CalDB PUBLIC Release Notes
SDP Installation Date: (UTC) 2008-09-10T15:30:00
Public Release Date: 08
SDP + PUBLIC
CalDB 3.5.0 includes two ACIS upgrades:
as well as a new HRMA Encircled Counts Fraction file (for use in
determining PSF sizes for source detections in L3,) which is assigned
the codename "REEF" for "Radial Encircled Energy Fraction":
- ACIS -120C T_GAIN Epoch 34
- ACIS -110C OSIP table mods (BI chips only).
"How CalDB 3.5.0 Affects Your Analysis"
II. Summary of Changes
A. ACIS -120C T_GAIN Epoch 34
Epoch 34 covers the 01 May through 31 July 2008 period. The T_GAIN
interpolation files effective for 2008-02-01 include both Epochs 33 and
34. The new files effective on 2008-05-01 are non-interpolating or
"flat" files including only Epoch 34. The four datasets follow the
current convention for T_GAIN
upgrades, in that the first and third files include data for S1 and S3
(the BI chips) that are NOT CTI corrected. The second and fourth files
have CTI-corrected BI chip data. For the FI data, the CTI correction is
always applied, as no
calibration for T_GAIN exists for the FI non-CTI case.
SDP/CIAO ACIS L1 processing tool "acis_process_events"
"Reprocessing Data to Create a New Level=2 Event File"
Why Topic: "ACIS Time-dependent Gain Correction (T_GAIN)"
B. ACIS -110C OSIP version N0007
With the addition of the ACIS -110C Epoch 0 T_GAIN file (applicable to
the BI chips only) and the corresponding addition of ACIS -110C
DET_GAIN version N0006 with CalDB 3.4.4 (March 26, 2008), there is an
incumbent upgrade in the order-sorting (OSIP) tables for the
BI chips during the same epoch. We have combined the unaltered OSIP
tables for the FI chips from version N0006 with those of the BI chips
in the file acisD2000-01-29osip_ctiN0005.fits, which were derived from
the BI chip responses in acisD2000-01-29p2_respN0005.fits, and the
corresponding BI chip detector gains above.
The new ACIS -110C OSIP file is applicable only to data taken
within the first year of Chandra normal operations and only when the
ACIS focal plane temperature (FP_TEMP in the data file headers) is set
to -110C, or 163.2K +/- 5K. Only the ACIS-S1 and ACIS-S3 OSIP data are
being upgraded with this release. The motivation for this upgrade is to
allow a more rigorous, predictable (rather than ad hoc) setting of the
ACIS-S1 order-sorting acceptance region than has existed since CalDB
2.26 (Feb. 2, 2004). Such an upgrade allows a more quantitative and
trustworthy estimate of the FRACRESP values in the OSIP table for
ACIS-S1, which translates to more accurate PHAFRAC values in resulting
grating ARFs derived therefrom.
Practically speaking, the changes, even to the ACIS-S1 MEG
order-sorting profile, are visible, but not very statistically
significant, compared with the previous (N0006) OSIP version for -110C
operations. However, as most the -110C data have not been reprocessed
since 2001, the user is well-advised to reprocess any ACIS data from
the period in question, so that the much-improved ACIS GAIN data are
applied particularly to ACIS-S1. Furthermore, an extension of the Repro
3 archival reprocessing effort will be completed during the coming
months, and the new GAIN, T_GAIN, and OSIP will be applied to all -110C
observations from 1999-09-16 through 2000-01-28. Hence, the new OSIP
will become the standard for this and any future processing of these
SDP/CIAO Grating L1.5 processing tool tg_resolve_events
"HETG/ACIS-S Grating Spectra: Generate a New
Level 1.5 Event File"
"LETG/ACIS-S Grating Spectra: Generate a New Level 1.5 Event File"
"HETG ACIS-S Grating ARFs: Run fullgarf"
"LETG ACIS-S Grating ARFs: Run fullgarf"
C. HRMA REEF version N0001 (Encircled Counts
The Encircled Energy Fraction (ECF) file is used in two separate tasks
in L3 pipeline processing:
- To filter out spurious sources when combining multiple
sourcelists from a single observation derived from different wavdetect
that use data from different energy bands and blocking factors.
For each source in each sourcelist, a psf size is estimated for that
source location and for a representative energy in the energy band,
using the ECF file. If the source size as reported from wavdetect for
that energy band and blocking factor is inconsistent with the psf
size, the source data for that source is excluded from the final
- To estimate source and background aperture sizes at
arbitrary locations in the field-of-view for each energy band. The ECF
file provides estimates of circular source and annular background
apertures that are approximately equal in area to the actual apertures
used in the aperture photometry tasks. These apertures
are used to estimate limiting sensitivity (minimum flux a source
must have to be included in the Chandra Source Catalog) for
arbitrary source position.
NOTICE TO USERS: The new REEF file is only accurate for a specific
HRC-I configuration and SIM_Z setting, and is never to be used as a
source for PSF size data for research or publications. It is only being
used to set the source selection sensitivity for L3 source cataloging.
PSF size values thus derived are not considered accurate or publishable
for scientific purposes. The file is NOT being released for use in CIAO
ACIS and HRC Level 3 pipeline tool mkpsfsize, and the Level 3
S-Lang PSF size module
To be determined with the CIAO 4.1 release.
III. Technical Details
A. ACIS Epoch 34 T_GAIN files
The new -120C T_GAIN data were derived from ACIS External
Source (ECS) data taken during radiation zone passes during May - July
2008. Here are the plotted corrections, showing the changes from
the previous files and the new ones, for the aimpoint of ACIS-I3
(Figure 1). At 5 keV, the new correction is about 5-10 eV and at 10keV
the correction is 9-18 eV (0.09-0.18%), depending on the interpolation
Figure 1: A comparison of old and new
corrections for the CTI-corrected FI chip I3 at the aimpoint. There is
not much of change with the new files.
Likewise, below is the comparison between old and new T_GAIN
for the aimpoint of ACIS-S3, with the BI chip CTI correctors
Figure 2: A comparison of old and new
corrections for the CTI-corrected BI chip ACIS-S3 at the
aimpoint. The changes here are minor as well, though a bit larger than
the other two illustrated cases.
Now consider the non-CTI corrected case for ACIS-S3, applied to
for which the CTI correction is unavailable or has not been applied to
the BI chips.
Figure 3: A comparison of old and new
corrections for the non-CTI-corrected BI chip ACIS-S3 at the
aimpoint. The changes here are minor as well.
Hence the new Epoch 34 T_GAIN corrections are basically
necessary for optimal application of the time-dependent gains for the
most recent Chandra ACIS data. Hence the continued need for our period
upgrades to these data.
B. ACIS -110C OSIP vN0007
The recent release of the ACIS -110C T_GAIN corrections for the BI
chips, and the incumbent DET_GAIN file upgrade (See CalDB
3.4.3 release notes)
has prompted the question of upgrading the corresponding OSIP (Order
Integrated Probability) tables for the BI
chips for that same ACIS Epoch (1999-09-16T00:00:00 to
2000-01-29T20:00:00, Epoch "0"). In
particular, given that the new DET_GAIN released with CalDB 3.4.3
eliminates the difficulties with the S1 gain function, a more
acceptance region, with better-definded PHA fractions in the resulting
Grating ARFs, is now
appropriate and available for installation in the CalDB.
The OSIP table depends on the ACIS RMF response generator and the
DET_GAIN settings for the chip and location in question. With CalDB
3.4.3 we have
introduced the following T_GAIN and DET_GAIN files (Table 1) derived
from the given
constituents. Table 2 gives the current and proposed OSIP -110C files
and their components. The new
file has a more precise and traceable derivation (not ad hoc limits for
the BI chips, and S1 in particular.
1: ACIS -110C T_GAIN and
DET_GAIN file released March
31, 2008, CalDB 3.4.3.
|FI CHIP INPUT
|BI CHIP INPUT
(There are no FI chip T_GAIN data for this epoch.)
2: ACIS -110C OSIP files
N0006 (current release) and
|FI CHIP DATA
INPUTS DATA SOURCE
Ad hoc adjustment to admit full order-sorting profile of HEG -1st
order. See CalDB
2.26 release notes.
The new OSIP file has been applied to several -110C ObsIDs from the
ACIS Epoch 0 period (1999-09-16 - 2000-01-28), including ObsID 1103
(Capella, HETG/ACIS-S). The figures below compare the order-sorting
images of three situations.
Fig 4a gives the order-sorting image (top: filtered, bottom:unfiltered)
for the Level 2 events file reprocessed with CalDB 3.4.2, using the
configuration of GAIN, T_GAIN and OSIP in Table 3, row 1. Fig 4b gives
the order-sorting image, but for the calibration configuration in Table
3, row 3, with CalDB 3.4.5. Fig. 4c gives the the OS image for the
latest calibration configurations, in CalDB 3.5.0, including the new
OSIP in this release. Fig. 4a shows the erratic gain function on
ACIS-S1, and the difficulty of doing order-sorting in that region. Fig.
4b illustrates the great improvement that results from using the latest
GAIN function for ACIS-S1. Fig. 4c gives the best available OSIP
filtering for the same S1 GAIN function, and these final results are
what we are moving forward with for ACIS -110C ObsIDs.
Fig. 4a: Order-sorting image of
Level 2 events file of ObsID 1103.
Previous GAIN and OSIP applied.
Fig. 4b: Order-sorting
image of Level 2 events file of ObsID 1103.
Latest GAIN and T_GAIN applied, with version N0006 OSIP. Note
the wiggles in the ACIS-S1 region are gone.
Fig. 4c: Order-sorting
image of Level 2 events file of ObsID 1103.
Latest GAIN, T_GAIN and OSIP applied. Best available
C. Radial Encircled Count Fraction tables (REEF2
format) for Level 3 source sensitivity and detection.
Level 3 processing requires
encircled counts fraction data from the
HRMA/HRC-I calibration product file
This file is linked to the HRMA Calibration
Web page http://cxc.harvard.edu/cal/Hrma/psf/index.html,
under "DATA FILE (06/24/2004)" in the first row of the first table
under the PSF heading, where it is described as "Enclosed Count
Fractions (ECF) using circular apertures based upon the SAOsac model of
the Chandra PSF."
Level 3 uses these data to approximate
the size of the PSF as a function of source energy and
the field-of-view. For the tasks described above a detailed
of the PSF is not necessary. The psf
sizes derived by Level 3 from these ECF data are not considered
scientifically meaningful as PSF sizes for the literature, and as such,
must not and will not be published for that purpose.
The CalDB REEF file:
is based on, and contains the data
file with a revised file structure
(see description in the file's ICD Specification, explicitly linked at supporting/ecficd.pdf)
the explicit reference to
the HRC-I in the INSTRUME keyword is removed, as these data will be
used in ACIS Level 3 processing as well. The file contains Chandra
parameterized as enclosed count fractions (ECF) within circular
apertures. The parameterization is given as RADIUS(ECF, THETA, PHI,
ENERGY) where ECF is a fraction of the total number of counts contained
within a given circle in a plane perpendicular to the telescope's
optical axis, THETA is the off axis angle, PHI is the azimuthal angle
(oriented in Mirror Spherical Coordinates), and ENERGY is the midpoint
energy bin (with the upper and lower ENERGY bounds ENERG_LO and
The calculation were carried out using
rays produced by SAOsac and
HRMA configuration file "orbit_XRCF+tilts_04". The simulations were
generated using the flat geometry of the HRC-I.
The radii were calculated for a set of energies (midpoint of the valid
energy range), off-axis angles, azimuthal angles (oriented in Mirror
Spherical Coordinates), and a set of ECFs:
1% steps from 0.01 to 1
Energies (in keV):
0.5, 1.0, 1.5, 2.0, 2.5, 3.0, 3.5, 4.0, 4.5, 5.0, 5.5, 6.0, 6.5, 7.0,
7.5, 8.0, 8.5, 9.0, 9.5, 10.0
Off axis angles (in arcminutes):
0.15, 0.25, 1.0, 1.5, 2.0, 3.0, 5.0, 7.5, 10.0, 12.5, 17.0, 20.0
Azimuthal angles (in degrees):
0., 45., 90., 135., 180., 225., 270., 315.
Each value for the radius is a median calculated form 100 SAOsac Monte
Carlo simulations of the PSF for a given
ECF, THETA, PHI, and ENERGY value. Statistical errors were derived from
these simulations and RADIUS_SMIN and RADIUS_SMAX contain the values
the central 68.26% of the realizations, indicating the variation in the
radius due to randomly sampling the model. No systematic errors are
included in these tables.
optics team has also run simulations
using the tilted ACIS geometry
and does not find any significant differences in the ECFs. Hence these
data are acceptable for ACIS Level 3 processing purposes as well,
though these data are not optimized ECF values for ACIS configurations.
Again, for Level 3, approximations of the PSF size are adequate.