Search

Performance, reliability and scalability in data-access are key issues in the context of the computing Grid and High Energy Physics data processing and analysis applications, in particular considering the large data size and I/O load that a Large Hadron Collider data centre has to support. In this paper we present the technical details and the results of a large scale validation and performance measurement employing different data-access platforms--namely CASTOR, dCache, GPFS and Scalla/Xrootd. The tests have been performed at the CNAF Tier-l, the central computing facility of the Italian National Institute for Nuclear Research (INFN). Our storage back-end was based on Fibre Channel disk-servers organized in a Storage Area Network, being the disk-servers connected to the computing farm via Gigabit LAN. We used 24 disk-servers, 260 TB of raw-disk space and 280 worker nodes as computing clients, able to run concurrently up to about 1100 jobs. The aim of the test was to perform sequential and random read/write accesses to the data, as well as more realistic access patterns, in order to evaluate efficiency, availability, robustness and performance of the various data-access solutions. Index Terms--Computer facilities, computer input-output, data handling, data processing, mass memories, system analysis and design.

Data are presented on preparation methods, properties, and main application fields of semiconductor chalcogenide scintillators (CS) on the basis of zinc selenide and other [A.sup.II][B.sup.VI] compounds. Peculiar features and mechanisms of the CS luminescence are discussed. CS luminescence spectra maximums are located within the range 590-640 nm, depending on the method of preparing the scintillator. The luminescence decay time ranges within 0.5-3 and 30-50 [micro]s. The afterglow level is less than 0.001-0.01% after 5-20 ms, and the radiation stability is [greater than or equal to] 5 x [10.sup.8] rad. Certain application fields of the new CS are examined taking into account their particular properties. Index Terms--[A.sup.II][B.sup.VI] compounds, chalcogenide scintillator, luminescence properties, radiation detector.

Carrier drift mobilities of CdZnTe (CZT) grown by high-pressure Bridgman (HPB) method and chlorine-doped CdTe grown by traveling heater method (THM) were measured by a time-of-flight (TOF) technique. A spectrometer grade crystal of HPB CZT shows room-temperature electron mobility of 960 [cm.sup.2]/Vs and hole mobility of 56 [cm.sup.2]/Vs. Both electron and hole mobility of THM CdTe: Cl crystals are higher (1100 [cm.sup.2]/Vs for electrons and 88 [cm.sup.2]/Vs for holes) than that of the HPB CZT crystal. Both materials show a saturation of the electron mobility at low temperature around 100 K and a strong decrease of the hole mobility with a lowering of the temperature. Theoretical mobility has been calculated by solving a Boltzmann transport equation assuming several scattering mechanisms such as polar optical phonon, ionized impurity, and alloy scattering. It is concluded from the comparison of the experimental and theoretical temperature dependence that shallow trap-controlled mobilities are observed in both spectrometer-grade CdZnTe and CdTe crystals. Index Terms--CdTe:Cl, CdZnTe, high-pressure Bridgman (HPB), mobility, time of fight drift.

An extrapolation chamber was designed and constructed, and its performance was studied in beta radiation fields. The characteristics of this chamber and the measurement procedures used for the calibration of beta-ray applicators are discussed in this work. The surface dose rate of a planar 90Sr + 90Y applicator was also obtained using the extrapolation chamber, and the results were compared with the Amersham calibration certificate. An agreement within 2.7% was verified. Index Terms - Beta dosimetry, calibration, extrapolation chamber.

We describe recent work on the development of a Compton scatter polarimeter for measuring the polarization of hard X-rays (100-300 keV) from astrophysical sources. Results from measurements with a laboratory prototype are summarized, along with comparisons to Monte Carlo simulations. We also present our latest design concept, which envisions a complete polarimeter module on the front end of a 5-inch position-sensitive PMT Although the emphasis of our development effort is towards measuring hard X-rays from solar flares, our latest design has the advantage that it is sensitive over a rather large field-of-view (> 1 steradian), a feature that makes the design especially attractive for [Gamma]-ray burst studies.

FAIR (FAst Intercrate Readout) is a synchronous ECL bus system dedicated to readout. It is based on a new trigger and readout hardware level protocol and an on a new control system that 'learns' how to setup and control modules. The hardware protocol along with the data structure allow both readout and event building at the same time at the rate of 22 ns/longword (1.44 Gbit/s) without the need of CPUs. It performs trigger management and full pipelining by using a multilevel FIFO structure. FAIR provides for a multi-crate front-end environment and uses an embedded serial network to accomplish front-end control and setup. The data transfer measured performances and the control system are presented in some detail.

A 32 channel 12 bit Time to Digital Converter (TDC) has been designed to fit in a single unit VME 6U. 32 Time to Amplitude Converter (TAC) sections and a multiplexer circuit are followed by a fast 12 bit ADC. The total handling and conversion time for all the channels is less than 10 [[micro]seconds]. The differential non linearity is kept below 2% of the average channel width. The Common Start/Stop mode, the Full Scale Range, the individual low and high threshold values and the self-test-mode are software selectable. The module has Zero Suppression and Chained Block Transfer protocol capabilities.

Absolute determination of thermal neutron flux is an important capability for several applications including neutron dosimetry, analytical chemistry, oil recovery, nuclear waste clean-up, and the determination of the neutron lifetime. Few methods exist that are calibrated to better than [approximately]0.5% accuracy. As a corollary to this statement, it is also [rue that few thermal neutron absorption cross sections are known to better than [approximately]0.5% accuracy. Several groups, including one at the National Institute of Standards and Technology (NIST) have worked for many years to develop methods for pushing the accuracy below the 0.1% level. This paper describes a calorimetric technique which avoids many of the problems associated with absolute counting methods. A description of the calorimetric method and results of system tests will be presented.

Two photomultiplier tubes with bialkali photocathodes and fused silica windows (2[inches] Philips XP2020Q and 1-1/8[inches] Hamamatsu R1668) were studied between room temperature and -160 [degrees] C, illuminated in dc mode in the UV/VUV wavelength range. The photomultipliers were operated as diodes, with direct measurement of the current of photoelectrons, in order to study in detail the effects associated to the photocathode. The influence of light intensity, applied voltage and illuminated area on the response as a function of temperature were investigated. The increase of the electrical resistivity of the emissive layer upon cooling proved to be responsible for a loss of linearity of the photocathode response, leading ultimately to the complete absence of signal at low temperature. A simplified model is presented that accounts for some experimental results. An unexpected increase of response below room temperature was also found at 170 nm.

The radiation detection properties of semiconductor detectors made of 4H silicon carbide were evaluated. Both Schottky and p-n junction devices were tested. Exposure to alpha particles from a 238Pu source led to robust signals from the detectors. The resolution of the Schottky SiC detector was 5.8% (FWHM) at an energy of 294 keV, while that of the p-n junction was 6.6% (FWHM) at 260 keV. No effect of temperature in the range of 22 to 89 [degrees] C was observed on the characteristics of the 238Pu alpha-induced signal from the SiC detector. In addition, testing in a gamma field of 10,000 rad-Si [h.sup.-1] showed that the alpha-induced signal was separable from the gamma signal.

In the framework of the EXCYT radioactive ion beam facility, now under development at LNS Catania, we have developed several devices for beam monitoring. These devices are mainly based on scintillator detectors and photosensors, since they should allow to reconstruct beam informations even when operating at very low intensity, where conventional electrical devices are no longer sensitive. This new class of devices promises to give a good reliability and robustness at a cheap price, and in some case it will also allow counting the beam particles one by one.

The scintillator energy resolution component which is due to light yield nonproportionality has been characterized for NaI(T1) and LSO. Results are based on a discrete convolution of measured electron response data and the electron energy distribution resulting from full-energy absorption events. The behavior of this energy resolution component as a function of energy is observed to be strongly dependent on the shape of the electron response. Furthermore, in some energy regions, the light yield nonproportionality component is observed to be larger than the resolution predicted by assuming Poisson photoelectron statistics. Characterization of this energy resolution component will facilitate deconvolution of other components from the total energy resolution.

We have conducted thermal measurements on a Silicon - Pyrolytic Graphite microstrip detector model similar to prototype modules being tested for the central detector of the ATLAS experiment at the future Large Hadron Colider (LHC). A thermally conductive pyrolytic graphite insert is sandwiched between two single-sided silicon detectors to extract Joule heat from the detector and associated electronics. We have measured a temperature profile across the detectors as a function of heating power at a coolant temperature of -10 [degrees] C and ambient temperatures of 0, -5, -10 [degrees] C and found that the temperature increase of the silicon detectors is below 3 [degrees] C for realistic power levels.

A large area (60 mm x 60 mm) n-bulk and n-strip readout silicon strip detector prototype was fabricated for the ATLAS SCT detector. Detector modules with a strip length of 12 cm were made by butting two detectors. One of the 12 cm modules was irradiated with protons to a fluence of 1.2 x [10.sup.14] p/[cm.sup.2], and a beam test was carded out for the non-irradiated and the irradiated detector modules. Efficiency and noise occupancy were analyzed using the beam test data. High efficiency was obtained for both detectors in the bias voltages down to about half the full depletion voltage. The noise occupancy was

As a part of R&D for the BELLE experiment at KEK-B[1], we measured the spatial resolution of silicon strip detectors for particles with incident angles ranging from 0 [degrees] to 75 [degrees]. These detectors have strips with pitches of 50, 125 and 250 [[micro]meter] on the ohmic side. We have obtained the incident angle dependence which agreed well with a Monte Carlo simulation. The resolution was found to be 11 [[micro]meter] for normal incidence with a pitch of 50 [[micro]meter], and 29 [[micro]meter] for incident angle of 75 [degrees] with a pitch of 250 [[micro]meter].

We have developed a SPICE model for the ohmic side of AC-coupled Si microstrip detectors where interstrip isolation is obtained via field plates. The interstrip isolation has been measured in various conditions by varying the field plate voltage. Simulations have been compared with experimental data in order to determine the values of the SPICE parameters for different voltages applied to the field plates. The model is able to predict correctly the frequency dependence of the coupling between adjacent strips. Furthermore, we have used such model for the study of the signal propagation along the detector when a current signal is injected in a strip. Only electrical coupling is considered here, without any contribution due to charge sharing derived from carrier diffusion. For this purpose, the AC pads of the strips have been connected to a read-out electronics and the current signal has been injected into a DC pad. Good agreement between measurements and simulations has been reached for the central strip and the first neighbours. Experimental tests and computer simulations have been performed for four different strip and field plate layouts, in order to investigate how the detector geometry affects the parameters of the SPICE model and the signal propagation. Furthermore we have considered neutron irradiated devices to study the modifications induced by radiation on the detector performance and the impact on the SPICE model.

The alignment systems required for the PHENIX muon tracking chamber system will be described. These systems include the system that has been developed to accurately electro-etch cathode strip foils with respect to external fiducial marks to a 25 [[micro]meter] accuracy, as well as the system that is being developed to measure the relative alignment of three stations of chambers spanning 4.5 meters, to an accuracy of 25 [[micro]meter].

A description is given of the principle and experimental verification of a new method which significantly reduces broadening of the position response, due to parallax, for radiation incident on a detector at finite angles of incidence. The technique, which can be implemented in gas-based position-sensitive detectors with planar geometry, substantially improves position resolution for scattering experiments using X-rays and neutrons, and will permit larger angular coverage than has previously been possible. An improvement of nearly a factor four in rms position resolution is predicted, which is confirmed by measurements using a gas proportional X-ray detector with delay-line position encoding.

Thin Gap Gas Chambers have been proposed for an upgrade of the endcaps of the DELPHI detector at LEP. Two full size chambers have been built and a study of the optimal operating conditions has been carried out. In this paper the main construction parameters are discussed and test results will be given. Tests of the electronic readout were performed and the general feasibility of the detector is demonstrated.

We describe operating principles and results obtained with a new detector element: the Gas Electrons Multiplier (GEM) [1]. Consisting of a thin composite sheet with two metal layers separated by a thin insulator, and pierced by a regular matrix of open channels, the GEM electrode, inserted on the path of electrons in a gas detector, allows the transfer of charge with an amplification factor approaching ten. Uniform response and high rate capability are demonstrated. Coupled to another device, multiwire or micro-strip chamber, the GEM electrode permits higher gains or less critical operation; separation of the sensitive (conversion) volume and the detection volume has other advantages: a built-in delay (useful for triggering purposes), and the possibility of applying high fields on the photo-cathode of ring imaging detectors to improve efficiency. Multiple GEM grids in the same gas volume allow large amplification factors to be achieved in a succession of steps, leading to the realization of an effective gas-filled photomultiplier.

The data-acquisition system for a silicon micro-vertex detector of the BELLE experiment is described with special emphasis placed on the front-end circuits and their interfaces. The description includes a 128-channel preamplifier chip, a silicon-based double-sided hybrid card, and readout control chips for digital interfaces, which are discussed in terms of both architectural and operational aspects.

We present a read-out system for large mass bolometric detectors. It consists of a differential voltage-sensitive preamplifier having very low series and parallel noise, thanks to the use of a pair of selected silicon JFET at the input. Very good electrical characteristics were obtained. DC power supply rejection ratio, thermal drift of input offset voltage and input leakage current are minimized by using novel circuit solutions. Parallel noise behavior was also investigated using a new measurement method which is sensitive to the shot noise generated by very small currents (tens of fA).

A new testing method has been used to evaluate the dynamic performance of several digitizing systems used in nuclear physics experiments. This method is useful for characterizing high resolution analog to digital converters when a pure enough signal source is not available. The signal to noise ratio of the digitizer, excluding the harmonic components from the noise, is the computed parameter. The use of this method to estimate jitter errors is also studied.

We describe a test chip for a pipelined, multihit time to digital converter, capable of operating in common stop mode, with double hit resolution of approximately 10 nanoseconds, maximum time range of 10 microseconds and least count of 50 picoseconds. This is constructed with a standard CMOS process using a novel application of the Vernier principle. The test chip demonstrates this application. This device has many potential applications in high energy and nuclear physics experiments, as well as other fields of research. We present results of measurements made on this test chip.

The track segment finding subsystem of the LEVEL 2 trigger in the CLAS detector has been designed and prototyped. Track segments will be found in the 35,076 wires of the drift chambers using a massively parallel array of 768 Xilinx XC-4005 FPGA's. These FPGA's are located on daughter cards attached to the front-end boards distributed around the detector. Each chip is responsible for finding tracks passing through a 4 x 6 slice of an axial superlayer, and reports two |segment found' bits, one for each pair of cells. The algorithm used finds segments even when one or two layers or cells along the track is missing (this number is programmable), while being highly resistant to false segments arising from noise hits. Adjacent chips share data to find tracks crossing cell and board boundaries. For maximum speed, fully combinatorial logic is used inside each chip, with the result that all segments in the detector are found within 150 ns. Segment collection boards gather track segments from each axial superlayer and pass them via a high speed link to the segment linking subsystem in an, additional 400 ns for typical events. The Xilinx chips are rambased and therefore reprogrammable, allowing for future upgrades and algorithm enhancements.

The design of a hardware pipelined system to collect, monitor and broadcast the trigger information in the CERN NA48 high-rate experiment is presented. Implementation details for a VME-based prototype are also presented. The solutions adopted to cope with a synchronous-asynchronous pipelined trigger system are discussed.

To replace an aging multiplexed observation system of about 1500 analog signals originating in 5 machines of the CERN PS Complex a VXI system consisting of various oscilloscope and multiplexer modules housed in local crates has been developed. The crates are lined together via an Ethernet cable and run under die VxWorks operating system. An on-line Oracle database allows for reference signals and observation settings to be saved and The signals are displayed within one hundred milliseconds after their capture on virtual oscilloscopes on DEC workstations; these oscilloscopes include all the usual feature plus the specific controls related to the CERN context Two system accessing about a hundred signals, were delivered in March 1993 as a part of the PS-SL controls rejuvenation project This paper will describe the hardware and software structure, the current and expected performances as well as costs and forecasts.

A portable swipe monitor has been developed at Los Alamos National Laboratory for the state of New Mexico. The monitor is intended for use at New Mexico ports of entry to detect removable alpha contamination on shipments bound for the proposed Waste Isolation Pilot Plant (WIPP). This detector uses Long-Range Alpha Detection (LRAD) technology, which monitors ionization created by alpha interactions with ambient air, and includes built-in background subtraction of external radiation fields and radon. Its sensitivity to alpha contamination is on the order of a few tens of dpm, and it has excellent linearity throughout its entire range of over 200,000 counts per minute of alpha activity. Designed to be used primarily in the field, the monitor is battery-powered and can operate continuously for up to five hours.

The calorimetry subsystem of the CLEO-III Trigger incorporates both analog and digital electronics to provide pipelined trigger information every 42 ns with a latency of approximately 2.5 [micro]s. This paper describes the pipelined signal-processing and pattern-recognition schemes used to provide calorimeter information to the experiment trigger, with somewhat greater emphasis on the analog components of the system. Analog processing is employed to address the quantization error caused by split energy deposition in adjacent calorimeter cells, and digital field programmable gate arrays are used extensively to filter and categorize the calorimeter energy topology. Timing, geographical, and energy information are all available for use in the calorimeter trigger. Index Terms--Analog processing circuits, calorimetry, digital circuits, field-programmable gate arrays (FPGAs), trigger circuits.

Nuclear weapons effects simulators are large pulsed power installations for generating intense X-ray fluxes, which are used in testing the susceptibility and hardening of military systems in the radiation environment. These simulators are expensive because of the requirement for very high power pulsed output to drive radiation loads that can provide the necessary spectral and temporal characteristics of the X-radiation suitable for testing. During the past 40 years, the Defense Threat Reduction Agency, as well as the Sandia National Laboratory, have developed pulsed power components and systems and improvements in load designs that are lowering the cost of X-ray simulators. The progress in pulsed power development encompasses elements from energy-storing capacitor banks to radiation loads and has led to the formulation of concept studies for future simulators with fewer components and with radiation loads as an integral part of the system. Index Terms--Bremsstrahlung, capacitive energy storage, electron beam transport, high-current switches, high-voltage insulators, inductive energy storage, intense X-rays, linear transformer devices, nuclear weapons effects simulators, opening switches, plasma K-line radiators, power conditioning, pulse forming lines, pulse generators.

A digital pulse shape discrimination data acquisition system developed around a commercially available digital oscilloscope card and data acquisition software was applied to on-line measurement of aqueous radioactivity. The radiation detector consisted of a flow-cell containing granular 63-90 [micro]m Parylene C polymer coated CsI:T1 and two Burle 8850 photomultiplier tubes (PMTs). The PMTs were connected in a software-realized coincidence mode. The digital data acquisition system consisted of a GaGe CompuScope 8012A/PCI 12-bit dual channel 50-MHz digital oscilloscope card residing in a Pentium III 450 MHz personal computer and the LabVIEW 5.1 data acquisition software. The individual fast anode scintillation pulses were digitized by the oscilloscope, transferred from on-board memory to computer memory, and processed by the LabVIEW Virtual Instrument (VI). The coincidence resolving time was set to 100 ns to reduce background. Pulse shape discrimination was achieved using the charge integration technique. Experimental tests showed very good pulse shape discrimination between alpha and beta particles. The alpha to beta and beta to alpha spillover were 2.1% and 1.4% respectively for single parameter analysis which yielded a figure of merit (FOM) of 1.65. The alpha to beta and beta to alpha spillover for dual parameter analysis were 1.3% and 0.3%, respectively. The alpha and beta detection efficiencies for the CsI:T1 flow-cell were typically 31% and 26%, respectively. This system was also coupled to a liquid ion chromatography system for online measurement of aqueous alpha/beta radionuclides.

The study of the performance and application of [PbWO.sub.4] in electromagnetic calorimetry at energies far below 1GeV has been continued. The significantly improved optical and scintillation properties of 15cm long Nb/La-doped crystals, optimized for the ECAL/CMS calorimeter, are documented. The lineshape, energy and time response of a 5x5 matrix are tested with monoenergetic photons up to 790MeV energy and compared to previous measurements. First attempts have been made to enhance the scintillation yield by suitable dopants (Mo, Tb) for applications at very low photon energies. As a first large scale project at medium energies, the proposed concept for a compact photon spectrometer to be implemented into the ANKE magnetic spectrometer at COSY (KFA Julich) is illustrated.

We have built and tested a mixed signal Multi-Chip Module (MCM) to be used in the Level-1 Pre-Processor system for the Calorimeter Trigger of the ATLAS experiment at CERN. The MCM performs high speed digital signal processing on four analogue input signals. Results are transmitted serially at a serial data rate of 800 MBd. Nine chips of different technologies are mounted on a four layer copper substrate. Analogue-to-digital converters and serialiser chips are the major consumers of electrical power on the MCM, which amounts to 9 Watts for all dies. Special cut-out areas are used to dissipate heat directly to the copper substrate. In this paper we report on design criteria, chosen MCM technology for substrate and die mounting, experiences with the MCM operation and measurement results.

Insensitive regions and spatial resolution of microdot (MDOT) detectors with 200 [micro]m, 100 [micro]m, and 50 [micro]m pitches were studied using a collimated X-ray beam as a probe. The insensitive zones measured by the probe-scan were compared with those estimated previously from the drift-voltage dependence of the count rate. The X-ray probe-scan along an isolated strip of a ganged linear array of MDOT anodes showed significant insensitive zones in the 200 [micro]m pitch detector, but less in the 100 [micro]m and 50 [micro]m pitch detectors. The measured projected spatial widths were 80 [micro]m, 55 [micro]m, and 51 [micro]m for the 200 [micro]m, 100 [micro]m, and 50 [micro]m pitch detectors, respectively. Together with gain measurements made previously, these results suggest that 100 [micro]m is the optimum pitch to minimize insensitive zones, and achieve good spatial resolution and good avalanche gain.

A technique for electron multiplication in gas is being developed that lends itself to easy fabrication with lithographic techniques. Moreover, the electrical characteristics of the resulting device are very stable. Comprising an array of anode pins that are recessed into a closely-packed hexagonal cathode, the MIcro-Pin Array (MIPA) represents a promising addition to micro-pattern detectors with two-dimensional sensitivity. We describe the electrostatic rationale behind the MIPA detectors design, the fabrication technique, and X-ray results that illustrate electrical stability and high rate capability.

We discuss the main operating features of GEM detectors, optimized for use as trackers in a high radiation environment. The construction, tests and performances of large prototypes for the COMPASS experiment are also described, as well as the results of an exposure to very high intensity beams.

This paper describes an implementation of the High Level Trigger for the ATLAS experiment on a large testbed based on a 32 node switch attached to data sources, event processing units, supervisor modules and interface to the level-1 trigger. We describe the functionality of each component and investigate its performance. We show the operation of the complete system and outline future activities.

The higher order modes (HOM's) of RF cavities at the Pohang Light Source (PLS) storage ring cause longitudinal coupled bunch mode instabilities. A longitudinal feedback system (LFS) is introduced to cure these instabilities. As a key component of the LFS, there is a single-ridged waveguide-overloaded cavity for a longitudinal bunch-by-bunch LFS kicker. Since the RF frequency is 500 MHz, the bandwidth of this kicker should be wider than 250 MHz to damp any coupled bunch mode. Also, a higher shunt impedance of the kicker is selected to use a lower power amplifier. One aluminum kicker is fabricated, and the bandwidth, HOM's, and the shunt impedance of the kicker are measured with a network analyzer. This kicker has several different features from the LFS kicker for DA[Phi]NE. First of all, there are four input/output ports to obtain a wider bandwidth. Second, there is a nose cone to obtain a higher shunt impedance. Third, the symmetric frequency response of the shunt impedance around the central frequency is provided. Finally, low beam coupling power (4 W per port at 400 mA) is obtained. According to the simulation result by the high frequency structure simulator (HFSS) code, we obtained the high shunt impedance of 621.3 [Omega] (transit time factor considered value) and the wide bandwidth of 255 MHz. These are compared with the measured result. Index Terms--Coupled bunch mode instability, HOM, longitudinal feedback system kicker, wake fields.

The level-3 trigger system of the STAR experiment will in the final stage consist of a farm of 24 ALPHA/Linux processors, interconnected by SCI (Scalable Coherent Interface). The system will perform online tracking of [N.sub.track] [is greater than or equal to] 8000 tracks per event ([N.sub.point] [is less than or equal to] 45 per track). The track data will be transfered to a global level-3 CPU (expected data transfer rate [nearly equals to] 48 MB/s), performing online event analysis tasks (e.g. invariant mass reconstruction) with a design trigger input rate of R=100 Hz (R=20 Hz for a prototype system).

The BABAR experiment is the particle detector at the new PEP-II 'B factory' facility at the Stanford Linear Accelerator Center. The experiment's goal is the detailed study of the fundamental phenomenon of charge-parity symmetry breaking in B meson decays. Within the detector's online system, the Online Event Processing subsystem (OEP) provides for near-real-time processing of data delivered from the data acquisition system's event builder. It supports four principal tasks: the final 'Level 3' software trigger, rapid-feedback fast data quality monitoring, event displays, and the final steps in online calibrations. In order to accommodate the expected 2000 Hz event rate, the OEP subsystem runs on computing farm, receiving events on each node into a shared memory buffer system and coordinating processes performing analysis on this data. Analysis tasks are assigned administrative characteristics providing for protection of data integrity and ensuring that indispensable tasks such as triggering are always performed, while others are performed as resources are available. Data access in OEP uses BABAR's standard offline reconstruction framework. Thus, notably, the Level 3 trigger software could be developed in advance in the full offline simulation environment and yet be suitable for near-real-time online running with minimal changes.

We report on soft error rates (SER) of 16 and 64 Mb dynamic memory chips induced by three types of elementary particles, neutrons, protons, and pions, with emphasis on results obtained with pion beams. Significant SER differences, up to a factor 1000, are seen between various manufacturers and cell technologies. We discuss reaction mechanisms and by comparing SER rates to nuclear reaction cross sections present guidelines for predicting failure rates. Index Terms--Circuit design, hadrons, memory testing, particle beams, soft error rates.

The trigger system for the AUGER fluorescence detector is presented. The main goals of the design were low price, high flexibility of the trigger, high reliability and good testability even at remote operation. The simulation of all boards using VHDL tools resulted in a design which is optimal with respect to these goals and highly independent of existing commercial solutions. The large-scale integration of modern FPGAs yielded a massive parallel system for recognition of tracks and suppression of background. The readout and control is carried out by low cost PCs under LINUX operating system.

Dynacore's project objective is to provide tools for the remote operation of large physics experiments [1], [2], [3]. Our group concentrates on the design of a remote control room for the Textor-94 plasma-physics experiment [4], which will incorporate a distributed database, accessible with Corba over Internet [5]. During normal operation, Textor-94 generates hundreds of megabytes of measurement data within the few seconds of a plasma shot. These shots are generated approximately every ten minutes, in total around 30 shots per day. A database will make the measurement data available to scientists, both at the experiment and in remote control rooms, which can be located in different countries, within minutes after the acquisition. Fast access to the database and a high data-storage speed are, therefore, of utmost importance. We have defined a set of data objects, which allow storage of virtually any kind of Textor measurement data. An Objectivity [6] object-oriented database is used to store these objects. We have also designed an architecture in which database clients access the measurement database solely via data managers with Corba-interfaces. The data managers introduce security into the database, and also hide details of the database implementation for the users. An object manager provides clients with a central starting point, and distributes the total database load in an intelligent way over all available data-managers. We have tested the performance of our architecture on a computer cluster consisting of eight Sun-Ultra-10 workstations [7]. These computers are interconnected with 1-gigabit/s Ethernet fiber-optic links over two Cabletron SSR-8000 IP routers, and also via 100 Mbit/s Ethernet UTP cables over one Cabletron SS-6000 [8]. This article will discuss the database model, the distributed architecture and the performance measurements results. The measurements showed that our high performance goals can be achieved with the presented architecture.

The D?? level 1 tracking trigger uses data from the scintillating fiber tracker, the central and forward preshower detectors, the muon system and the calorimeter. Tracks are found in the scintillating fiber tracker with transverse momentum greater than 1.5 GeV/c. The tracks are matched with hits in the central preshower detector and the muon system for electron and muon tagging. Preshower clusters are also used for identifying photon candidates. These multi detector triggers are then sent to the level 1 Trigger Framework where they are further combined with the calorimeter to create the final level 1 trigger. This paper presents an overview of the level 1 trigger system with emphasis on the use of large programmable logic devices (PLD's) in an extensible system architecture that allows complex, multi detector triggers.

The multi-level trigger system of the HADES detector is intended to select dilepton events in heavy ion reactions and to reduce minimum bias event rates by a factor of 10000. After a first level trigger based on the charged particle multiplicity the second level is the essential dilepton trigger which identifies electron signatures in several detector components and correlates them to dileptons. The third level trigger verifies these dilepton candidates using additional tracking information. Programmable logic devices, FPGAs and CPLDs, as well as DSPs are widely used to gain flexibility and provide common solutions. The trigger concept and fully functional versions of the readout boards and the second level trigger are presented.

The paper presents the design of piecewise constant periodic output feedback control for a discrete-time singularly perturbed system resulting from the discretization of a continuous-time standard singularly perturbed system. By a suitable linear transformation of state variables, the given continuous-time singularly perturbed model is converted into a block triangular form in which the fast subsystem is decoupled. The discrete-time model corresponding to the transformed model also exhibits a two time scale property if sampling period is larger than the parameter [Epsilon]. Now, an output injection matrix is found that stabilizes the slow subsystem. The periodic output feedback gain is then calculated only for the slow subsystem and the same for the fast subsystem is set equal to zero. Finally the periodic output feedback gain for the composite system is obtained using the periodic output feedback gains computed separately for the slow and fast subsystems. An approach has been suggested whereby the determination of periodic output feedback gain for the slow subsystem can be converted into an optimization problem. By minimization of the suggested performance index the closed loop system behavior is improved. The method has been applied to a large pressurized heavy water reactor (PHWR) for control of xenon-induced spatial oscillations. A particular grouping of state variables has been suggested for obtaining the model in standard singularly perturbed form. The periodic output feedback gain is then calculated. The efficacy of control has been demonstrated by simulation of transient behavior of the nonlinear model of the PHWR. Index Terms--Large heavy water reactor, periodic output feedback, singular perturbation, spatial control.

Recent studies of the level-two (LVL2) trigger of the ATLAS detector [1] show that it will be possible to run the trigger algorithms at high luminosity with a reasonable number of general-purpose processors, using a sequential selection scheme and guidance from the Region-of-Interest (RoI) provided by the LVL1 trigger. The computing power requirements for B-physics, which is studied at low luminosity, are much greater than those at high-luminosity as there is no LVL1-guidance for the track finding algorithms. Instead, track finding is performed for the entire Inner Detector volume. Currently, 2500 commodity CPUs would be required to supply the necessary computing power for the B-physics trigger. We describe a system of only 200 computing nodes which would be capable of performing the B-physics triggering(2). Each of these nodes is made up of a commodity PC and a FPGA co-processor board. Each node processes an entire event. The different tasks are allocated to the appropriate hardware device (CPU or FPGA). Track reconstruction requires a variety of different steps, some of which are suited to parallel processing, whereas others require sequential execution. For some tasks, floating-point arithmetic is needed. The flexibility of the PC/FPGA combination meets these varied requirements well.

We describe the architecture of the Data Archiving and Distribution of the Virgo antenna for gravitational wave detection. The main characteristic of this system is the modularity of the architecture. This solution allows system upgrades without dramatic changes of the hardware and software components. The main performances are: 1. maximum sustained data flow of 10 Mbyte/s on DLT tapes (35/70 Gbyte) for the raw data archiving; 2. up to 1 Tbyte data archiving capacity on disk at a maximum sustained data flow of 25 Mbyte/s for the on-line data distribution; 3. up to 10 Mbyte/s data retrieval flow for the on-line data distribution. The basic architecture of the system consists of two sections: an acquisition and storage section and a data management section. The former is a LynxOS based system with the disks directly connected to the CPU slave boards, the latter is a DEC-Unix Alpha Server (Data Server), NFS mounting the LynxOS disks through a Fast Ethernet network.

The NA48 experiment at the CERN SPS aims to measure the parameter Re([Epsilon]'/[Epsilon]) of direct CP violation in the neutral kaon system with an accuracy of 2 x [10.sup.-4]. Based on the requirements of: * high event rates (up to 10 kHz) with negligible dead time * support for a variety of detectors with very wide variation in the number of readout channels * data rates of up to 150 MByte/s sustained over the beam burst * level-3 filtering and remote data logging in the CERN computer center the collaboration has designed and built a modular pipelined data flow system with 40 MHz sampling rate. The architecture combines custom-designed components with commercially available hardware for cost effectiveness and flexibility. To increase the available data bandwidth and to add filtering and monitoring capabilities, the original custom-built event builder hardware has been replaced by a farm of 24 Intel PentiumII based PCs running the Linux operating system during the shutdown between the 1997 and 1998 data taking periods. During the data taking period 1998 the system has been successfully operated taking ca. 70 Terabyte of data.