This directory contains a dummy HEXTE PHA file for XSPEC simulations of 
HEXTE spectra, accompanied by response and background files. 

NOTE:  These files apply for Cycle 3 and later (including current Cycle)

Specifically, you'll find:

dummy_hexte.pha   - dummy PHA file for setting up the response, background
and bad PHA channels in XSPEC.

hexte_cl_nra3.bak - HEXTE cluster background spectrum (average of clusters A
and B)

hexte_cl_nra3.rsp  - HEXTE cluster response matrix (average of clusters A
and B, excluding the detector PWB2 which no longer produces spectral
information)

Instructions for using the files in XSPEC follow these important notes:

NOTES
=====
1. The HEXTE dead-time fraction of 40% has already been accounted for in the
background and response files; users simulating spectra should concern
themselves only with the total exposure time to be proposed (and the
fraction thereof spent on source and background, see below).

2. The background spectrum is a cluster average produced by Dan MacDonald at
UCSD by accumulating background data for ~200ks live-time. For simulation
purposes, statistical errors have been removed from the PHA file, but
systematic error estimates at the expected level of 8% of the cosmic x-ray
background have been included. However, the XSPEC "fakeit" command ignores
systematic errors.

EXAMPLE OF AN XSPEC SIMULATION FOR A HEXTE OBSERVATION
======================================================

1. Type "data dummy_hexte.pha" to read in the dummy PHA file. This will also
cause the correct response (hexte_cl_nra3.rsp) and background file
(hexte_cl_nra3.bak) to be read in automatically, as well as the bad PHA
channel ranges.

2. Define an XSPEC model with the correct normalization for your source
of interest.

3. Decide on the total proposed exposure time. This will be the same
number that you specify for the PCA. However, the HEXTE background and
response files are for 1 "average" HEXTE cluster. Since the total number
of counts received is proportional to (# clusters * time) for each
position, use of these 1-cluster files for simulating HEXTE datasets is
different for two cases; consider for example a HEXTE observation of
100ks... 

3A: two-cluster 16s source/background beamswitching. In this case, one
HEXTE cluster is always observing on-source while one is observing
background simultaneously (even though the individual clusters swap roles
every 16s). So for the on-source PHA file simulation, enter 100ks (=1
cluster * 100ks) For the background simulated PHA file, one should account
for the 4s per 16s background dwell which is lost due to cluster motion.
Therefore, simulate a background PHA file of exposure 75ks(= 1 cluster (*
100ks * (16-12)/16 to account for cluster motion). For beamswitch dwell
cycles longer than 16s, eg. 32, 64 or 128s, subsitute the dwell cycle
duration for "16" in this formula. 

3B: two-cluster STARE observation. In this case, both clusters are
on-source for the entire 100ks, which is equivalent to a single cluster
observation of 200ks. So use the files to simulate a 200ks exposure (=2
clusters * 100 ks). Or perform 2 simulations of 100ks

Other combinations are possible, eg. one cluster beamswitching while
another stares at either source or background. In this case, use the
appropriate cluster exposure (number of clusters * time) for
the source and background spectra respectively. 

4. Type "fakeit" to initiate the simulation. Hit return to enter the
defaults for the first two questions, and then a name for the on-source
PHA file. Finally, enter an exposure time according cluster's rocking
pattern: beamswitching (case 3A above), or "staring" (case 3B) eg. for a
100 ks observation with source/background beamswitching:

XSPEC> fakeit
       mkn841_1e5.fak
       1.0E5       ! beamswitching case; use full exposure

For the two-cluster staring case (3B above), enter 2.0E5 for the exposure
and you are done with the simulation stage - skip to note 6 below.

4. For source/background beamswitching observations, the next step is to
create a simulated background spectrum with the appropriate statistics. Set
the model normalization(s) to zero, and type "fakeit" again. Personally, I
like to include the exposure times in the background file-names so that I
can keep track of and re-use them in simulations for other sources using the
same exposure times, eg. for the 100ks beamswitching observation above

XSPEC> fakeit
       back_1e5.fak
       7.5E4    

Note that the off-source exposure is 1.0E5 * (16-2*2) / 16 s = 7.5E4s due
to the 2s lost each time the cluster moves on- or off-source. For slower
beamswitching, subsititute the dwell time (32, 64, or 128) into this
formula. 

5. Now read in the simulated source and background spectra - note that it is
essential to read in the SIMULATED background spectrum, to override the
default, noise-free hexte_cl_nra3.bak; otherwise your statistics will be too
optimistic.

XSPEC> data mkn841_1e5.fak
XSPEC> back mkn841_1e5.bak       ! VERY IMPORTANT TO DO THIS

(Users familiar with the Ftools GRPPHA may change the BACKFILE keyword
the on-source file to read in the simulated background automatically).

XSPEC> show files

This will confirm which source and background files are currently being
used by XSPEC.

6. Ignore the predefined HEXTE bad PHA channels (currently channels 0-12).

XSPEC> ignore bad

If you choose to raise the HEXTE's Lower Energy Bound above its default
value of PHA channel 12 (eg. to reduce a bright source's count rate to
stay within telemetry limits), you should also ignore those channels up to
the Lower Energy Bound, since they will contain no data in practice. Note
that the PHA channel number = photon energy in keV apprximately, so
for a 30 keV lower bound you would type "ignore 1-30". 

7.  You're now ready to analyze your simulated HEXTE spectrum! 

Philip Blanco, CASS/UCSD.
Please send questions and comments to xtehelp@athena.gsfc.nasa.gov