|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
BEFS Time Tagged DataThough the BEFS standard data-extraction routine produces a single spectral file (SPD1) for each exposure, there is a mechanism for producing a time-tagged spectral-photon list for any spectrum. The program extract_tt, (available in both c and idl with other routines from archive.stsci.edu/pub/orfeus/befs/software/), reads a photon-list file (LST1), extracts those photons falling within a user-defined region of the detector, and writes detector X and T for each photon to a binary table extension in a FITS-format output file (TTD1). To run the c version of extract_tt (information on the IDL version can be found in the program header), you must first compile and install the rdpar and cfitsio libraries, which live in the directory SOFTWARE/lib_src. The source code for extract_tt is in SOFTWARE/src. The program expects a single command-line argument: the name of a parameter file containing all of the other input parameters. The complete set of files used on the ORFEUS-I data set live in SOFTWARE/parfiles. For example, to extract a time-tagged FUV spectrum of HD 88115, use the parameter file hd88115.pfd1. Make sure that it points to the appropriate calibration files, which live in SOFTWARE/calfiles. In fact, the only calibration file actually used by extract_tt is the list of Y values that define the center of the extracted spectral window, called "d1 y centers" in hd88115.pfd1. The output of this program is a FITS file (e.g., BEFS1058.TTD1) with a single binary table extension (EXTNAME = EVENTS) containing (X, T) for each photon in the spectral window. The data are neither calibrated nor background subtracted, but do include only those periods deemed "good times" during our preliminary analysis. One or more short (1-second) data drop-outs are possible. The files can be read into IDL using the FITS binary table i/o routines of the IDL Astronomy User's Library, available at http://idlastro.gsfc.nasa.gov/homepage.html. For example, to plot the total count rate as a function of time, enter fxbopen, unit, 'BEFS1058.SPD1', 'EVENTS', header fxbread, unit, time, 'TIME' fxbclose, unit plot, histogram(time) Because the total count rate may be dominated by airglow, it is probably more profitable to consider only subsets of the data, plotting spectra for a series of 100-s integrations, or plotting the count rate in a restricted spectral region. You can calibrate your mini-spectra using information from the corresponding .SPD1 file, but there is currently no easy way to perform a background subtraction of your time-resolved spectra. NOTE: Events with values of X between 0 and 1023 do not represent real photons, but are stim pin (stimulation pin) events used to determine the dead-time correction (DTC). Once you have binned your data, compare the stim-pin rate in each time segment with the mean value for the observation. To properly flux-calibrate your rebinned spectra, you must re-scale the DTC for each subset of the data (the DTC goes up when the stim pin rate goes down). This effect should be very small.
This information was all provided by Dr. Van Dixon of the
Space Sciences Laboratory, The University of California, Berkeley.
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|