A GUIDE TO THE INTERACTIVE ANALYSIS SYSTEM
1. General
In the period 01/12 to 31/12, several groups of visitors to the
Observatory made use of the interactive analysis system and their experience,
together with that of the Observatory Team, has led to a few general guidelines
for the organization and proper exploitation of the system.
Users are reminded that the IA system was implemented primarily as
a facility
whereby the community could analyze EXOSAT data in more detail at the
Observatory. There is no question of the Observatory Team undertaking a
systematic interactive analysis (in the absence of the user) of data
as a service
- this is the function of the automatic analysis. It is recommended
that users of
the system should as far as possible be familiar with general X-ray
data analysis
techniques, since the support provided by the Observatory Team is geared in
general towards a demonstration of the facilities rather than a
'supervision' of
the entire analysis. First time users are expected to require a few hours full
time advice from the Duty Scientists followed by consultancy support
on specific
details.
Real time interactive analysis of LE data will not
be supported because of the excessive time required to linearise the image.
During the system trial period, which will terminate on 15.03.85,
no real time
booking will be permitted, i.e. P.I.'s attending their current observation must
request IA system time in advance in the normal manner (EXPRESS no. 7, p. 43).
Only one 'observer' will be allowed access to the system at any one time a
although one observer means in practice one or two 'bodies'. From 15.03.85
to 31.03.85 the Observatory Team require 100% access to the computer system for
configuration, coordination of final updates and testing and documentation such
that a full service will be provided from 01.04.85.
Throughout the
night period (23.00-08.00) no support of the IA system is in principle
available,
although the Observatory Controller will be able to offer some advice,
consistent
with the real time DCR operations having priority at all times.
Note
that the software comprising the IA system is not an Observatory product, but
rather a means of production and as such is not available for distribution
outside the Observatory.
2. The ME Interactive Analysis System
A description is given of the facilities that are
supported, or will shortly be supported by the ME
Interactive Analysis System. The two basic end
products of the ME interactive analysis are spectral
and rates files. These are used by a series of
analysis programs, common to all the instruments, for
spectral fitting, period searching, plotting etc.
Data from the following OBC modes can be analysed
with the system: HER2, HER3, HER4, HER5, HER6, HTR3,
PULS, PLS2, DIR2. Note that for the pulsar programs
only spectral analysis can be carried out at present
and for DIR2 only a very simple analysis is possible.
a) Spectra
The data used to create a spectral file can be
selected in the following ways:
1 - Argon or Xenon data
2 - Particular Detectors
3 - Time, phase or count level windows
Background subtraction can be carried out
using:
1 - Slew data
2 - Array swap data, corrected for offset
3 - Standard CCF spectra
4 - Leicester University background correlations
The detector gains are stored on the spectral file
at the time of creation so that the effects of
detector gain drifts and pre-amplifier gain setting
changes are 'hidden' from the user. In addition,
deadtime correction factors can be measured or
calculated and applied to the spectra. Once a spectral
file has been produced it can be plotted or used as an
input to the spectral fitting program.
b) Light Curves
Light curve or 'rates' files can be produced with
the following options:
1 - Argon or Xenon
2 - Integration interval
3 - Energy channels (or bins)
4 - Method of background subtraction (as for spectra)
In addition, the difference in intrinsic background
counting rates can be corrected for using either array
swap or slew data or standard background files. The
file can be deadtime corrected and for some OBC modes
the deadtime correction factors can be a function of
time. The times stored in a rates file can be
corrected to heliocentric times; note that Madrid is
assumed to be at the centre of the world and the Sun
at the solar system barycentre.
The rates files can be used as inputs to the
following programs:
1 - Fourier Analysis
2 - Folding Analysis
3 - 'Publication Quality' Plotting
4 - Interactive Timing analysis facility
c) Other Facilities
Other important facilities supported by the ME
Interactive Analysis include:
1) Directory listings - Listings of times, OBC
modes pointings of all
the observations on a FOT
etc.
2) Observation Information - More detailed listing for a selected
observation
3) HK Analysis - Plots and listings of
any spacecraft or
experiment HK parameters
4) Macros - Commands can be executed
from a file. Parameters can
be set at execution time
allowing the use of
standard macros
5) Help files - On-line documentation of each command
d) Future Improvements
The ME Interactive Analysis was designed to be used
by a small number of scientists each of whom would be
very familiar with all aspects of the operation of -
the ME, the OBC, the HP computers and the OP analysis
environment. As many of the astronomers coming to use the interactive
analysis will not be familiar with these, much work is
currently underway to improve the help files and level
of documentation. In addition, analysis of high time
resolution data is very time consuming mainly because
of the low CPU power of the HP computers. A link to
the ESOC Fujitsu main-frame computer has been
established and it is expected that many of the CPU
intensive processes will be transferred from the HP's
to the Fujitsu
2. The LE Interactive Analysis System
The LE Interactive Analysis System consists of more
than 10 programs to enable users to perform all the
basic operations needed for the detailed analysis of
CMA data.
The following is a list of the main options offered
by the system:
Accumulate
images
Deblur
images
Accumulate sum signal histograms
Calculate efficiency correction (see p. 27 for
further details)
Accumulate rate files*
Rotate images
Overlay images
Automatic source detection
Source centroid
Signal to noise ratio optimisation
Source intensity estimation and
significance of detection
Hot spot removal
Gaussian smoothing
Image contouring
Pixel <--> celestial coordinates conversion
Count rate <-->flux conversion
* Accurate timing analysis of CMA data can be
performed using LE rate files as inputs to the ME or
GS timing programs or to the timing interactive system
that is currently in preparation.
Several interactive programs suitable for the
control of colour display devices are also available.
Documentation describing the LE interactive system
is currently being written and will be available at
the Observatory in the near future. A two-level help
file and some documentation macros are already
available on the HP4 computer system at ESOC.
4. The GSPC Interactive Analysis System
Facilities available:
- - Accumulate spectra for HEBL2 data
HEBL4
DIRECT (old)
- - Accumulate rate buffers for HEBL2, HEBL4, HK
data.
- - Background subtract spectral buffers and
determination of detector gain.
5. General Facilities
- Spectral Fitting - One or two detectors up to
a maximum of 256 channels. Detectors that are
currently supported are GSPC, ME and all CMA
filters (the latter are counted as one
detector).
- FFT - a 4096 point fast Fourier transform can
be applied to the rates buffers generated by
the various IA programs. For data sets longer
than 4096 points the FFT's are summed
together.
- Folding - An unlimited amount of data can be
folded over 75 trial periods and a plot of
as a function of period obtained together with
a plot of the light curve of the period with
the best .
- Light curves - a general purpose folding (at a
single period) and light curve program will
data for three detectors at once.
- Timing analysis interactive system (currently
in preparation and testing) common to all
instruments with the following objectives:
- - 'chain' the data (bad data, fill in gaps
with simulated data etc. and eliminate long term
trends).
- - test the hypothesis of constancy by various statistical tests
(, % variability) and determination of the source statistical moments.
- - search for flares/bursts (down to msec).
- - search for periodicities and determination of folded light curves.
- - phase fitting for pulsating sources.
- - characterise aperiodic variability via auto-correlation function
techniques.
- - investigate amount of variability as a function of energy and
spectral changes.
- - correlate data from different energy bands or
experiments via cross correlation function
techniques.
D. ANDREWS
P. GIOMMI.
A.N. PARMAR
L. STELLA
N.E. WHITE
[EXOSAT Home]
[About EXOSAT]
[Archive]
[Software]
[Gallery]
[Publications]