Pierre Auger Observatory Overview of the Acquisition Systems Cyril Lachaud for the Auger Collaboration LPCC/CDF 11 place Marcelin Berthelot 75231 Paris Cedex 05 (FRANCE) Phone: (+33)1 44 27 15 20 Fax: (+33)1 43 54 69 89 Email: cyril.lachaud@in2p3.fr Abstract The Pierre Auger Observatory is the next generation ground based cosmic ray detector fully devoted to the study of the ultra-high energy cosmic rays (UHECR) with energies above 10 18 ev. Two detection technics are used jointly to detect the air showers generated in the atmosphere, a ground array of water Čerenkov tanks and fluorescence telescopes. The size of the observatory (3000 km 2 ) implies a strong autonomy of all the detectors (using radio telecommunications for data transfer and GPS for time synchronization) coordinated by a central acquisition center. We will review the main features of Auger, focusing on the data acquisition systems, local (at the level of each water tank and fluorescence telescope) and central (how the local triggers are combined to detect the air showers). I. INTRODUCTION The Pierre Auger Observatory [1] is the next generation ground based detector fully devoted to the study of the ultrahigh energy cosmic rays (UHECR). The Observatory has been conceived to measure the flux, arrival direction distribution and mass composition of cosmic rays over the whole sky from 10 18 ev to the very highest energies with high statistical significance. Recent reviews of the physics of UHECR are available [2], [3]. Two sites, one in each hemisphere, of 3000 km 2 of instrumented area will study the Extensive Air Showers (EAS) generated by the UHECR in the atmosphere. The design calls for an hybrid detection of the EAS : Surface Detector (SD) : 1600 water Čerenkov detectors (arranged on a 1.5 km side triangular grid) sample the shower signal at the ground level Fluorescence Detector (FD) : 4 optical stations, each containing 6 telescopes, measure the emission from atmospheric nitrogen, which is excited by the charged particles of the shower as they traverse the atmosphere Presently the southern hemisphere Observatory is being built in Argentina (figure 1). The first phase of the project, the construction and operation of a prototype system, known as the Engineering Array (32 Čerenkov detectors and 2 fluorescence telescopes, figure 2), has now been completed [4]. This has allowed all of the subsystems that will be used in the full instrument to be tested under field conditions. Fig. 1. Layout of the southern Observatory, where 1600 surface detectors will be deployed in the dotted area. The four telescope sites are shown. The lines encompass 30 angles and define the azimuth acceptance of each telescope. Fig. 2. The Engineering Array. The in-filled stations (stations located between two regular stations) are in magenta.
II. THE FLUORESCENCE DETECTOR (FD) A. System overview Figure 3 shows a schematic view of a fluorescence telescope unit [5]. An array of 20 22 hexagonal photomultiplier tubes (the camera) is mounted on a quasi-spherical support located at the focal surface of a segmented mirror. Each PMT overlooks a region of the sky of 1.5 deg 2. The telescope aperture has a diameter of 2.20 m and features an optical filter (MUG-6) to select fluorescence photons in the range 300-400 nm. A Schmidt corrector ring allows to double the collection area without increasing the effects of optical aberrations. The above Schmidt geometry results in a 30 30 field of view for each telescope unit. The final design envisages four eyes of six telescopes. Shower Track Aperture Fig. 4. Searched patterns to fire the Second Level Trigger. TABLE I FLUORESCENCE TRIGGERS ALGORITHMS AND RATE. Trigger level Reduction by Rate (Hz) T1 boxcar running sum and threshold 100 Hz/pixel T2 pattern search 0.3 Hz/pixel T3 basic filters 8 10 4 Hz A. General description III. THE SURFACE DETECTOR (SD) A Local Station (LS) (SD unit) can be seen in figure 5 and a corresponding schematic diagram is shown in figure 6. Camera Corrector Ring Segmented Mirror Fig. 5. View of a Local Station (LS). Fig. 3. Schematic view of a fluorescence telescope unit. B. Acquisition The current signal coming from each phototube is sampled at 10 MHz with 12 bit resolution and 15 bit dynamic range. The First Level Trigger system performs a boxcar running sum of ten samples. When the sum exceeds a threshold, the trigger is fired. This threshold is determined by the trigger rate itself, and is regulated to keep it close to 100 Hz. Every microsecond, the camera is scanned for patterns of fired pixels that are consistent with a track induced by the shower s fluorescence light (figure 4). This is the Second Level Trigger, it was running in the Engineering Array run at a rate of 0.3 Hz and was dominated by muons hitting directly the camera and random noise. These components, as well as lightnings, are then filtered by the sotware Third Level Trigger (looking for time ordered long tracks), yielding a rate of 8 10 4 Hz (one event every 20 minutes). Each Third Level Trigger is sent to the Central Data Acquisition System (see V) to form an hybrid trigger. Solar panels Phototube 12 m³ of water Fig. 6. Comms and GPS Antennas Electronic box Schematic view of a Local Station (LS). Tyvek Liner Batteries Each unit consists of a 3.6 m diameter, 1.55 m high rotationally moulded polyethylene tank, enclosing a liner filled with 12000 litres of very pure water. Three 8 or 9 photomultiplier tubes view the water through windows in the liner. These are used to detect Čerenkov light produced by particles crossing the tank. Signals from the
photomultipliers are extracted from the last dynode and from the anode, with an amplification of 32 applied on the dynode signal to match the dynamic range and are read by the electronics mounted locally at the station. Power is provided by batteries connected to two solar panels, and time synchronisation relies on a commercial GPS receiver using a technique described in [6]. A specially designed radio system is used to provide communication with the Central Data Acquisition System (see V). Each tank forms an autonomous unit, recording signals from the ambient cosmic ray flux, independent of the signals registered by any other tanks in the SD array.. B. Front-end electronics and first level trigger The readout of the the 6 signals from each tank (the signals from the anode and amplified dynode of each of the three photomultiplier) is accomplished using front-end electronics with six 10 bit FADCs (Fast Analog to Digital Converter) running at 40 MHz. The digitized signals are sent to a daughter PLD (Programmable Logic Device) board where first triggering decisions are made. In the EA, we use two different first trigger modes : Threshold : three-fold coincidence of 1.75 VEM (Vertical Equivalent Muon, see III-D). The time of the trigger is given by the bin that goes beyond the threshold. Time Over Threshold (TOT) : two-fold coincidence of 12 bins at 0.2 VEM within a 3 µs window. The time of the trigger is given by the bin that fires the trigger (the 12 th bin). Whenever a trigger (T1) is realised, a time span of 19.2 µs (768 bins) of FADC are copied to a buffer (256 pre-trigger bins and 512 post-trigger bins) which can be accessed by the Station Controller. Figure 7 shows an example of such FADC traces obtained with the different trigger modes. The TOT and Threshold trigger algorithms can be changed from CDAS. TOT triggers are driven by physics and occur at the level of 1 Hz. Threshold triggers parameters are choosen to keep the total first level trigger rate close to 100 Hz. C. Station Controller, second level trigger and GPS time The local electronic is controlled by a CPU board hosting a Power PC 403GCX at 40 MHz running under the OS9 operating system. The Station Controller is used to select from the T1 triggers, the most interesting ones and to send them to CDAS. For the EA, all TOT triggers are promoted as second level triggers (T2). The Threshold T1 are designated as T2 if the threshold energy is greater than 3.2 VEM (to keep the second level trigger rate close to 20 Hz). The time at which the LS triggers is crucial for determining the shower direction. It is measured at each LS using a commercial Motorola GPS board (OnCore UT). The signal is processed by the CPU board and a Time Tagging board which provides the event time with a precision of 8 ns (measured both in the laboratory and in the field). Fig. 7. FADC traces obtained for each photomultiplier from Threshold and TOT triggers. D. Calibration All the LS are constantly calibrated with the omni-present flux of atmospheric muons. Important calibration parameters can be extracted with this method [7], in particular the conversion between ADC counts and 1 VEM (Vertical Equivalent Muon), where 1 VEM is the signal associated with a centered and vertical travelling high energy muon that crosses a tank. IV. THE COMMUNICATION SYSTEM The Auger data communication system [8] consists of two integrated radio networks organised as a 2-layer hierarchy. A microwave backbone network of high capacity supports communication from the fluorescence detector sites. It uses a standard 34 Mbps telecommunication architecture based on commercially available microwave point-to-point equipment operating in the 7 GHz band. The microwave backbone provides also a series of distributed concentration nodes for data coming to and from LS via the Surface Detector wireless Local Area Network (LAN). The wireless LAN operates in the 902-928 MHz band (ISM). It is achieved in a manner similar to a cellular telephone system, whereby the area containing the detectors is divided into a number of sectors ( 60 LS by sector), communications within each sector are co-coordinated by a base station. LS send/receive data to/from the base station one time per second, the time synchronisation within each sector is achieved
using the 1 pps timing mark derived from the GPS units. The physics data exchange for each LS is limited to 1200 bits s 1. This limitation in the data exchange is responsible for most of the features of the data acquisition for the Surface Detector. V. THE CENTRAL DATA ACQUISITION SYSTEM (CDAS) A. Overview The Central Data Acquisition System (CDAS) was designed to form the third level trigger from the different local detector triggers, to allow control of the surface detector units and to organize the storage of data. The CDAS system runs on a network of 6 servers, each equipped with two 450 MHz Intel processors and 256 MB of memory. A RAID-5 disk storage system of 500 GB forms the central storage facility. The data that are sent by the LS to CDAS belong to several streams : LS T2 (high priority) : it contains time stamps and the type of trigger (Threshold or TOT). It is forwarded to the Central trigger (Ct) application Control (high priority) : it concerns configuration control of the LS from the CDAS (software downloads are possible, enabling upgrades of the LS acquisition software) Shower data and calibration (normal priority) : it corresponds to FADC traces and calibration informations sent to CDAS when a T3 event has been fired by Ct (see V- B). A specific application (Event builder : Eb) build the definitive event data by combining those data. Calibration and monitoring (low priority) : LS are sending monitoring and calibration informations every 6 minutes (forwarded to and stored by the Monitoring recorder (Mr) application) (n = 4) and the time requirement describe above Once a trigger has been identified a message requesting that all FADC trace information corresponding to the selected T2 is built by Ct and sent by the CDAS. In the EA, the T3 request is sent to all the stations in the array. FADC traces corresponding to a T1, found in a 30 µs window centered on the average time of the selected T2, will be returned (for the full array this request will be made only to nearby stations). Moreover Ct uses 2 other algorithms for specific studies : Random : to record random coincidences over the full array every 30 minutes C-M : 2-fold coincidence within 1 µs of the two neighbouring tanks named Carmen and Miranda. These occur at 0.8 Hz and are scaled to 0.0017 Hz for transmission. With the arrangements described, the total trigger rate during the first 8 months of 2002 was, on average, a little less than 0.01 Hz and about 200 000 SD events were recorded. TABLE II SURFACE DETECTOR TRIGGERS ALGORITHMS AND RATE. Trigger level Reduction by Rate (Hz) T1 Threshold > 1.75 VEM 100 Hz/LS TOT 1 Hz/LS T2 Threshold > 3.2 VEM 20 Hz/LS TOT 1 Hz/LS T3 Central trigger (Ct) 0.01 Hz (EA) The Fluorescence Detector send only one data type : FD T3 (high priority) : it contains the necessary informations for the Fd application to build a T3 trigger for the Surface Detector (see V-C) The primary role of the CDAS is to combine local triggers information from the SD stations to form the central trigger. B. The Central Trigger The Central trigger (Ct) combines the T2 sent by the LS each second to form the T3 trigger. The algorithm used by Ct to form a T3 trigger is the following : 1) order all the T2 in time 2) look for a 3-fold coincidence within the first 2 hexagons centered on the central-t2 tank with a time difference with respect to this central station of less than (6 + 5n) µs, where n is the hexagon number, where n = 0 defines the central detector of the set 3) once the 3-fold coincidence is established, Ct adds T2 from other LS if they are within the first 4 hexagons C. The Hybrid Trigger When a shower has been recorded by the Fluorescence Detector, the corresponding T3 is sent to the CDAS. The time of impact of the shower at a ground position in the region of the EA is computed by the Fd application and a corresponding SD-event T3 is constructed. All FADC traces recorded within 20 µs of the computed time are assembled and stored in a unique event, together with the identification of the corresponding FD T3 trigger. Data from those triggers form the hybrid data set and are merged and analyzed off-line. VI. CONCLUSION In this paper we have presented an overview of the different acquisition systems used in the prototype of the Auger observatory : the Engineering Array. The next phase of Auger consists in a pre-production array of 100 additional tanks and 2 6 telescope fluorescence detectors will be operational during 2003 with completion of the full instrument at the Southern site expected in 2005.
SD FD LS T2 FD T3 Ct Data Fd CDAS Eb LS T2 FD T3 SD T3 Data Fig. 8. CDAS input and output data streams and their related CDAS applications. REFERENCES [1] Pierre Auger Project Design Report. Fermi Laboratory, 1997. [2] M. Nagano and A. A. Watson, Observations and implications of the ultrahigh-energy cosmic rays, Rev. Mod. Phys., vol. 72, pp. 689 732, 2000. [3] X. Bertou, M. Boratav, and A. Letessier-Selvon, Physics of extremely high energy cosmic rays, Int. J. Mod. Phys., vol. A15, pp. 2181 2224, 2000. [4] Properties and performance of the prototype instrument for the pierre auger observatory, Nucl. Instrum. Meth., to be published. [5] S. Argiro, Performance of the pierre auger fluorescence detector and analysis of well reconstructed events, in Proc. 28th International Cosmic Ray Conference, Tsukuba, Japan, Aug. 2003. [6] C. L. Pryke and J. Lloyd-Evans, A high performance gps based autonomous event time tagging system with application in a next generation extensive air shower array, Nucl. Instrum. Meth., vol. A354, pp. 560 566, 1995. [7] X. Bertou, Calibration and monitoring of the pierre auger surface detectors, in Proc. 28th International Cosmic Ray Conference, Tsukuba, Japan, Aug. 2003. [8] P. Clark and D. Nitz, Communications in the auger observatory, in Proc. 27th International Cosmic Ray Conference, Hamburg, Germany, Aug. 2001. ACKNOWLEDGMENT The author would like to thank all the members of the Pierre Auger Collaboration, for all the hard work done to build the most important cosmic ray detector in the world. In particular I would like to thank Murat Boratav for giving me the opportunity to give this talk and Antoine Letessier-Selvon for long and friendly discussions.