[Date Prev][Date Next][Thread Prev][Thread Next][Search] [Main Index] [Thread Index] [HEASARC Archives]

FTOOLS v3.6 released


We are pleased to announce the release of FTOOLS v3.6.  In this
message, we summarize:

	1. How to obtain FTOOLS v3.6
	2. Impact on ASCA users
		2a. Differences between v3.5.2 and v3.6
		2b. Differences between v3.5 and v3.6
	3. Known bugs/deficiencies of FTOOLS v3.6

1. How to obtain FTOOLS v3.6
   FTOOLS v3.6 has been released, and can be obtained from the
FTOOLS web-page at:


Notices of the FTOOLS Releases, online HTML versions of FTOOLS
documentation (like this document), the official FAQ, and other
goodies are available on the World Wide Web at:


This is where the most up to date FTOOLS information is available,
including the continually updated FAQ. Users should always check
here first for answers to questions.

Any problems installing FTOOLS should be reported to:


Any problems using FTOOLS for ASCA data analysis should be reported
to the ftoolsheelp address above or to:


If the problem relates to the scientific contents of the FTOOLS,
the latter address is more appropriate.

2. Impact on ASCA users

There are relatively few changes in the contents of ASCA FTOOLS
since ASCA patch 3.5.2; changes since the original v3.5 release
are quite substantial.  These are listed separately below.

                2a. New since v3.5.2

* We now support stand-alone FTOOLS under Linux, and IRAF version
  of FTOOLS under Solaris.

* We have dramatically improved our testing of FTOOLS on different
  architectures; previous releases were extensively tested only under
  Alpha/OSF and Sun/SunOS.  With increased testing under Sun/Solaris,
  SGI and other types of machines, we have flushed out many architecture
  or compiler dependent bugs (e.g., non-standard extensions supported on
  most, but not all, machines).  Consequently, we believe this version of
  FTOOLS are far more robust, particularly on Solaris machines.

Changes to individual software:

* We have been notified of an error in the treatment of area discrimination
  for SIS Fast mode data ('select fast inside' 'select fast outside' commands).
  We have fixed this problem in xselect.

* We have fixed a bug in extractor which resulted in incorrect wmaps when
  the extraction  region was specified in sky coordinates.  However, see
  for the limitations of the current algorithm.

* Sisrmg has been converted to use native C version of fitsio.

* Ghkcurve has been updated to treat combined medium and high bit rate
  data correctly.

                2b. Changes since original v3.5 release

ascaarf    - uses new XRT calibration files (xrt_ea_v2_0.fits,
             xrt_psf_v2_0.fits) and new "filter" modification
             to the effective area curve that provides power-law
             fit to GIS Crab spectra consistent with Toor & Seward
             (1979; AJ 79, 995). This filter can be switched off using
             the 'filter=no' option. All calibration files can be read
             from either the CALDB or the refdata area. The default is
             the AUTO setting which first checks for CALDB files and if
             they are not found then looks in the refdata area. If CALDB
             is given as the value of the input RMF parameter then ascaarf
             checks the CALDB for an appropriate RMF to use to set the
             energies on which the ARF is calculated.

             Note that the "filter" modification is applied to both GIS
             and SIS data however it has only been calibrated against the
             GIS Crab data.

ascaeffmap - uses new XRT calibration files and AUTO argument for all
             calibration files.

ascalin - now can redetermine GIS/SIS PI values using manually
          input new gain and/or offset numbers.  To use this option,
          use the 'gain fit' command in XSPEC and use the resultant
	  gain and offset numbers on the command line (gainnorm=0.98,
          gainoff=0.03 etc.)  Note, however, the parameter file will
          remember the values supplied; it is recommended to reset these
          to 1.0 and 0.0 respectively, after use.

          Please note that this update is to enable manual adjustment of
          the gain, specifically to allow for the GIS gain mis-correction
          announced recently (see GIS team memo Apr 12, 1996). The update 
          corresponding to the July 10th announcement by the GIS team
          will be available in a future release.

fastwmap - bug fix of the FTOOLS v3.5 release version (this fix was made
           available shortly after the v3.5 release)

ldeadtime - carry out deadtime correction for GIS light curves

sisrmg - several user interface bugs have been fixed.  It should now
         default to grade 0234 for BRIGHT/BRIGHT2 mode data, unless
         the 'grades' hidden parameter is used to override the default.
         For fast mode data, it is recommended that you specify the grade
	 explicitly (normally grades=0) on the command line; without this
	 line, sisrmg will still crash on Alpha/OSF machines for fast mode
         data (this version does run on Alpha/OSF machines, provided that
         you supply the grades=0 option on the command line).  However,
         this is believed to be the last (minor) remaining user-interface bug.

3. Known bugs and defficiencies of FTOOLS v3.6

* The current scheme used for spectral extractions in sky coordinates have
  limitations.  See
  for details.

* When running on PI spectral files, sisrmg does not treat the event
  threshold/lower level discrimination correctly (PHA vs PI inconsistency).
  This may result in slight inaccuracy for the lowest channels.

* The formula to correct for the GIS-3 gain problem (apparently caused by
  a change in the gain map, as detailed in the 1996 July 10th announcement
  by the GIS team) is not yet available in ascalin.

* The current version of SISPI does not record the nominal gain value.
  For the usual case of 3.65 eV per channel, the FITS header will have:

GAIN_NOM=           0.00365000 / Nominal gain value (ev/channel) for PI column

* The standard region selection (to exclude the calibration source) for
  GIS 64x64 spatial resolution mode fails.  However, a point source near
  the center can be extracted without problems.