E. Costa1, F. Frontera2,3, D. Dal Fiume3, L. Amati1, M.N. Cinti1, P. Collina2, M. Feroci1, L. Nicastro3,

M. Orlandini3, E. Palazzi3, M. Rapisarda4, and G. Zavattini2.


1 Istituto di Astrofisica Spaziale, CNR, Via E.Fermi 21, I-00044 Frascati, Italy

2 Dipartimento di Fisica, Universita' di Ferrara, Via Paradiso 12, I-44100 Ferrara, Italy

3 Istituto Tecnologie e Studio Radiazioni Extraterrestri, CNR, Via P.Gobetti 101, I-40129 Bologna, Italy

4 Divisione di Neutronica, C.R.E., ENEA, Via E.Fermi 27, I-00044 Frascati, Italy




The four Lateral Shields of the PDS experiment aboard the Beppo-SAX satellite are equipped with dedicated electronics and used as a Gamma Ray Burst Monitor (GRBM). On a suitable trigger time histories are recorded and transmitted. The large geometric area and the low background of these detectors provide a remarkable capability which is limited, on the other hand, by the no uniform exposure to different direction, due to the shielding of the other experiments of SAX. Here we discuss the on-ground calibrations and the performance simulations in progress, aimed to reconstruct the cosmic X-ray fluxes impinging on the satellite from the recorded counting rates.



SAX (Satellite di Astronomia X) is a joint venture of Italian and Dutch Space Agencies (ASI and NIVR) for a broad band (0.1-300 keV) study of the X-ray sky (Boella et al. 1996). The high energy part of this band is covered by the Phoswich Detection System (PDS) designed by ITESRE-CNR, IAS-CNR and University of Ferrara. It consists (Frontera et al. 1996) of four Phoswich detectors [NaI(Tl)/CsI(Na)], with a pair of passive collimators that can be rocked in order to monitor the background while observing a source. The whole assembly is surrounded by four active shields, disposed as a square well with the goal of reducing the background and its modulation along the orbit. Each shield is made of a large area CsI(Na) detector (about 275 x 413 x 10 mm3 size) seen by two Hamamatsu R2238 photomultipliers (PMT). For each lateral shield a light pulser, made with an Am241 a source on a NaI(Tl) crystal, is mounted amid the two PMT and used as a gain monitor during the mission. The signals from the two PMT of each shield are amplified and analogically summed before being fed into a single-channel discriminator that generates the veto signal for the main detectors and, in parallel, to a window discriminator that generates the enable signal for the ratemeter counting and the AD conversion, and possibly the selection logic of the GRBM trigger. The anticoincidence threshold (ACT) and the lower (LLT) and upper (ULT) thresholds of the GRBM function are independently programmable on each shield (ACT: 7 steps from nominal 100 to 300 keV; LLT: 15 steps from nominal 20 to 90 keV; ULT: 7 steps from nominal 200 to 600 keV) .


The GRBM Dedicated Channel

The counts from signals within the LLT and ULT of each shield are used to detect the presence of a gamma-ray burst (GRB) candidate event and trigger the production of the time history records. All the relevant parameters to determine the trigger conditions can be set from the ground station by telecommand. The count rate of each detector is accumulated on a short integration time, programmable from 4.8 to 4000 ms. The background level is continuously computed by the moving average of the counting rate on a fixed time interval, programmable from 8 to 200 s. If the counting rate exceeds, simultaneously on two shields, the level of n× s (n programmable to 4, 8 or 16 by command), the GRBM logic is triggered. The temporal profile of the GRB candidate is then stored in the onboard memory with the following structure of the time bins: 7.8 ms for the last 10 s before the trigger, 0.48 ms for the first 10 s after the trigger and 7.8 ms for the following 110 s.

The Relevant Housekeeping Data

Useful additional information is included in some housekeeping data. Ratemeters of GRB counts and anticoincidence counts for each shield are transmitted with a 1 s time resolution. Moreover the pulse height spectra of each shield in the GRBM energy band are continuously accumulated in 256 channels every 128 s.


We show in Figure 1 the relative position of the various detecting units and components of the PDS. The GRBM detectors have been conceived and designed as anticoincidence for the central phoswiches thus, for instance, their crystal frames are flat inside (since they must allow for the collimators rocking) and strengthened outside. Moreover the PDS, which is the heaviest experiment of SAX is located at the center of the satellite. As a consequence the transparency of the entrance window for Gamma Bursts is not even, but is determined from the mass distribution of the PDS experiment, and in general, of the SAX satellite. In Figure 2 we show the PDS relative position with respect to the other SAX experiments from which their shadowing effect on the GRBM can be seen. As a consequence each direction with respect to the satellite frame, and hence every direction in the sky for a certain satellite pointing, is sampled with a different effective area. This condition is far from an ideal, dedicated GRB experiment. Nevertheless the large area, the low background intensity and modulation guarantee a high throughput of information, in particular in the timing domain. Therefore we invested relevant efforts to arrive to a reasonable knowledge of the response of SAX GRBM to photons impinging from different directions.



On-ground Calibrations

Before the integration in the PDS experiment, the Lateral Shields (LS) spatial responsivities in 30 different points on the crystals surfaces were calibrated with collimated radioactive sources and standard laboratory electronics. After the integration on PDS, one of the shields was calibrated with the flight electronics and with radioactive sources located at a distance of 4 m, also in order to check the impact of thresholds and of event selection logic. After the full integration of SAX (except for the solar panels) the GRBM was calibrated with radioactive sources of previously measured activity, in the 100 mCi range, positioned at a 5 m distance. The sources were mounted in a low-backscattering holder on a graduated vertical bar. SAX was slowly (» 10 arcmin/s) rotated about its pointing axis. Lateral Shields and GRBM ratemeters were recorded, and the energy spectra were collected averaged over 20°, which is suitable to interpret the main variations in the ratemeters in terms of spectral features and to scale the ratemeter calibrations for possible changes of threshold. In this way the azimuth dependence of the GRBM efficiency was explored in detail. By suspending the sources at different heights also a set of polar angles was covered by this calibration. In Figure 3 we show the counting rate of the four GRBM counters as a function of the azimuth angle with a Co57 source positioned at the same height of the center of the shields. The count rate from each slab has its maximum when directly exposed to the source but a significant amount of side and back scattering is also present. Some of this is due to the environment of the testing hall where the calibration was performed, and was calibrated with a small detector set in the position of SAX after its removal. In Figure 3 the heavy absorption on LS #2 from the High Presssure Gas Scintillation Proportional Counter and on LS #4 from the two Wide Field Cameras and their masks is apparent. On the other hand LS #1 and #3 see the sky through the carbon fiber tubes of the MECS and LECS which are relatively transparent at high energies. One very appealing conclusion is that the two "good" shields (#1 and #3) are co-aligned with the WFCs (whose field of view is 20° x 20° and energy range 3-30 keV), giving some chance to simultaneously detect a GRB in both instruments.



In order to interpret the calibration data, correct them for the finite distance of the source and properly interpolate and extrapolate results to directions not covered from the measurements, we set up a complementary activity of Monte Carlo simulation. The code is based on the release 4.2 of the MCNP Los Alamos code (Forster et al. 1990). A large effort was devoted to include as much detail as possible in the description of SAX. Nowadays the description of SAX includes 168 objects of different materials (excluded collimator cells) and is continuously improving, also with a feed-back from the analysis of calibration data. In Figure 4 we show the simulation data describing the same situation of calibration in Figure 3. Even if the simulation is not yet complete it is already providing the major features of the calibration results. In order to handle the calibration data and have an analytical description of their trends we are also trying to fit detectors efficiencies dependence on energy and direction of incident photons with complex functional forms, partially derived from the underlying Physics, partially semiempirical.


The expected performance of the SAX GRBM was studied at the early stage of its development by Pamini et al. (1990), based on Monte Carlo simulations. The sensitivity was found to be a function of the GRB direction and intensity. Now the on-ground calibration data, the more refined Monte Carlo simulation, and the in-flight calibration data will be used to establish the real sensitivity of the GRBM. At present the cross analysis of the calibration data and of simulations is still in progress, and we can therefore only give a rough estimate of the sensitivity of the GRBM to cosmic events. Using the on-ground calibrations data and the radioactive source activity we have in a first evaluation derived the effective area of LS #3 in their normal direction to be about 185 cm2 at the Am241 energy (60 keV), 610 cm2 at the Co57 energy (122-136 keV), 760 cm2 at the Ce139 energy (166 keV), 680 cm2 at the Hg203 energy (279 keV), 630 cm2 at the Sn113 energy (392 keV) and 520 cm2 at the Cs137 energy (662 keV). The secondary low energy lines of these radioactive sources were subtracted, when not negligible. The above effective areas should be considered as a first order evaluation, since they do not include a correct estimate of the scattering effects in the calibration set-up, that will be studied in detail by means of the Monte Carlo simulation. Similar results are derived for LS #1, while for LS #2 and #4 a more detailed study of the angular dependence is in progress because of the presence of the HPGSPC and WFC, respectively, in their field of view.

From a very preliminary analysis of the in-flight data we derived the background level to be of the order of 0.7 counts cm-2 s-1 in the 50-600 keV range and an excellent stability (± 7% within one whole orbit). Since a simultaneous trigger from two shields is needed for the GRB detection, we expect that the limiting sensitivity will be driven by the shields which are performing worse.




Boella, G., R.C. Butler, G.C. Perola, L. Piro, L. Scarsi, et al., BeppoSAX, the wide band mission for X-Ray

Astronomy, Astron. & Astrophys Suppl. Ser., 122, 299 (1997).

Forster, R.A., R.C. Little, J.F. Briesmeister and J.S. Hendricks, IEEE Trans. on Nucl. Science NS-37, 1378 (1990)

Frontera, F., E. Costa, D. Dal Fiume, M. Feroci, L. Nicastro, et al., The high energy instrument PDS on-board the

BeppoSAX X-ray astronomy satellite, Astron. & Astrophys. Suppl. Ser., 122, 357 (1997).

Pamini, M., L. Natalucci, D. Dal Fiume, F. Frontera, E. Costa, et al., The Gamma-Ray Burst Monitor on Board

the SAX Satellite, Il Nuovo Cimento C, Vol. 13 N. 2, 337 (1990).