Next: Data Format: Type I Up: THE OGIP STANDARD PHA Previous: THE OGIP STANDARD PHA
The header for the PHA file Data Extension includes all the information required by XSPEC (and by extension any other general spectral analysis package) to describe the stored PHA data. This consists of the following (mandatory) keywords/values:
The POISSERR keyword is not required if a STAT_ERR column is present however it is recommended in this case for clarity. If STAT_ERR is to be used for the errors then POISSERR is set to false.
The CHANTYPE keyword specifies whether the channels referred to in the file are as assigned by the detector electronics at the time of data collection (in which case CHANTYPE = PHA), or whether any corrections have been applied. An example of the latter case is when the data has been re-mapped onto a standard "pulse-Invariant" (PI) channel grid, in which case CHANTYPE = PI (see also CAL/GEN/92-002, George et al. 1992, Section 7).
The following keywords are mandatory if the channel numbering-scheme for an instrument does NOT start at unity:
The following optional keywords supply further information:
The XFLTXXXX keyword is used by XSPEC to aide joint spectral/temporal or spectral/spatial fitting when the input consists of several individual PHA files. The keyword will contain the values of the appropriate time (or phase etc. ), or spatial parameters describing the PHA dataset. Thus, in the case of joint spectral/temporal analysis for example, the XFLT0001 keyword will consist of a string containing the start & stop time of the PHA dataset. These times will be in the units required by the spectral/temporal model to be used within XSPEC (and thus could be in units of orbital phase of a binary system). It is the responsibility of the User to consult the relevant GTI and/or Detector extensions (below), and place the required information into the relevant XFLTXXXX keyword. Further examples of such complex spectral analysis tasks will be given elsewhere.
A number of optional keywords are provided to supply the user with information on the precise start & stop times of the integration of the PHA dataset. These will be listed elsewhere and adhere to the standard OGIP method for defining times (see Angelini et al. 1992). Such keywords are necessary since times calculated using the standard (DATE-OBS + TIME-OBS) & (DATE-END + TIME-END) keywords are not guaranteed exact (since they involve a conversion from spacecraft clock to UT etc. ), and in any case do not allow sub-second temporal resolution. Times specified by these keywords are in the original temporal units supplied by the spacecraft/pre-processing system. Thus, in the case of (say) sub-second, phase-resolved spectroscopy users are able to cross-check that their selected PHA dataset does indeed correspond to that required from the ephemeris, taking into account any glitches in the spacecraft clock, leap seconds, baryocentric corrections etc. It is strongly recommended that the information supplied by these keywords is included in all PHA datasets, even if not strictly necessary from the scientific viewpoint. Of course, the exception to this recommendation is when the times have no meaning -- e.g. when a phase-folded or source-intensity-selected PHA dataset has been constructed.
For type II data extensions with multiple spectra some of these keywords may have different values for each spectrum. In this case the values will be stored in a column instead of a keyword. The standard keywords which may be turned into columns are EXPOSURE, BACKFILE, CORRFILE, CORRSCAL, RESPFILE, ANCRFILE, XFLTXXXX, DATE-OBS, TIME-OBS, DATE-END, TIME-END.
Next: Data Format: Type I Up: THE OGIP STANDARD PHA Previous: THE OGIP STANDARD PHA Keith Arnaud 2009-05-06