This Legacy journal article was published in Volume 4, February 1994, and has not been updated since publication. Please use the search facility above to find regularly-updated information about this topic elsewhere on the HEASARC site.

The SAS-2 and COS-B

Gamma-Ray Databases

Paul Barrett & Brendan Perry


1. Introduction

In Volume 5 of the `Annual Review of Astronomy and Astrophysics' (1967), Giovanni Fazio wrote that "... up until now, no photons of energy greater than 100 keV originating from beyond the solar system have definitely been detected." During the next few years, many groups using balloon-borne detectors reported detections of gamma-ray sources. However, such experiments were hampered by the high level of atmospheric gamma-rays due to cosmic ray interactions in the atmosphere, making the statistical significance of such sources low and their existence doubtful. Though several gamma-ray detectors were placed in orbit, the first certain detection of celestial gamma-rays came from OSO-3, which detected emission of gamma-rays with energies greater than 50 MeV from the galactic disk with a peak intensity toward the galactic center.

The first unambiguous detection of high-energy (>50 MeV) gamma-ray sources was achieved with the SAS-2 satellite. Unfortunately, the lifetime of the SAS-2 mission was cut short after about six months by the failure of a low-voltage power supply. The European Space Agency's COS-B satellite was the next major high-energy gamma-ray detector. It's lifetime was considerably longer than that of SAS-2 and was turned off after about 8 years when on-board resources were exhausted. The second COS-B catalog containing the positions of 25 sources was the most complete listing of high-energy gamma-ray sources until the recent release of the Compton/EGRET catalog (Fichtel et al. 1993).

Since the launch of the Compton Observatory in April 1991, there has been renewed interest in the SAS-2 and COS-B data. For example, the data has been used to check for long-term (of order years) variability of sources discovered in the EGRET all-sky survey and to improve the ephemeris of the Geminga pulsar (see e.g. Legacy, No.2) by extending its timeline. Therefore, HEASARC has decided to make this data available publicly and this article may be considered an announcement of its availability. The intention of this article is to give an overview of the two datasets. Details of the restoration and FITS formats for the data, and associated calibration data, can be found in other OGIP documents located on the FTP server legacy (see the article in Legacy, No.3 by Drake and O'Neel).

2. The SAS-2 Mission


The second NASA Small Astronomy Satellite (SAS-2) was dedicated to gamma-ray astronomy in the energy range above 35 MeV. The satellite carried a single telescope using a 32-level wire spark-chamber. The satellite was spin stabilized with the telescope axis along the spin axis. SAS-2 was launched on 1972 November 15 and became operational on 1972 November 19. On 1973 June 8, a failure of the low-voltage power supply ended the collection of data. During the approximately six months of the mission, 27 pointed observations, typically of a week duration, were made resulting in about 55 percent of the sky being observed, including most of the galactic plane. The field-of-view of the telescope is about 35 degrees (full width at half maximum) with an angular resolution of a few degrees. In addition to the general galactic emission, high-energy gamma-rays were also seen from the Crab and Vela pulsars (see e.g. Fichtel et al 1975).

The low fluxes involved in the study of gamma-ray sources make it desirable to minimize the background flux from cosmic-rays. Therefore a low Earth equatorial orbit was chosen having a 2 degree inclination, an apogee and perigee of 610 km and 440 km, respectively, and an orbital period of about 95 minutes. The sensitivity of the sparkchamber was noticed to decrease during the lifetime of the mission as the sparks from the gamma-ray event caused the gas to crack producing unwanted by-products. Do to the degradation of the gas and the inability to flush and replace it with new gas, the lifetime of the mission would not have been more than about 9 months, in any case. The calibration of the SAS-2 experiment was done using both the flight unit and an identical flight spare unit. The range of energy studied at the National Bureau of Standards (NBS) Synchrotron, Gaithersburg, Maryland, was approximately 20 to 114 MeV. The energy range between 200 to 1000 MeV was studied at the Deutsches Elektronen-Synchrotron (DESY), Hamburg, West Germany.

The Data

The format of the SAS-2 data is based on a ROSAT Events FITS file. Briefly, such a file contains a primary header and four extensions. No data is contained in the primary header. All data resides in the four extensions which are named ALLGTI, STDGTI, EVENTS, and TSI. The ALLGTI and STDGTI extensions contain the exposure information. The ALLGTI gives the time interval when the instrument was on and ready to receive data, whereas the STDGTI gives the time interval when data, deemed to be good by the standard processing software, were received by the instrument. Therefore, the total exposure from the STDGTI is usually less than that from the ALLGTI. The EVENTS extension contains a list of all the events for an observation, and the TSI extension is for the housekeeping data or the instrumental status.

In order to restore fully the SAS-2 data (i.e.. provide event, exposure, and housekeeping data), access to the original data was necessary. The SAS-2 event list was commonly used by researchers, but the exposure information was never made available. The original data provided us with the orbital parameters and status of the spacecraft, from which we were able to calculate the necessary exposure data and extract the auxiliary housekeeping data. Thus, for the first time the SAS-2 data has been put into a form that is compatible with that of COS-B and EGRET data. The SAS-2 data has been divided into 27 files, one for each pointing. The total number of events that satisfied the trigger logic of the spark-chamber and post-flight human and software selection was 13056.

In addition to the event data, calibration data also has been made available. There are four types of calibration files: 1) the redistribution matrix or energy dispersion data, 2) the effective area or sensitivity data, 3) the point-spread-function data, and 4) the energy boundary data. The calibration data was not available in the original data, but was gleaned from the literature (see Fichtel et al. 1975). More detailed information about these files can be obtained from the `SAS-2 Calibration Guide' in the legacy calibration database.

3. The COS-B Mission


The European Space Agency's satellite COS-B was dedicated to gamma-ray astronomy in the energy range 50 MeV to 5 GeV and carried a single experiment: a spark-chamber telescope, developed by six European institutes in collaboration. The experiment became operational on 1975 August 17 and was switched off on 1982 April 25 when on-board resources were exhausted. During this period, 65 observations, typically of a month's duration, were performed. The satellite was spin-stabilized with the telescope axis along the spin axis. Circular sky regions of about 40 degrees in diameter were covered in each observation. The majority of the pointings were distributed along the galactic equator, 15 observations were devoted to regions at high (>20 degree) galactic latitudes. Several regions of specific interest were observed repeatedly.

The highly eccentric polar orbit of COS-B, with an apogee around 90,000 km (chosen to maximize useful observation time while allowing real-time data transmission) exposed the experiment to the solar modulated interplanetary cosmic-ray flux. The unexpectedly long operational life of the experiment, specifically of the spark-chamber, was accompanied by a long-term degradation and by short-term disturbances of its performances and, consequently, of the experimental sensitivity. The variation and sensitivity of the instrumental background were investigated thoroughly and integrated into the database. The possible impact of their statistical and systematic uncertainties must be considered in any type of analysis.

The Data

The format of the COS-B data, like the previously discussed SAS-2 data, is based on a ROSAT Events FITS file. The COS-B FITS files are based on the final COS-B database (see H.A. Meyer-Hasselwander et al., 1986 for a detailed description). The original database contained three files: the observation file, the exposure file, and the event-list file, with indices in one file pointing to records in another file. The design of this database is similar to the extensions in a ROSAT Events FITS file, making the creation of the COS-B FITS files easier than those of SAS-2. The completed database contains 65 files corresponding to the 65 observation periods containing all of the exposure and event information for that period.

During the processing of the data, we decided to scan each of the three original files for data integrity. For the event-list file, we checked the following parameters for their proper values (as described in the "Explanatory Supplement to the COS-B Final Database" by H.A. Meyer-Hasselwander): gamma class, edit class, photon energy, right ascension, and declination. We found that some events have a `photon energy' of 0 MeV, a `declination' of +95 degrees and an `edit class' larger than three. These 750 events were obviously not meaningful and were deleted from the database. Additionally, the `gamma class' was found to have more than the three cases (2, 22, and 3) as specified in the "Explanatory Supplement." Since the energy and coordinate of the event was acceptable, these events were not deleted from the database. Details of other minor discrepancies can be found in the "HEASARC COS-B Database Document", which is currently in preparation.

4. Using the Data

The main reason for restoring the SAS-2 and COS-B datasets is to make them available in a standard FITS format that is useful to the High Energy Astrophysics community. The rationalized FITS format developed by the HEASARC and the ROSAT Guest Observer Facility was adopted as the standard. The ROSAT and ASCA missions are using this format as their standard and EGRET data also will be translated into this format, as will most future HEA photon event data. The availability of SAS-2, COS-B, and EGRET data in this format allows the data to be imported easily into the IRAF/PROS software package and XANADU software package, including XSELECT, a set of tasks for analyzing HEA data in FITS format.

In addition a program called FADMAP, which is based on a program of the same name for analyzing COS-B data, will be made available soon. FADMAP produces three maps or images: a background counts map, a source counts map, and an exposure map. A fourth map, a flux map, is produced from these three maps. By using FTOOLS tasks, it will be possible to combine data from SAS-2, COS-B, and EGRET into a FITS file that then can be read into FADMAP or any of the other previously mentioned data analysis software.

Anonymous ftp

The data are currently available via HEASARC's anonymous ftp account and are located in the public directories, cosb/ and sas2/, on the host, ( Each directory contains several subdirectories, data/, doc/, and calib_data/, along with .message and README files describing the contents and status of the directory. The calib_data/ subdirectory also has several subdirectories including a doc/ subdirectory that currently contains LaTeX and PostScript files of the SAS-2 and COS-B Calibration Guide.

Access to these data is possible via anonymous ftp or Gopher. For more information about using ftp or Gopher to access the HEASARC anonymous ftp account, see the article by Drake et al. in this issue of Legacy.

5. Future Directions

Possible Revisions

The SAS-2 and COS-B FITS files may undergo some revision in the future to make the data more accessible to the high-energy astrophysics community, though we expect these changes to be minor. Aspects that may affect these changes are:

1) the compatibility of the two datasets with each other. Because SAS-2 and COS-B were different instruments from different experimental groups, the event data and housekeeping data were likewise different. Some incompatibility still exists between the two datasets mostly due to differences of the housekeeping data. Such changes may be resolved in the future allowing data analysis software to access the data in a more consistent manner.

2) the compatibility of the two datasets with the Compton Gamma-Ray Observatory Science Support Center's data analysis software. Some of the data analysis software is designed to analyze EGRET data. Changes to the SAS-2 and COS-B FITS files also would allow this data to be analyzed using the same data analysis software.

In a future issue of Legacy, we plan to describe how to use the tasks available in the XANADU software package, including FADMAP, to analyze the COS-B and SAS-2 data.


"The HEASARC's Newly Consolidated Anonymous FTP Account", 1993, S. Drake & B. O'Neel, Legacy, 3, 53.

"Gamma Radiation from Celestial Objects", 1967, G.G. Fazio, , Ann. Rev. Astr. Ap., 5, 481.

"High-Energy Gamma-Ray Results from the Second Small Astronomy Satellite", 1975, C.E. Fichtel, R.C. Hartman, D.A. Kniffen, D.J. Thompson, G.F. Bignami, H. Ogelman, M.E. Ozel, & T. Tumer, Ap. J., 198, 163.

"First Energetic Gamma-Ray Experiment Telescope Catalog",1993, C.E. Fichtel et al. , Ap. J. Supp., submitted.

"Explanatory Supplement to the COS-B Final Database", 1986, H.A. MeyerHasselwander et al., in Proc. Cosmic Ray Conf., La Jolla, ESA Publication.

Next Proceed to the next article Previous Return to the previous article

Contents Select another article

HEASARC Home | Observatories | Archive | Calibration | Software | Tools | Students/Teachers/Public

Last modified: Monday, 19-Jun-2006 11:40:52 EDT