O. Tokarev - BINP, Novosibirsk, Russia

Initially the VEPP-5 control system consisted of a number of unconnected subsystems, each one having a different design. Lots of fieldbuses were used, including homemade ones. In the last two years the control system was almost completely reengineered. Upper level consists of Linux PCs, which are connected via Ethernet. Low level consists of CAMAC hardware. The main fieldbus now is Ethernet, and CANBUS is begin being used. The control system software is network-transparent, and it complies the 'standard model' [1]. In 2001 we began using Linux-based CAMAC controllers, which made the control system software much more unified and easy to maintain.
[1] A. Goetz et al. Experience with a standard model'91 based control system at the ESRF. Proc. ICALEPCS 93 (Berlin). NIM 352A (1994) 22-27.

P. Paisuwan, R. Apiwatwaja, *G. Isoyama, S. Rugmai, S. Rujirawat, N. Sanguansak and T. Ishii - Siam Photon Laboratory, Thailand; *Alsoat ISIR, Osaka University, 8-1 Mihogaoka, Ibaraki-shiOsaka 567-0047, Japan

The commissioning of an accelerator complex is related with the operation of its machine control system directly or indirectly. The Siam Photon Source is the second-hand machine transferred from the SORTEC Laboratory, although some parts, particularly in the storage ring, have been appreciably modified. Thus, the problems inherent in the second-hand machine exist and they must be overcome for successful commissioning. In this report such issues are described. A few examples are described below. (1) The most serious problem we encountered was the broken timing system. First an obvious symptom is found. The battery is dead after a long period of the storage of the system in a package without being used. Then the disappearance of the memories installed there was found. Further, we found a computer installed in the synchronous pulse generator did not work. Eventually, we found the logic in a part of electronic system had fatal errors. This sort of trouble shooting proceeded step by step and was quite time consuming. Finally we decided to abandon the timer system of the SORTEC ring. (2) We found another mistake in the circuit design of the tuner control in the RF power supply in the storage ring. A circuit in which the negative feedback should be supplied was running with the positive feedback.This results in the undesirable vibration of the tuner. (3) When the synchrotron was operated, the 5th order harmonics generated is reflected into the primary line. This did not only affected various control systems in the machine but gave bad influence on the outside facilities. The cause was found. The filter did not work because of too large current of the 5th order harmonics due to the resonance. The work for the cure is underway. (4) A serious problem is the heterogeneous floor settlement. This causes large COD resulting in a short lifetime of the stored beam. The work for cure is underway. (5) Various other problems encountered will be reported.

H. Kleines, J. Sarkadi, K. Zwoll, F. Rathmann, H. Seyfarth, M. Mikirtytchiants, M. Nekipelov - FZ Juelich; A.Vassiliev, P. Kravtsov - PNPI

For the ANKE experiment at the medium-energy ring COSY at Forschungszentrum Jülich, a hydrogen/deuterium internal gas target, the Atomic Beam Source (ABS), was built. The total amount of process signals of the heterogeneous front-end equipment of this target is about 700. The interlock system has been implemented with Siemens S7 PLCs. A PC with Windows NT serves as controlling computer. The process control software was implemented on this PC on the base of the SCADA system WinCC from Siemens. Basic design decisions and structure of the control system will be presented. After a discussion of the WinCC features, implementation issues and experiences will be reported.

G. Cox, M.T. Heron, B.G. Martlew - CLRC, Daresbury Laboratory, UK

The SRS is a 2GeV synchrotron light source with 14 beamlines serving approximately 34 experimental stations. Control of the major elements of the beamlines (vacuum pumps, gauges, valves and radiation stops) is the responsibility of the main SRS Control System. As part of the long-term upgrade plan for the SRS Control System, the existing system has been replaced with a PC based client/server architecture. A parameter caching system was developed to manage server loading and maintain the SRS Control System's response time. The cache system was implemented using ActiveX/COM objects written in Visual C++ and Visual Basic. This paper describes the major components of the system and the techniques used during development.

Stefan Weisse, Ilia Chachkhunashvili, Gunter Trowitzsch - DESY Zeuthen, Germany

PITZ (Photo Injector Test Facility Zeuthen) is a test facility at DESY Zeuthen for research and development on laser driven electron sources for Free Electron Lasers (FEL) and linear colliders. The optimization of an electron gun is only possible based on an extended diagnostic system including a Video-system. The goal is to measure the electron beam position and the profile of the beam at different places and by different diagnostic tools along the beam line. The poster describes the requirements and the chosen implementation. Based on commodity PC hardware the system offers the complete remote control of the video cameras and the grabber hardware and provides the full video information for all the different client applications. The TINE protocol is used to control the system from different types of user interfaces. The video information (768x574 pixels) is transferred by direct TCP socket connections and is loss-less compressed with a ratio of at least 2:1. The client API is available for Win32, Solaris and ROOT.

O. Heber 1, S. Ghelberg 2, Y. Ben-Dov 3, D. Berkovits 3, C. Bordeanu 1, G. Gitliz 1, M. Hass 1, M. Paul 2, Y. Shahar 1 and G. Verri 2 - 1 Dept. of Particle Physics, Weizmann Institute of Science, Rehovot, Israel 76100; 2 Racah Institute of Physics, Hebrew University, Jerusalem, Israel 91904, 3 Soreq NRC, Yavne, Israel

The Accelerator Lab at the Weizmann Institute includes the 14UD Pelletron accelerator, the 3 MV Van de Graaff accelerator and the Molecular Physics Electrostatic Ion Traps Lab. All the three facilities are controlled with Group3 Control system hardware connected via several fiber optic loops to three PC computers. The major recent development was to separate between the users and the hardware by converting the system to server client setup. Each loop control PC was converted to a server. Clients can operate each system via the local network without the overhead caused by the communication with hardware. Each server is managing security, databases and alarms. The software was upgraded to LabVIEW 6.1 with the DSC module. In this module a tag represents each real hardware point and the server client communications are done with these tags. A special script language was written using LabVIEW to perform routine measurements for Accelerator Mass Spectrometry experiments. In this script language a sequence of automatic control commands is performed. The script includes control commands such as: Faraday cups in/out, accelerator optic setup and change between different ions, measurements, scans, communication with Data Acquisition computer, calculations and more.

Wolfgang Mexner, Karl-Heinz Cerff, Randolf Butzbach - Synchrotron Radiation Research Group, Forschungszentrum Karlsruhe GmbH, Germany

For the synchrotron radiation source ANKA, Karlsruhe/Germany 9 beamlines have been constructed simultaneusly. This allowed for the unique possibility to design a standardized beamline control system for all 9 beamlines. The solution is based on a VME crate with the realtime operating system OS/9 running on a PowerPC or 68K system as a server. The I/O signals and the stepper motors are controlled by the VME crate via CAN-Fieldbus. The different beamline components like slits, DCM, mirrors, vacuum controller etc. were developed as independent standardized nodes for the OS/9 system using the PC based Integrated Process Control Development System "Gamma" from RST, Munic/Germany. For the automatic experiment control as well as for the GUI for manual control, all functions can be called over the LAN via a DLL from multiple PCs simultaneously.

James Patrick - Fermilab, USA

This paper describes plans for future development of the Fermilab accelerator control system. The middle and high level portions of the system have for many years been based on VAX/VMS, a custom graphics library, and a custom communications protocol Acnet. Migration of these portions to a system based on PCs and other commodity processors, coded primarily in the Java language, is in progress. Plans and status of this migration are discussed.

G. Chiozzi, B. Gustafsson, B. Jeram - ESO; J. Dovc, M. Kadunc, I. Kriznar, M. Plesko, A. Pucelj, M. Sekoranja, G. Tkacik, I. Verstovsek, D. Vitas - IJS and Cosylab, Slovenia

The ACS is a CORBA-based control system framework with all features expected from a modern control system. It has been recently installed at the ANKA light source in Karlsruhe, Germany and is being used to develop the ALMA control system. ALMA is a joint project between astronomical organisations in Europe, USA and Japan and will consist of 64 12-meter sub-millimetre radio telescopes. ACS provides a powerful XML-based configuration database, synchronous and asynchronous communication, configurable monitors and alarms that automatically reconnect after a server crash, run-time name/location resolution, archiving, error system and logging system. Furthermore, ACS has built-in management, which allows centralized control over processes with commands such as start/stop/reload, send message, disconnect client, etc. and is fine-grained to the level of single devices. ACS comes with all necessary generic GUI applications and tools for management, display of logs and alarms and a generic object explorer, which discovers all CORBA objects, their attributes and commands at run-time and allows the user to invoke any command. A Visual configuration database editor is under development. An XML/XSLT generator creates an Abeans plug for each controlled object, giving access to all Abeans applications such as snapshot, table, GUI panels, and allowing to use the CosyBeans GUI components for creating Java applications. For those that write their own control system, ACS allows to define own types of controlled data and own models od communication, yet use powerful support libraries as long as one aadheres to some rules in the form of programming patterns. ACS uses several standard CORBA services such as notification service, naming service, interface repository and implementation repository. ACS hides all details of the underlying mechanisms, which use many complex features of CORBA, queuing, asynchronous communication, thread pooling, life-cycle management, etc. Written in C++ and using the free ORB TAO, which is based on the operating system abstraction platform ACE, ACS has been ported to Windows, Linux, Solaris and VxWorks. The applications are written in Java and run on any JVM-enabled platform. ACS is based on the experience accumulated with similar projects in the astronomical and particle accelerator communities, reusing and extending concepts and components of implementation. Although designed for ALMA, ACS has the potential for being reused in other new control systems, as proven by the nearly seamless installation at ANKA this spring.

Ya. Shashel, M. Silkin, N.G. Styervoyedov - Kharkov National University, Ukraina

One of the tasks of the electrons accelerator control is the maximal transfer of power of HF-supply sources to accelerated beam. Its decision consists in creation of optimum conditions of resonant exchange of energy between electromagnetic field and accelerated particles beam. It is necessary exactly to adjust and to support the necessary phase of accelerating HF-wave related accelerated bunches for it. Such extreme regulator of phase has been developed. The information parameter for automatic control of the phase shifter is the acoustic signal, which is excited as a result of interaction of HF-field with absorbing load.

Yoshinobu Shibasaki, Masakatsu Mutoh, Masashi Nanao, Katsuhiro Shinto, Isamu Abe*, and Hiroyuki Hama - Laboratory of Nuclear Science, Graduate School of Phisics, Tohoku University, Japan; *KEK, High Energy Accelerator Research Organization, Japan

A control system for a 300MeV-electron linear accelerator at the Laboratory of Nuclear Science (LNS), Tohoku University has been renewed. The new system consists of PCs running on Windows2000 and PLCs as IO Controllers (IOCs). COACK is employed to be management software to control the accelerator, and a database handling function is added so as to enhance the flexibility. Software of IOC and user-interface are composed of programs developed in the Visual Basic and commercially available components (ActiveX Components). Construction of the basic framework has been completed, and up to now the connections of devices more than 50% of the whole accelerator system has been replaced to new one from the old control. We are going to introduce the new control system.

Karl-Heinz Cerff, Wolfgang Mexner, Ralph Steininger - Synchrotron Radiation Research Group Forschungszentrum Karlsruhe Gmbh, Germany

For up to 32 beamlines of the synchrotron radiation source ANKA, Karlsruhe/Germany, a modular Beamline Personnel Safety System (BPSS) was planed and is erected simultaneously now for 9 beamlines. The design criteria are:
secure and reliable operation by use of a divert redundant (three processor) PLC controller from the manufacturer PILZ,Stuttgart/Germany with the capability of self-diagnose, also monitoring the state of the peripheral safety-hardware.
A standardized modular and by legal authorities certified(!) soft- and hardware to manage the safety needs of complex beamlines, means beamline branching or multiple beamline radiation hutches and shutters, further to manage the failure safe interaction with the beamline control system and the machine Interlock under injection, beamline maintenance and erection, normal operation, and alarm conditions.
Certification of all soft- and hardware components under the European Standard EN 954-1 risk category 4 makes the safety comissioning of a beamline fast and easy.

Kay Rehlich, DESY Hamburg, Germany

The next generation of large accelerators facilities will be produced and operated by an international collaboration. To meet this challenge the Global Accelerator Network concept was developed to make the best use of the world-wide competence and resources. The idea is to design, build, maintain and operate the facility by a distributed organization of many institutes located in different countries. Adequate methods and tools are needed to support the information, communication and controls tasks over the life-time of such a project.
At the TESLA Test Facility we started the remote operation of the accelerator. The requirements, implementations and the experience together with an outlook will be discussed.

L. Catani, A. Cianchi INFN-Roma2, Roma, Italy; G. Di Pirro INFN-LNF, Frascati, Italy

Staring from beginning of year 2003, the TESLA Test Facility accelerator (TTF) will be upgraded to TTF2 with final energy of 1 GeV and a SASE-FEL facility. Most of its components will be upgraded as well because of the increased size of the machine and as consequence of new requirements in terms of features and operability. The Optical Diagnostic System will consist of 25 (mainly OTR) stations along the linac for beam profile, emittance and energy/energy spread measurements. Its acquisition and control system (ODCS) will also be redesigned to be a beam images server for local and remote operation and measurements, as well as for documentation and data sharing system. It will be based on CompactPCI/PXI, Window 2000 Server, and developed with LabVIEW(tm) taking advantage of its features like, powerful image acquisition, processing and analysis library (IMAQ), multi-platform compatibility, integrated networking for remote operations and data sharing. The VI Server mechanism will be used to access to the system functionalities either from "local" control room consoles (Unix) or from remote computer (Unix/Mac/PC) A web interface will also be available for data publishing and restricted operations. During the presentation the design and features of the new Optical Diagnostic Control System for TTF2 will be discussed.

Elisabetta Pace - INFN Frascati, Italy

The DAFNE-Light Facility (Frascati, Italy) uses the synchrotron radiation light emitted at DAFNE storage ring. Presently, two beam lines are under commissioning: an X-ray beam line (spited in two different beam lines, DRX1 And DRX2) and an infrared beam line(DIR. The control system is based on FieldPoints and LabVIEW by National Instruments. In this article, the solution adopted and that is currently working is described.

Stefano Piano - Finuda Collaboration - INFN - Sez. di Trieste, Italy

A parallel data acquisition system based on VME has been designed for FINUDA. FINUDA is an experiment on hypernuclear physics at DAFNE, the Frascati Phi-factory. The acquisition software runs on embedded CES RIO 8062 CPU's, which use the Lynx OS operating system. Events from the different layers of FINUDA are collected through a differential PVIC bus by a run-control program, which runs on a bi-processor INTEL PIII computer based on the Linux OS. A DAQ system for the silicon vertex detector (I/OSIM) has been developed, this system requires that a large number of channels are suppressed by an efficient and fast algorithm. The algorithm also considers the coherent oscillations of the I/OSIM readout channels. The data are initially skimmed and, then, sent to a run-control program through an optical fiber PVIC bus via DMA.

L. David & the Ganil control group - GANIL, France

For almost ten years, the Ganil control system has been based on VMS workstations and Camac crates running Vaxeln on RtVax controllers, with Ada as a common language. When Digital Equipment (now Compaq) gave up with RtVax processors, we decided to move to Vme crates with PowerPC controllers running VxWorks. By the way, we also wanted to try some use of Linux, to get rid of the links remaining with Vms in the beam tuning programs and to be able to use some free software tools that were not very powerful in a Vms/Motif environment.
This paper describes the milestones we are underway to perform :
- graphical user interfaces using GTK instead of Motif
- Replacement of commercial XRT widgets by Gtk+
- Data Base access with SQL requests, programmed in Ada
- TCP/IP communication with VxWorks real time crates
- Common use of development tools with Linux

Reinhard Bacher, DESY, Germany

The Global Accelerator Network (GAN) is a proposed model to design, build, operate and maintain a future large accelerator by partners of an international collaboration. This novel approach has many legal, sociological and technical issues. This paper will describe the general ideas of the GAN model and report on first results of a workshop series recently established to form a world-wide community and to discuss the various aspects.

Distributed Control Systems with Virtual Engineering Stations and Controllers as Integrated Network Nodes with Single IP-protocol

Vinogradov V.I. - Institute for Nuclear Research Russian Academy of Science, Russia

Integrated modern control systems require distributed computer nodes, compact embedded cores and High-performance network protocols for interaction between distributed nodes. A lot of Fieldbuses today have effective trends to single IP-protocol networks. There are two basic general problems in development distributed microcomputer systems with flexible topologies for DAQ and Control applications. The first is effective (price/performance) Architecture on he base of week interactions of microprocessor core and modern compact Processor board. The second is high-performance systems on the base on strong interactions between processor cores. The first approach is more effective for Control systems, the second for high-performance DAQ applications. Week and strong interactions between processor cores in modular RT-systems with effective distributed Architecture for different applications can be used both for most reliable effective systems.Distributed systems with modular hardware and software for DAQ and Control should be consist of embedded microprocessor based front-end compact Terminal Systems (TS) for Real-time tasks integrated with Virtual Engineering Stations (VES) by tradition networks (Ethernet 1/100/1000) with weak interaction between Processor Cores (message passing) or by System Oriented Networks (SON) with strong interaction between Processor Cores (direct access to distributed node memory). System Architecture with week and strong interaction between processor cores in modular RT-systems with effective distributed processor nodes for complex topologies detectors are proposed and discussed The best platform for the first embedded applications is new compact single board microcomputers with Ethernet 10/100/1000 on the board. Two parallel networks should be good platform for reliable high-speed network-based virtual Control Systems. The second approach based on Distributed memory multiprocessor architecture is for high-performance data-flow acquisition and processing can be used with weak interaction network for installation, management, monitoring and remote access. One of the advanced platform for instrumentations in both applications should be compact MB, cPCI and PXI for this modular systems. General problems and proposals on main of this Approaches are discussed in this paper. Integrated web technologies as basic approach for future access interface to collected data in distributed nodes are proposed and discussed.

XueTao, Shao Beibei - Department of Engineering Physics,Tsinghua University, P.R.China

A system based on the embedded micro processor ColdFire 5307 has been well developed for BEPC-II in Beijing. The system includes 16M Bytes SDRAM, 2M Bytes Flash ROM and also a 10BaseT Ethernet interface. The free software GNU gcc, gdb is used as compiler and debug tools. And the embedded free OS uClinux is ported to our system; for the real time applications, the RTOS uCOS-II is ported successfully and the TCP/IP stack based on uCOS-II has been also developed. This system is useful for the sub system control of the accelerator of BEPC-II, and is also useful for the data acquisitions of other physical experiments. As an example, the bus access ability of our system was tested for a VME system. The I/O read and write capacity is much better than the commercial VME control board based on the PowerPC micro controller. The host of this system is PC, the data taken from the system will be sent to the host PC through the Ethernet. Now the net interface is 10Mbps , and a 100Mbps Ethernet interface will be developed.

Boriskin V.N., Ayzatsky N.I., Gurin V.A., Kushnir V.A., Mitrochenko V.V., Reshetnyak N.G., Romasko V.P., Savchenko A.N., Stepin D.L., Zakutin V.V., Zhiglo V.F. - Scientific-Research Complex ; Accelerator; National Science Center; Kharkov Institute of Physics and Technology Kharkov, Ukraine

Crossed-Field Secondary Emission electron sources have been investigated in the "Accelerator" R&D Production Establishment since 1993 [1]. The main purpose of these investigations is the study of performances of such sources and the search of possible application fields. Experimental set-ups developed for investigations at the "Accelerator" R&D Production Establishment comprise: vacuum chamber with an assembly of secondary emission metallic cathodes under investigation, pulse high-voltage generator, assembly of magnetic field generation and measuring system. In the choice on the methods of measurement, it has been taken into account that the object to be studied has the following characteristics: - the output beam of the gun is a set of parallel electron annular beams with d from 5 to 8 mm (8 beams at the first stage) arranged on the diameter D = 44 mm (the number of beams and their mutual arrangement may vary);
- total pulsed current of all the beams about 100 A;
- current pulse duration; from 1 to 5's;
- pulse-repetition frequency; from 1 to 12.5 Hz;
- electron energy; up to 50 keV;
- the electron beam source and the collector are in the constant magnetic field from 0.1 to 0.35 T.
To carry out above-described scientific-research developments we have designed and manufactured a 12-channel information measuring system. A hardware complex of the multichannel pulse-signal recording system is built up from a personal computer, crate CAMAC with modules, an interface unit and a synchronization system. The software of IMS operates in the Windows environment. It was written with the use of a visual object-oriented programming system C++Builder 5.
REFERENCES
1. Ajzatsky N.I., Boriskin V.N., Dovbnya A.N. et al. Electron beam production in multicathode secondary-emission systems. Proc. EPAC'02 (Paris, June 2002).

Lu Ping, Sun Baogen, Li Jingyi, Liu Jianhong, Wang Junhua - NSRL, University of Science and Technology of China, Hefei, China

The VXI-Based DAQ system for DCCT, COD, and photo beam position monitor of NSRL is described in the paper. The VXI controller, the VXI DAQ module and the conditioner are from National Instrument corporation, the Labview drivers are supplied with the modules, it is very convenient to control them by Labview program. The control system of NSRL is based on EPICS, we implement the DAQ system as an EPICS IOC by using the Channel Access VIs developed by K.U.Kasemir, and it works well.

F. Bal, S.Gonzalve, A. Guirao, S. Jolly - CERN, Switzerland

The FLIC card (FLexible I/o Card) is a general purpose PCI card with programmable logic, allowing to implement specific readout controllers with a large on-card buffer. Mezzanine cards are used on the FLIC for signal or protocal adaptations to external systems. These mezzanines may either comply with the PMC standard or with the CERN SLink convention, or they may implement any user-specific logic. The first applications were PC-based data acquisition systems for old CAMAC and VME electronics, the application discussed here is a high speed test system for pixel chips, where Slink is used to acquire data from the 40 MHz chip test station and store it at high rate into the local SDRAM buffer. From where it can be accessed at PCI bus speed by Labview based pixel analysis programs. Both the SDRAM controller and the SLINK readout controller are implemented in a Lattice Field-Programmable System Chip (FPSC). The SDRAM is mapped via the PCI BAR mechanism to the system memory and can hence be read out as such either under Linux, using the mmap() functions, or under Windows, using a specially developed DDL.

V. T. Lazurik, V. M. Lazurik, G. F. Popov, Yu. V. Rogov - Kharkov National University, Kharkiv, Ukraine

Authors have integrated the software ModeRTL (Modeling of Radiation-Technological Line) into the control system of the RTL incorporating an industrial linear electron accelerator with a scanner of electron beam and a conveyor. An operation of the control system is based on monitoring of electron beam and radiation facilities parameters, calculations of the absorbed dose distribution in an irradiated materials, and elaboration of control signals on value of deviation between calculated and calibration data for the specific radiation-technological process. The control system uses the thermoacoustic emission, generated by the pulsed electron beam in thin rods and plates, as a source of primary information about the current status of characteristics of each electron pulse of scanning electron beam. The main function of the software ModeRTL in control system are the following: processing of the acoustic response amplitude from radiation-acoustic dosimeter and restoration of electron intensity and a spatial profile of radiation field on the surface of irradiated materials; calculation of the absorbed dose, charge and temperature distribution in an irradiated materials; comparison of calculated absorbed dose and calibration data that are stored in data base; preparation and verification of control commands for radiation facilities. Due to fast analytical methods, the software ModeRTL execute all functions in real-time regime. The general architecture, the main components, and the mathematical aspects of the software ModeRTL will be discussed in the paper.

Anze Zupanc, Igor Kriznar, Mark Plesko, Urban Simoncic - Jozef Stefan Institute, Slovenia

In my article I present integration of new orbit correction method with frequency into Orbit Correction application. The Orbit Correction has grown from simple application to a platform for different orbit manipulation tools in form of plug-ins. It is daily used at synchrotron light source ANKA, Karlsruhe. Basic orbit operations are implemented so far, such as closed orbit correction, local bumps. These plug-ins in form of OrbitCorrectionModel-s, are using DataBush thus making calculation correction with semi-real-time control system properties. Using DataBush library gives programmer intuitive access to devices and elements relevant to machine physics, while access to the control system is provided with Abeans. Modular construction of Orbit Correction offers physicists an opportunity to easily write their own orbit manipulation methods with minimum concern about control system details.

Hideki Ogawa, Tomoyuki Inoue, Kiwamu Tsukada, Mitsuhiro Yokota, Mamoru Kojima, Jun-ichi Kodaira, Akimitsu Nishizawa, Takeshi Ido, Yoshiaki Kawasumi, Shinji Kato, Kozo Yamazaki - National Institute for Fusion Science, Japan; Isamu Abe, Takashi Kosuge, Kazuyuki Nigorikawa - KEK, High Energy Accelerator Research Organization, Japan; Masakatsu Mutoh, Yoshinobu Shibasaki - Laboratory of Nuclear Science, Tohoku Univ., Japan

The PC based control system has newly developed for Heavy Ion Beam Probe(HIBP) by using COACK(Component Oriented Advanced Control Kernel). The Plasma diagnostics by HIBP is an important diagnostics for the local measurement of the electric field and for the study of plasma turbulence of the density and the electric field. HIBP consists of an Au+ beam source, a tandem type 6Mev accelerator, and a detector. We have introduced COACK for the upper layer of the control system and developed operation console, historical operation record logging and device I/O functions. On the other hand, lower layer consists of PLCs, Function-Generators, and CAMAC. The total control points is over 400. As a result of introducing COACK, we could complete the development of the HIBP control system for exceedingly short period with low cost, and now the system is running stably. The detail and next target of the development of the HIBP control system is presented in this paper.

D. Bolkhovityanov, R. Gromov, BINP, Novosibirsk, Russia

In late 90s a need to replace aging "Odrenok" crate controllers in VEPP-5 control system has become apparent. In 2001 BINP electronics department designed a Motorola mc5200-based controller which has onboard 100M Ethernet and runs uClinux (Linux version for microcontrollers w/o MMU). In the course of a year VEPP-5 team has tested this controller and integrated it into control system infrastructure. This work had shown both pros of new controller (mainly due to unification of control system environment) and its cons (CPU architecture differs from that of host PCs, uClinux lacks some important features of Linux).

D. Bolkhovityanov, O. Tokarev, BINP, Novosibirsk, Russia

CANBUS is used in VEPP-5 control system for various types of slow controls, such as power supplies management. Finding a PC CANBUS controller was a complicated task, since most of these cards lack a properly supported Linux driver. So, for some time inhouse-made CAMAC-based controllers were used. Currently these are being replaced with new PCI controllers. This greatly simplified the software and there are plans to use CANBUS in more areas. VEPP-5 experience with these hardware played a key role in other BINP labs' decisions to use it.

Philip Duval - DESY, Hamburg Germany; Vladimir Yarygin - IHEP Protvino, Russia

One of the main goals of the control system engineer is to provide functional, intuitive, easy-to-use-and-maintain application programs to the operators, machine physicists, and hardware engineers. One way to do this is to avoid writing code entirely and systematically generate applications via databases and configuration files, where servers are database driven and clients are wide-interface or widget driven. The resulting application tend to be primitive, to extent that client applications can do little more than get data and display it, or issue commands. That is, it is difficult if not impossible in such schemes for widget A to react to an event (a mouse click) and tell widget B to do something, simple because there is no opportunity to program. On the other hand programming is itself a daunting task for many potential developers. A framework is needed where the same primitive application can be generated but where it is nonetheless possible to expand the application via simple programming. We propose here the generation of applications by control system wizards, which react to input criteria and render fully functional code to the desired platform. This code can be run as is and/or be used as a starting point for further development. Specifically, the TINE[1] server wizard will be presented, which generates server-side projects in C or Visual Basic. The TINE client wizard will also be presented, which generates client-side projects in Visual Basic, Java, and DDD (DOOCS[2] Data Devices). In the case of the client wizard, an application exists as a meta-application in XML format, specifically UIML (User Interface Markup Language).
[1] Philip Duval, The TINE Control System Protocol: Status Report, Proceedings PCaPAC 2000, 2000.
[2] Kay Rehlich, Experience with an object oriented control system at DESY, Proceedings PCaPAC'99, 1999.

Philip Duval and Honggong Wu - DESY, Hamburg, Germany

The ACOP[1] (Accelerator Component Oriented Programming) ActiveX control has been extensively used in control applications not only at the DESY accelerators but at many other laboratories as well. A new ACOP Java Bean reusable component has recently been developed and tested. This new component keeps same user interfaces as the ActiveX control. As in the latter case, it offers plugs for different data exchange protocols, such as TINE and Channel Access. Its graphic package was also adopted from ActiveX control, offering the same features fine-tuned for accelerator application development. The ACOP Java Bean can be used in a very similar fashion in a Visual Java environment just as the case for ActiveX in Visual Basic. Detailed development issues as well as a common user interface will be presented. In particular, we shall present benchmark results concerning ActiveX and Java Bean performance as well as ease of development.
[1] I.Deloose, P.Duval, H.Wu, The Use of ACOP Tools in Writing Control System Software, Proceedings ICALEPCS'97, 1997. and http://desyntwww.desy.de/acop

S. Baird, J. Fullerton, J-J. Gras, A. Guerrero, S. Jackson, L. Jensen (CERN-SL/BI), H.Hiller (B&S), Switzerland

To enable the use of SPS as an LHC injector, and to comply with the standards and directives laid down by the SL Controls group, most of the existing software associated with the SPS Beam Instrumentation will require re-engineering during the next 2 years. Faced with the magnitude of this software re-development task, we developed a set of internal rules, generic tools and templates. These facilitate, and in some cases automate, the development of the entire software chain from the real-time to the user interfaces. The benefits of our approach, and the results obtained so far, will be presented.

Tao Xiaokui, Sun Jun, Tong Dechun - Department of Engineering Physics, Tsinghua University, P.R. China

An automatic measuring system for cavities tuning has been developed recently to improve the efficiency and accuracy. In this system, stepper motors which are positioned well and automatically controlled easily are used to drive measuring antennae. In order to measure automatically, a controlling program has been developed based on LabVIEW. Using this system, the resonant frequency, the quality factor, the coupling factor between two cavities, the pass band performance, the field distribution and the shunting impedance can be all gotten easily and quickly.

Masakatsu Mutoh, Yoshinobu Shibasaki - Laboratory of Nuclear Science, Graduate School of Phisics, Tohoku University, Japan; Isamu Abe - High Energy Accelerator Research Organization (KEK), Japan

A generalized device layer for the COACK system has been developed. Software architecture for accelerator control systems commonly consists of three layers. COACK is a generalized control kernel having highly useful functions, and is installed in a middle-layer of the three layers. In regard to the lower device layer, a unique device layer has been developed that is adaptable to the target control system under various conditions. In the accelerator control system, PLCs or instruments using GPIB are employed to control various devices such as magnet power supplies, RF components, vacuum components, beam monitors, and so on. If each device configuration and control procedure could be standardized, a generalized device layer could be constructed. In order to realize the generalization of the device layer, XML is used to express the device configuration and the control procedure. The device layer is installed automatically according to its configuration and control procedure. Consequently, the developed device layer can be applied to various control systems by means of only rewriting files described the device configuration and control procedure in XML without any modifications.

J. McDonald, H. Nishimura - Lawrence Berkeley National Laboratory, USA

Portability of GUI controls programs that use the EPICS-based system has become an issue in the ALS hybrid control system. In the area where the run-time performance is not required, interpreters like Tcl/Tk, Perl/Tk and Python/Tk are a solution. However, once run-time performance is required, programs have to be natively compiled. This talk is on our recent experience on developing GUI controls application programs using Delphi on Windows and Kylix on Linux to share the common source codes including GUI, database and machine controls using a cross-platform SCA component.

H. Nishimura, J. McDonald - Lawrence Berkeley National Laboaratory, USA

A series of software tools to enable interactive access to the EPICS channels have been developed by using off-the-shelf components in C++ Builder, Delphi and Kylix. They cover wide range of the fundamental tasks from a low-level channel browsing to a high-level data analysis that are complementary to our Matlab-based automation programs. One of the examples will be a generic template program that is visually customized by machine operators to create control programs. Our paper will focus on the use of modern, high-level components for rapid development.

C. Timossi, H. Nishimura, J. McDonald - Lawrence Berkeley National Laboaratory, USA

The Advanced Light Source at Lawrence Berkeley National Lab utilizes Microsoft Windows operating system for much of it's graphical interface both for accelerator and beamline control. The need to support multiple development platforms such as National Instrument's Labview, Microsoft's Visual Basic and Borland's Delphi is met by deploying an ActiveX Control that wraps the interface to the EPICS system used for machine control. Experiences using this control for a variety of controls purposes will be presented.

A. Drago, A. Stella - INFN-LNF, Frascati, Italy

In this paper, a data acquisition system (DAS) developed to track dynamically the beams in DAFNE PHI-factory is presented. From a hardware point of view, the DAS is composed by one or two oscilloscopes and a personal computer, connected together through IEEE 488 interface. The software is written in LABVIEW 6.0 and it is portable on different platforms. As output, it produces a database that can be read by any spreadsheet or text editor.
A LECROY LC584AM oscilloscope, built around a PowerPC working at 96MHz, can acquire 8-bit data coming from four channels. Special feature of this oscilloscope is the capability to be synchronized with an external clock in the range 50 - 500 MHz. The clock signal comes from the DAFNE timing system and, through a phase shifter, is used to select the bunch.
The oscilloscope samples the difference and sum signal in horizontal and vertical from a Beam Position Monitor (BPM). A second oscilloscope can sample synchronously four signals from another BPM (this feature is under development). It is possible to store up to 500k contiguous fixed-point values for each channel; these data correspond to a record of 25k consecutive bunch turns in the Main Ring. The system is standalone, even though the operator can work remotely from the consoles of the DAFNE control system.
If the record is considered valid, data are downloaded to the computer; the transferring of up to 1Mbytes for each channel has to be done as quickly as possible to perform many measurements without delaying the machine operations. Then the measurements are converted to millimeter; besides other beam data are calculated using machine parameters or the Hilbert transform. At the end, data are automatically inserted in a database tree that is built using time stamps and can be remotely accessed by the users. Off-line retrieval and presentation tools are also developed in LABVIEW.

Developing a Control System from a Divan Bed

Grega Milcinski, Janez Dovc, Mark Plesko, Gasper Tkacik - Josef Stefan Institute, Slovenia

A dream thought to be made possible by Internet - to develop at home, to manage projects from anywhere in the world - should become a reality with high speed connections and new "productivity" tools. Did it work for you or do you find yourself clogged down with hundreds of unread e-mails and the ones read sorted (seemingly not enough) into several tens of folders on the basis of severity, projects or people? Our experience shows that e-mail collaboration is ineffective with a group of five people or more: tasks are left undone and messages are lost. Two solutions are possible: either to use a commercial package for project management, which always has "that crucial feature missing", or to adapt an open source application for your needs. We opted for the later and modified ResourceTracker. The article is about this tool and our adaptations, which integrate it with automated build procedures, web page generation and other e-management tools all backed by an SQL database. We present a sample project flow from receiving an order to delivering a product. The system works with a group of students and their professional results satisfy the most demanding customers coming from Europe, USA and Japan; therefore we believe it to work for a group of responsible programmers as well.

Klemen Zagar, Anze Vodovnik - Jozef Stefan Institute, Slovenia

Development of a control system typically requires that many software artifacts are kept in sync with one another. For example, if a hardware device is replaced with a newer model which exposes some additional functionality, this might require changes or additions in device driver code, in networking protocol, the operator's user interface, and the configuration database. Such modifications are often trivial: addition of functions, graphical user interface elements, and database columns, all based on established templates. Traditionally, software engineers would be assigned the task of affecting these changes, which is error prone -- e.g., forgetting to make a modification in the configuration database -- and time consuming -- performing regression tests to check that nothing has been broken by the widespread modification. To ameliorate these issues, we propose describing devices using XML files with a well defined schema and generating all the software artifacts from these descriptions using templates. We have developed a specialized language called XPGL (eXtensible Program Generator Language) for the purposes of defining the templates, which is presented in this article.

Dragan Vitas, Mark Plesko, Gasper Tkacik, Ales Pucelj and Igor Kriznar - J. Stefan Instititute and Cosylab Ltd. , Slovenia

The modern approach in software development for control systems requires RAD tools like the CosyBeans/Abeans framework. Those tools enable the developer to quickly produce an application with well-tested functionality. This shifts the developer's focus to device integration, which is always case specific and needs good testing. When the new application is deployed on the real system for the first time, everybody expects that it will run smoothly and without any strange behavior. In practice this is not a usual case. Thorough tests are still necessary to verify that the application behaves according to specifications, in particular the graphical user interface. This can be done only in a realistic environment. Unfortunately, it is a common case that developers are not allowed to interfere with a running system.

We have therefore developed a generic simulator that tries to simulate as realistically as possible a running system. Even erroneous behavior or situations that are not likely to happen ever on real system can be easily produced with simulator to test how the application will respond to them. The application programmer interface of the simulator allows to adapt it to different control systems and to write simuletts, small routines that simulate a particular behavior. In this paper, we present the architecture of the simulator and different areas where we have successfully used it. Using the simulator, our developers have polished their masterpiece, improved its performance and got rid of bugs before anyone even notice that they had them.

Kazuyuki Nigorikawa, Junichi Kishiro, Takashi Kosuge, Isamu Abe and Shin-ichi Kurokawa - High Energy Accelerator Research Organization (KEK), Japan

The beam monitor system in the KEK proton synchrotron is consisted of a lot of module interface called WE7000 series. The control and data presentation program coded by using "Control API". However, this API does not support Client/Serve model. We developed "WE Manager" by Visual Basic in order to realize the Client/Server model for these modules. The manager has two components. One is used to control the WE7000 itself ("Control Component"), the other is used to establish to network connection ("Network Component"). The network component was developed to have an extendable connection not only to TCP/IP but also connection to other protocol, such as COACK (Component Oriented Advanced Control Kernel), STARS (Simple Transmission and Retrieval System) etc. The test operation is now undergoing aiming to make it as a standard I/O server for monitor program.

Gerhard Grygiel - Deutsches Elektronen Synchrotron, Germany

With respect to the TESLA project it's required to install, configure and run thousands of frontend computers. These procedures should be automated and predictable.The first installation, the upgrade of a system and the exchange of a defect node should be completely automatic. The lose of data should be minimal. With the Distributed Object Oriented Control System (DOOCS) and the Debian GNU/Linux Operating System we have demonstrated such as a system at the TESLA Test Facility. It's realised with the Full Automatic Installation (FAI) which is based on a class concept. To install a new node it's only necessary to register the MAC address, define the node name and the FAI classes. All software dependencies are resolved by the Debian package system and all Control System configuration are defined in FAI classes. In case of a reinstallation, the configuration is retrieved from a backup system.

Takashi Kosuge, Yuuki Saito, Kazuyuki Nigorikawa, Hiroaki Katagiri, Akihiro Shirakawa, Hiromitsu Nakajima, Kenji Ito, Junichi Kishiro and Shin-ichi Kurokawa - KEK(High Energy Accelerator Research Organization), Japan

A central control system for beamline interlock in the Photon Factory was developed using COACK (Component Oriented Accelerator Control Kernel) in 2000, and it is working stably without system failure. The central control system is connected to a private LAN, which has a firewall, so we can monitor or control the system anywhere within the LAN. Since it is sometimes requested to access the monitoring system from outside of the firewall, it was decided to implement other COACK server for outside clients. One of the important issue for multi COACK-servers is the synchronization between these, and we have find that STARS (Simple Transmission and Retrieval System) performs effectively for the synchronization. STARS is a transferring message software for small control systems with TCP/IP socket, and it works on many kind of operating systems. Originally, STARS was developed as an interface program of COACK for non-Windows system. Then it is very easy to implement the STARS architecture in the COACK. We will describe detail of multi COACK server system with STARS here.

M. Kadunc, I. Kriznar, M. Plesko, G. Tkacik - J. Stefan Institute and Cosylab Ltd., Slovenia

Object Explorer is a generic application that discovers at run time all devices, properties or channels that can be controlled, analyses them using CORBA and Java introspection tools and displays methods, fields, commands and other actions in a GUI. The user can select and invoke these actions, including entering arbitrary parameters. The application was first developed by the KGB group at JSI for testing the ANKA control system. A completely new version has now been developed for the ACS control system software and the Abeans API. The application is divided into two parts: The engine part is control system-specific and communicates with the underlying data or server layer, providing data for the GUI part and passing user requests to the control system. This layer is relatively thin and can easily be implemented for many server protocols such as TINE, EPICS etc. The GUI part is independent of the engine implementation and of the communication protocol used. It receives data from the engine and shows all the available information to the user. This article describes the functionality of the Object Explorer, some of the software designs used, and discusses some ideas which might be implemented in future releases of the application.

Igor Verstovsek, Janez Dovc, Miha Kadunc, Jernej Kamenik, Igor Kriznar, Gasper Pajor, Mark Plesko, Ales Pucelj, Gasper Tkacik - Jozef Stefan Institute, Slovenia

The New Abeans and CosyBeans are in their third generation, developed through experience gained from several projects, such as ANKA light source, ALMA radio telescope project at ESO and NRAO and from cooperation with control groups from DESY, Riken and elsewhere. Abeans form an application framework and CosyBeans form a GUI framework, designed for developers of control systems of large experimental setups. These allow simple development of complex applications, providing all the benefits of the underlying Java platform, such as Java logging, high-speed graphics of Java 1.4 and easy installation and maintenance with Java Webstart. Abeans is a library that provides simple Java beans for connection with the control system. At the same time, it provides several useful services: logging, exception handling, configuration and data resource loaders, authentication, and policy management. CosyBeans provide clear and consistent visualization of dynamic data with standardized presentation of alarms, monitors and connection status. By means of Adapters that act as a mediator between the underlying remote data layer and top-level GUI, CosyBeans decouple presentation of data from the actual data sources. In this manner, greater portability and consistency are achieved. Both frameworks can be used alone independently but really excel when used together, enabling a coherent look of applications that run on different protocols. We have provided plugs for CORBA, EPICS and TINE. We present some of the main Abeans and CosyBeans concepts, such as Services, Abeans Engine, Models, Plugs, Displayer, Adapter, CosyPanel and Plug-in.

Rok Sabjan, Matej Sekoranja - J. Stefan Instititute and Cosylab Ltd., Slovenia

Visual DCT is an EPICS configuration tool completely written in Java and therefore supported in various systems. It was developed to provide features missing in existing configuration tools as Capfast and GDCT. Visually Visual DCT resembles GDCT - records can be created, moved and linked, fields and links can be easily modified. But Visual DCT offers more: using groups, records can be grouped together in a logical block, which allows a hierarchical design. Additionally indication of data flow direction using arrows makes the design easier to understand. Visual DCT has a powerful DB parser, which allows importing existing DB and DBD files. Great effort has been taken and many tricks used to optimize the performance in order to compensate for the fact that Java is an interpreted language. Visual DCT is fast becoming the most popular tool among EPICS community, constantly being updated for performance tunning. In this paper we present our experiences with using VDCT to develop production EPICS databases. Problems and suggestions for improvements will be discussed.

Kirsten Hinsch, Ursula Lauströer, Rüdiger Schmitz, Winfried Schütte - DESY Hamburg, Germany

Here at DESY we have nine large particle accelerators and correspondingly a large and complex control system. It consists of subsystems of many generations of hard and software. Until recently we even had to support Norsk Data mini computers. The number of applications is considerably more than a thousand. The control team consists of a dozen people aged typically in the fifties. This large, diverse and understaffed system introduces a considerable amount of (healthy) inertia. On the other hand we have a fast hard- and software upgrade cycle in the general computing scene. There is every few years a major operating system upgrade. This conflict produces tension and stimulation.

Custom HMI for Commercial SCADA

V. Aleinikov, G. Dreyzina, A. Nikiforov - Joint Institute for Nuclear Research , Russia

A new cyclotron complex for the production of radioactive nuclear beams (DRIBs &endash; Dubna Radioactive Ion Beams) was developed and manufactured in Flerov's Laboratory of Nuclear Reactions. The control system for beam transportation system uses commercial SCADA FlexControl running under RTOS QNX4. Because of good interaction possibility it was able to develop self-made visualization server. It has highly improved flexibility and look and we use it instead of the regular one. Basic ideas for integrating custom HMI and commercial SCADA are described.

Jianjun Hu*, Stefania Canella - *IMP-CAS, Lanzhou, China - L.N.L.-I.N.F.N. Legnaro, Italy

GridBagLayout is a very sophisticated, flexible layout manager provided by the Java platform. It can lead to better overall GUI development and a really robust final product. In this paper, important topics about GridBagLayout will be mainly discussed on the cross-platform compatibility of GUI and the Look and Feel in different runtime environment. At the end of this article, an example GUI of steerer control for ECR ion source with layout structure tree is presented to show how to organize the GUI primarily with GridBagLayout layout manager. At last, the GridBagLayout layout manager can lead to GUI that more precisely meet the original specifications and the expected run-time requirements for accelerator control system.

M. Kadunc, J. Kamenik, M. Plesko, A. Pucelj, I. Vertovsek - J. Stefan Institute and Cosylab Ltd., Slovenia

Control system applications, usually developed using modern RAD tools, are often inconsistent and do not meet the visual or functional requirements of the users. To avoid such problems, Cosylab adopted a set of guidelines for writing a good user interface. Most of these guidelines were gathered by studying various user interface approaches in control system and common PC applications, identifying specific problems and finding the most acceptable solutions. During this talk some exemplary GUI widgets developed at Cosylab will be displayed and commented. The talk briefly discusses the choice of a color scheme for control widgets, labels and other components often used in control panels, as well as their layout on screen or in a window. Font, case and language of different texts used in applications are also suggested. Furthermore, the talk focuses on the interaction between the user and the computer, first by enumerating and evaluating ways of displaying messages to the user, and then by addressing accessibility issues, such as keyboard shortcuts, contextual (pop-up) and window menus, toolbars etc. Finally, it addresses the customization of application workspace and the problem of finding a balance between user-specified and default look and feel of the applications.

G. Pajor, J. Dovc, M. Kadunc, I. Kriznar, M. Plesko, G. Tkacik, I. Verstovsek - J. Stefan Instititute and Cosylab Ltd., Slovenia

In this article we present our basic GUI components and how we build GUI panels using visual and conventional programming. The new additions to the GUI components are discussed in light of their flexible design. Building an application as a front-end for control system is a responsible task that merges end-users' expectations of easy-to-use tool and developers' skill to find the right solutions. The most efficient way of writing applications that have certain similarities is obviously to create re-usable components. These components should be generic enough to fit into wide array of possible application specifications. On the other hand it should be possible to meet the specifications completely with very little additional coding. Re-usable components in ANKA accelerator project case were designed at the time of the project and were success by itself. About two dozens applications were built and virtually no application-based adjustments have been made to the components due to their generic nature. The slightly different aspect of our past work was presented to us as we began working on another project, beamline control application for Japanese commercial customer. We built the graphical user interface for the application and upon received screenshot they developed a quite clear idea of what the application should look and behave like. Their specifications went beyond the implemented functionality of components at that time. Due to good engineering of the components we added the requested functionality without problems. Decision should be made carefully which of the new features can be implemented generically enough to put them into the re-usable components instead of implementing it in application itself.

Timofei Bolshakov, Andrey Petrov, Fermi National Accelerators Laboratory, USA

Synoptic Display is a Java application for flexible online graphical representation of data received from a data acquisition system. We consider this project as next generation of programs such as ACNET Lexigraphics and EPICS MEDM. Synoptic Display projects (an equivalent of MEDM screens) are rendered on major Web browsers (for monitoring purposes) or started in a web-startable console Java application (both for monitoring and control). Small bandwidth (as low as 100-1000 byte/sec) is required between client and server sides due to usage of Scalable Vector Graphics (SVG) for data transfer. Synoptic Display components (data sources, processing pipes, visualization widgets) can be graphically arranged and logically interconnected in a web-startable Project Builder; and projects can be saved on a server-side repository. XML is used as a project format. A Runtime Project Engine (RPE) on the server side handles user's requests, downloads projects from the repository, launches data acquisition jobs, and generates SVG pictures. Tomcat Servlets and JSPs (Java Server Pages) are used as RPE web-tier. At present time ACNET Java Data Acquisition Engine is a primary data source for the Synoptic Display, since this is corporate Fermilab standard; there are no limitations to create new types of data sources.

Timofei Bolshakov - Fermi National Accelerators Laboratory, USA

SVG (Scalable Vector Graphics) is a language for describing two-dimensional graphics in XML. This standard has been issued by the Web Consortium. The author presents how this standard is used in FermiLab for the web-based projects "Data Logger Plotter", "SDA (Shot Data Analysis) Intensities" and "SDA Reports". All of these projects are based on Java servlets and JSPs (Java Server Pages) and work under the Jakarta Tomcat Servlet Engine.

A Gateway to Fix32 System

J.M. Nogiec - Fermi National Accelerator Laboratory, Batavia, IL, USA

The Fermilab's Magnet Test Facility (MTF) performs tests of superconducting accelerator magnets and components with help of two systems: the Distributed Monitoring and Control System (DMCS) and the Intellution Fix32 control system. The former system monitors test subjects and is responsible for quench detection and protection. It also controls test conditions, provides data visualization, and archives monitoring and quench data for on-line and off-line analysis. The latter system controls the MTF's refrigerator and is responsible for providing cryogenic conditions required for testing. In order to conduct tests and correctly interpret test results, data from both systems have to be simultaneously available for analysis and presentation. To satisfy this requirement a gateway system between both the DMCS and the Fix32 system has been developed. The gateway allows for bi-directional access to cryogenic data stored by Fix32 from the DMCS scanning system. It offers both an API and script-level access. Since the gateway is a stand-alone system it can be used to integrate Fix32-based systems with systems other than DMCS. The article will present the architecture of the gateway and discuss its implementation and functional aspects.

Toshiya Tanabe, Toshikatsu Masuoka, Motonobu Takano and Masayuki Kase - RIKEN-Beam Technology Division, Japan

New control systems for the first phase of RIKEN-RI Beam Factory (RIBF) will be constructed in coming years. Vender independence, software sharing and interoperability are the main aims of the design. Distributed objects through CORBA communication layer are used for the framework and Java is preferred for GUIs. Software sharing with Jozef Stefan Institute(JSI) of Slovenia via CORBA/JavaBeans has been successfully implemented. With the exception of field controllers, which are VMEs and some PLCs, PCs are used for not only for the clients, but also the application server and a RDBMS. The details of system description and the results of our R&D work are to be reported.

LiPing Yang, MingChuan Zhan, YuLan Yu, YongGang Li, JiHua Lu - Istitute of High Energy Physics, Chinese Academy of Sciences, P.R.China

The old data acquisition system is based on the CAMAC system of BEPC control system. Its speed of the acquired data is about 1 time / 16S, and the CAMAC bus is occupied by it long time during acquiring data through the CAMAC modules and the DVM. Therefore the performance of whole control system is influenced deeply. In order to change this state, we use the hardware products and LabView of National Instrument (NI) company and build the new fast acquisition system of main B,Q power supplies with an Industry PC (IPC) as an intelligence local station. The precisions (< 5/100,000) and speed (max.: 85 times/S and normal: 1 time/S) of the data acquisitions of new system absolutely are superior to old system (its original precision: < 1/10,000 and speed: 1 time/16S). The fast acquisition system also can be used to monitor the whole ramping process for diagnosing the control system and characteristic of main B,Q power supplies.

P. Duval, Z. Kakucs - DESY, Germany; I.Kriznar, M. Plesko, A. Pucelj, G. Tkacik - JSI and Cosylab, Slovenia

The standardization in PC and network technology has produced a distinct (for want of a better term) "babylonization", where islands of control exist in perfect ignorance of each other even though they might belong to the very same facility. This is due in part to commercial equipment which often comes with its own control software and to the many excellent but different solutions for control systems which have been developed in the accelerator control community. A control systems integrator frequently has to make decisions with long-term and far-reaching consequences. Often a pragmatic approach is to allow resourceful engineers to use the best available tools to solve controls problems and then to integrate their solutions into the control system. It usually turns out that integration, if not done systematically, amounts for the largest part of the work. There are usually many ways to do this, for instance defining a software bus, using gateways, or simply allowing apples and oranges to peacefully coexist. In this paper, we will examine most of the available tools in our community for the integration of control systems, detailing the merits of each approach as well as some popular controls systems and components. We will demonstrate that it is possible to mix them in order to benefit from the best part of each.

M. Drochner, H. Kleines, P. Wuestner, K. Zwoll - FZ Juelich / ZEL, Germany

In our lab, PC technology is used for a couple of years in developments for experiment control and data acquisition where less mainstream equipment (embedded, realtime, VAX) was dominating before. In neutron scattering experiments, spectrometer control and detector readout are done by CompactPCI systems, partly using custom peripherals. In particle physics data acquisition, passive backplane systems, tailored to our needs, are used. Experiences gathered in these two application fields, and future directions and expectations will be presented.

P.J. Yogendran, E. Klassen, K.S. Lee, M.M. Mouat, J.J. Pon - TRIUMF, Vancouver, Canada

A PC based link connecting a VME crate and a CAMAC crate has been configured. This setup is used to provide control and data acquisition between CAMAC in the Central Control System and specialized VME-based hardware used with wirescanners. Commercial hardware and software drivers together with in-house written software have been assembled to provide a bi-direction link. The hardware and software configurations, performance, and other aspects are described.

A. Pucelj, R. Sabjan, G. Tkacik - Jozef Stefan Institute, Slovenia

Accelerator community offers wide range of low-level communication protocols, from minimal solutions such as EPICS channel access to CORBA object model. This makes it natural to attempt their integration into single framework. Such generalizations are inherently susceptible to performance handicaps arising from either inappropriate abstraction or platform dependence of individual protocols. We present our experience in this area along with solutions and open questions to the problem. This article presents examples of integration of EPICS channel access and TINE protocol into Abeans framework using pluggable model. Focus is placed on the user/developer aspects of such implementation outlining the advantages of such approach. First, we show how simple access to the control system can be made in a general way. Comparison of different approaches shows the advantages of using a single interface and benefits for the developer. Next, we look at examples of actual plug implementation and the common points between the systems. Lastly, performance impact is analyzed. This is especially important, since our implementation is Java based, which can under circumstances be slow and memory consuming. Several methods are presented that eliminate this overhead. This is confirmed with actual examples.

Norihiko Kamikubota, Kazuro Furukawa, Junsei Chiba, Hidetoshi Nakagawa, Susumu Yoshida*, and Makoto Tokagi* - High Energy Accelerator Research Organization (KEK); *Kanto Information Service Co. Ltd., Japan

The construction of High-Intensity Proton Accelerator Facility, which consists of the 400-MeV linac, the 600-MeV superconducting linac, 3-GeV rapid-cycling synchrotron, and 50-GeV synchrotron ring, started in 2001 as a 5-year project. This is a joint project between JAERI (Tokai, Ibaraki) and KEK (Tsukuba, Ibaraki). The construction site is Tokai. However, the firsr 60-MeV of the linac has been under construction at Tsukuba site, and early studies with proton beams started recently. A prototype of the control system based on EPICS toolkit has been developed for the 60-MeV linac. Though we need X-window consoles, the only choice is to use PC platform: because nowadays less number of X-terminals is available in commercial markets. Up to now, we have studied the consoles with three operating systems, Linux, Free-BSD, and Windows, at the 60-MeV linac. In this report, we will discuss a realistic perspective of a PC-based console system suitable for our project.

G. Di Pirro, A. Stecchi, I. Sfiligoi, G. Mazzitelli - INFN-LNF, Frascati, Italy

The DAFNE Control System is based on a central shared memory where many distributed processors continuously store information and where user applications fetch the data from.
Such memory is made by merging all the distributed processors RAM in a common VME address space through point to point optical interfaces. This structure, together with a software based on LabVIEW, proved to be extremely flexible during the upgrade of the 2nd and 1st system levels. In this occasion, the multiplatform feature of LabVIEW allowed to reuse, with only few modification, all the 2nd and 1st levels software in spite of the drastic hardware change.
Now we are facing a similar challenge: the 3rd level VME processors, based on low-cost Macintosh logic board, become obsolete and we are going to install Pentium Embedded processor with Linux instead.
This paper describe the strategy that we have followed to fit a different platform into the existing 3rd level framework still reusing the existing software.

Socket-based Communication for Accelerator Control Using Linux, NT and Java

Stefania Canella, Jianjun Hu* - L.N.L.-I.N.F.N.; *IMP-CAS

Network communication is a very important topic in the software development for accelerator control systems. Socket-based interconnection provides the foundation for most of today's connectivity at application level. This paper describes how sockets have been used in a set of programs for the control system of an ECR ion source. These applications run in different environments(Linux, NT and Java) and have been writen in C language(for the server side) and Java-Swing( for the GUI client side).

S. Ramakrishnan - Scuola Superiore Sant'Anna, Pisa, Italy

The paper main focus is to identify weaknesses in present wireless network and to formulate attack trees to present them. The paper hopes to be a bridge to make a systematic effort to ensure wireless security. We start with the basics of attack trees and their use. Then all the present known problems in wireless network are dealt in detail. In section 4 then we summarize the weaknesses and design a high-level attack tree for wireless security. On the basis of the attack tree then a set of possible problems solutions are suggested. Finally I conclude with possible things that can be refined or dealt more from where the paper leaves them.

Matej Sekoranja - Jozef Stefan Institute, Slovenia

Distributed computing provide many solutions but choosing the correct programming model has become difficult. A comparison of various distributed computing technologies Java/RMI, DCOM, .NET, One, CORBA and RT-CORBA gives us the following results: some perform well only for passing simpler data types, some are too basic for a large system design, and others are only visions that are not ready for major applications. The only technology that is ready today for mission-critical, multiple platform, multiple language projects is RT-CORBA. RT-CORBA is an optional set of extensions to CORBA tailored to equip ORBs (Object Request Brokers) to be used as a component of a real-time system. It adds QoS (Quality-of-Service) control to regular CORBA improving the application end-to-end predictability. A brief overview, features and our experiences including some applications and tests of RT-CORBA are presented.

Steve Herb, Rüdiger Schmitz, Rolf Schröder, Harald Schulz - DESY Hamburg, Germany

The control system at DESY for the Hera, Petra, and Doris accelerators, and the associated pre-accelerators, includes about 500 computers (mostly PCs) using the IP and IPX protocols over ethernet for real-time communications. The control system network has recently undergone a major upgrade, with the BNC-based cabling and ~10 year old Router/repeater technology having been replaced by a structured cabling infrastructure using modern switches and routers from Cisco. The new system is built in conformance with, and is fully integrated into the DESY site network, but can also operate in stand-alone mode. Using the same structure as the site network has brought major benefits, as well as some understanding of the how the requirements for controls are different. We discuss this, as well as some of the tools which were (or would have been) useful for the commissioning and maintenance.

Hierarchical Data Archival System for EMS

J.M. Nogiec, K. Trombly-Freytag, D.G.C. Walbridge - Fermi National Accelerator Laboratory, Batavia, IL, USA

The Extensible Measurement System (EMS) has been developed at Fermilab's Magnet Test Facility to allow for standardization of systems and to increase reuse of software. It is a component-based framework for building data acquisition and analysis applications. As a truly open system, EMS does not enforce one specific way of storing data. As a truly extendable system, it is capable of accommodating various archiving methods and formats. The user configuring an application can choose a preferred method by selecting appropriate components. The article presents one such solution, a two-stage, hierarchical storage system, which incorporates both object-oriented and relational databases. This solution accommodates the needs for performance, program-level access to data, and user-oriented access. The architecture, results of performance studies, and a comparison with other approaches are presented.

I. Krylov, E. Ushkova - RRC Kurchatov Institute, Russia

Kurchatov Center of Synchrotron Radiation - the electron accelerators facility, including the linear accelerator, the synchrotron - booster at energy 450 MeV and the main ring at 2.5 GeV, is a dedicated synchrotron radiation source. The lower level of control system consists of CAMAC hardware for the data acquisition and control. The operators level consists of the PC network. The system of data acquisition about parameters of a beam and technological systems of KSRS facility is described and examples of network applications are given. Programs of creation of archive of the parameters are described. Examples of using of archives for the analysis of a facility status are given.

D. Quock, F. Munson, S. Dean, R. Enshiwat - Argonne National Laboratory, USA

ATLAS (Argonne Tandem Linac Accelerator System) accelerates heavy-ion beams, and consists of 3 separate ion source injectors, 64 superconducting resonators, and 3 target areas. As with any complex accelerator facility, database management systems are used extensively to support both the archiving and retrieval requirements of operational data, and the real-time control needs of the accelerator. The integration of Web technology to the already existing ATLAS Oracle Rdb database system and control system documentation has recently been deployed to enhance communication to the operators, control system engineers, and other technical support staff at ATLAS. This paper reviews the historical development of the control system database structure and user interface, and the Internet software developer tools used to create a new online interface.

D. Filimonov, S. Karnaev, B. Levichev, S. Smirnov, A.Veklov, Budker Institute of Nuclear Physics, Novosibirsk, Russia

The VEPP-4 control system was designed almost 20 years ago as CAMAC-based distributed control system [1]. The CAMAC-embedded 24-bit Odrenok computers are used in the system as main control machines. The main problem of the control system further development is the change the obsolete computers on PC. For ensuring the access to equipment and data for programs, working in PC computers, the host-based database is designed. The communication software is Linux-based and includes the Application Server in the host machine and Device Servers. The Application Server provides channel and database access for user applications. Device Servers provide interactions between Application Server and executive programs working in executive computers, connected to electronics. The backend database server is based on PostgreSQL. This paper describes a structure of the VEPP-4 control system database server and the database access procedures.
[1] A.Aleshaev, et al., 'VEPP-4 Control System', ICALEPCS'95, Chicago, USA

Ivanka Valova - Institute of Control and System Research, Bulgarian Academy of Sciences, Bulgaria

This paper describes an approach for organizing data using the multidimensional paradigm in the form of data cubes. A multidimensional cube is constructed based on set static variables, dimensions, and a set of dynamic variables, measures and consists a collection of cells. Dimensions, measures and data cubes with their operators are formally introduced. We have also considered two approaches for implementations: relational approach and object - oriented approach. With the first model a cube model can be implemented by mapping the operations to the appropriate SQL queries. We give example operations and corresponding SQL code.

In the other alternative dimensions, measure schema, and cube schema are grouped into a term of metadata, which will be mapped into object-oriented database in term of classes. In this direction we have also considered a modelling of MDDM in term of classes by means of UML. Finally a multidimensional model of multi- agent system for control, diagnostic and monitoring is presented.

Andreas Luedeke - PSI, Switzerland

The reference of actor / sensor values in control systems is in general done by names, so called channels. High level application have these channel names either hard-coded or include them by configuration files. The availability of a channel is normally checked at runtime. But it is difficult to do proper runtime testing, since a change or an error on a channel server can affect all application using these channels and it is inconvenient to start all applications for a test. Therefore an automatic availability check of all channels used in any configuration file of a high level applications is desired for the SLS. The check has three branches: Each channel server registers all served channel names in a central database. Each configuration file is uploaded to the same database and can be downloaded to the control system from the database. Both tables are compared periodically and missing channels are reported with a description, when they vanished from which server. The naming convention of the SLS constructs a channel name from a devices name and a property. In a third branch the available and the required channels are checked against the valid device names defined in the so called master-table and the properties defined for the property class hierarchy of the device.

Michael Boege, Jan Chrin - Paul Scherrer Institute, Switzerland

An interface to the Oracle database, implemented using the Common Object Request Broker Architecture (CORBA), provides clients with a set of methods that execute prepared Structured Query Language (SQL) statements and return result sets within the context of a CORBA Object. Such an interface has the distinct advantage in that clients may perform database operations in a more user-friendly manner and without the need to acquire knowledge of SQL syntax or the vendor's Application Programme Interface (API). Conversely, a set of CORBA application objects that permit generic database retrieval and modification operations has also been provided for clients that require the increased flexibility attained through direct execution of SQL commands. The application objects have been implemented using the C++ programming language for optimal performance with database queries performed using the Oracle Call Interface (OCI) through the Oracle Template Library (OTL). Client programs accessing the database server are typically written in Tcl/Tk, Java or C++, thereby exploiting the language independent object inter-operability features offered uniquely by CORBA. The database server-client software model is presented, including a description of a mapping of relational database tables to CORBA objects, together with examples of usage in the first year of SLS operation, during which time both static and real-time data have been stored and accessed.

Matthias Clausen, Phil Duval, Zoltan Kakucs, DESY Hamburg, Germany

MACE (Monitored Archiving Controlled by Event) is an archive-server and data-server in-one running under Solaris. The data from the monitored channels (in case EPICS) is stored using an event number got from the event-server by request. The user can refind the stored data in a very centralized storage pool. Each archiver can handle the incomming triggers from the event-server. The result is an easily scaleable archiver system. Details of this archiving system are presented in this paper.

A Proposal for a Central Error Server Using Web Services

O. Hensler - DESY, Germany

As the DOOCS control systems, operating the TTF Linac at DESY, is a widely distributed system, running in TTF phase 2 on approximately 50 front-end- and middle-layer computer with 5 to 10 server processes on each processor, a system with local log files is not sufficient anymore. In addition a central logging of all error and warnings will be needed. We will propose a system based on Web Services running JAVA servlets on the central server(s) allowing the operator to get an overview of the machine status, as well as getting information of single devices by browsing through a tree like structure closely connected to the TTF nomenclature . This server stores the information in XML files, allowing the usage of standard XML tools like searching functions. The communication from the front-ends will be implemented reliable, independent and stateless on a common standard to be portable to different architectures. First implementations will be discussed.

Norihiko Kamikubota, Shiro Kusano*, and Kazuro Furukawa - High Energy Accelerator Research Organization (KEK); *Mitsubishi Electric System and Service Co. Ltd., Japan

A beam-status display for the KEK Injector Linac has been developed as a Java applet. Actually this applet is a CORBA client, which communicates with the CORBA server of the KEK Linac control system, to update the information by every ten- second interval. As a result, anyone can watch at the linac beam-status in real-time by using a web-browser of remote PC. Recent updates of this system will be reviewed and future prospects will be given.

Olaf Hensler, Raimund Kammering, Anna Petrosyan, Kay Rehlich - DESY, Hamburg, Germany

Since October 2001 the documentation of all accelerator operation relevant information at the TESLA Test Facility is accomplished by using the "electronic logbook". Due to the broad acceptance and usage of this service this e-logbook is now also used at the HERA accelerator and for other applications. Every standard web browser can function as a user interface for the input of text and retrieval of information. Graphical data is inserted by low level postscript print services to offer a platform independent input interface. The generation of PDF is provided for high quality printouts. The data is stored in the today widely used XML format to allow high performance searches and also to allow interfacing with other web based services. A standard web server like Apache is generating dynamic content by use of JAVA servlet technology. The key technologies used for this e-logbook will be discussed and a practical overview of its working will be demonstrated.

V. Kocharyan - YerPhI; K. Rehlich - DESY, Germany

To give a possibility to access data from different computer platforms we have developed a client part of DOOCS library in Java language (JDOOCS). It is fully identical to the DOOCS Client API and gives a possibility to read and write data from/to servers supporting DOOCS protocol and TINE protocol. For testing of the JDOOCS library we have written a generic application in Java that allows to access all data in the control system.

TINE Release 3.20: Status Report

Philip Duval - DESY, Hamburg, Germany

The dominant control system in the HERA accelerator is TINE[1]. TINE is playing an increasingly important role in the pre-accelerators at DESY as well as in Zeuthen. Furthermore, recent efforts have made the EPICS2TINE translator[2] a viable alternative to EPICS over Channel Access. TINE has also been successfully integrated into DOOCS[3] so that DOOCS clients and servers can run happily over TINE as opposed to SUN RPC. TINE plugs for Abeans[4] as well as COACK[5] have also been developed. As one might surmise then, demands on and from control system components have made TINE extremely flexible and adaptable to a wide variety of operating systems and platforms. The most recent release of TINE (Version 3.20) offers many new features, some of which are completely unknown in other control system protocols (such as network subscriptions via multicast). In this report, we shall give a status report on the TINE control system highlighting the latest features available in release 3.20.
[1] Philip Duval, The TINE Control System Protocol: Status Report, Proceedings PCaPAC 2000, 2000.
[2] Z. Kakucs, P. Duval, M. Clausen, "An EPICS to TINE Translator", Proceedings ICALEPCS 2001, 2001.
[3] Kay Rehlich, Experience with an object oriented control system at DESY, Proceedings PCaPAC'99, 1999.
[4] Gasper Tkacik, Java Beans of Accelerator Devices for Rapid Application Development, Proceedings PCaPAC'99, 1999.
[5] I. Abe, et al., "Recent Status of COACK and its Functions," Proceedings PCaPAC 2000, 2000.