2.1 AE Version 2011-03-15

• Small changes were made to accomodate CIAO 4.3.

• Due to a conflict between IDL version 8 and the historical AstroLib, if you are using IDL version 8 you need to update your AstroLib to the Sept. 2010 or later version (http://idlastro.gsfc.nasa.gov/ftp/), and update your copy of our TARA package (if you use that) to the 2011mar16 version (http://www.astro.psu.edu/xray/docs/TARA/code/tara2011mar16.tar.gz).

• Any time you need to run a tool in the files acis_extract_tools.pro or ae_recipe.pro you should now load all the AE code with the IDL command
  .run ae
  rather than with any of these now obsolete methods:
  .run acis_extract_tools
  .run ae_recipe
  .run aet
  .run aer

• A discussion of ObsIDs that include both FI and BI detectors has been added to §5.5.

• The process of generating FITS and L^ATEX tables that collate AE results has been improved. The COLLATE stage of AE still produces a FITS table with vector columns, in order to record source properties over many energy bands. We have introduced the new tool ae_flatten_collation that transforms the AE collation into a more convenient and scientifically useful FITS table.
  • Results are retained for only the three energy bands we typically care about. Those quantities are stored in independent columns, eliminating all vector columns.

  • The unfriendly column names used by AE are replaced with more meaningful names.

  • The CIAO tool aprates is used to estimate confidence intervals on NET_CNTS, replacing the less accurate AE photometry confidence intervals.

  • The observed energy flux is very simply estimated by multiplying AE's photon flux by the median observed event energy.

  • The confidence intervals on spectral model parameters produced by XSPEC are formatted into L^ATEX strings, ready for use in a L^ATEX table of fitting results (either the ones we provide with AE, or custom tables produced by the observer).

  • XSPEC confidence intervals on model normalizations are converted to emission measures and formatted into L^ATEX strings. XSPEC fluxes are converted to luminosities.

• Our own recipe for using AE (§7.1) has been revised.

• The XSPEC fitting scripts supplied with AE have been revised.
  • XSPEC version 12.6 or higher is now required.

  • The scripts have always saved information about the fitted model using FITS keywords. The 8-character limit imposed on FITS keywords forces us to use some very unreadable keyword names. Unfortunately, we historically have not chosen these names very carefully. Thus, over time, the names for some keywords have evolved.

    For example, keywords holding XSPEC fluxes have to encode the energy range over which the flux was computed. We originally used a naming convention like F0P5_8 to mean flux in the 0.5 to 8 keV range; now we use the letter ``H'' as an abbreviation for 0.5 and we omit the underscore, resulting in the replacement name FH8.

    For another example, we originally represented parameters from the first component of thermal models with keywords that omitted the number ``1'', e.g. ``NH'', ``KT'', ``NORM''; now we include the ``1'' in the names.

  • The design of the scripts has been modified to reduce the labor required to implement a new spectral model. These new scripts have NOT yet been well tested.

• The tool ae_radial_profile (§7.20.5) has been simplified.

• Units are more consistently recorded for source properties and shown on plots. The format follows the standard defined by the document ``Specification of Physical Units within OGIP FITS files''.

• The execution speed of image reconstructions (CHECK_POSITIONS stage) has been improved by reducing neighborhood sizes and cropping PSF images.

• The AstroLib routines used for image reconstruction have evolved over time. With some versions of the AstroLib, the source position estimates AE derives via image reconstruction may have systematic errors. In this and subsequent versions of AE, the CHECK_POSITIONS stage begins by reconstructing a fake data image with a fake PSF, and then verifying that the astrometry of the result is correct. This defensive step should detect situations where AE and the AstroLib are making different assumptions regarding the astrometry of PSF images.

• An AE user (G. Sivakoff) has generously contributed IDL code that will query the CXC's archive to determine the aimpoint of an ObsID; see §7.6.