<Return to CalDB Release Notes>
CalDB PUBLIC Release Notes
Effective Date: (UTC) 2008-09-05T16:00:00
SDP Installation Date: (UTC) 2008-09-10T15:30:00
Public Release Date: 08 September 2008
Version: 3.5.0
Release Type: SDP + PUBLIC

I. Introduction

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":

User Documentation:
"How CalDB 3.5.0 Affects Your Analysis"

II. Summary of Changes

A. ACIS -120C T_GAIN Epoch 34

LOCATION: $CALDB/data/chandra/acis/bcf/t_gain/
FILENAMES: acisD2008-02-01t_gainN0005.fits

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.

Pipes/Tools affected:
SDP/CIAO ACIS L1 processing tool "acis_process_events"

Threads affected:
"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

LOCATION: $CALDB/data/chandra/acis/cpf/osip/
FILENAMES: acisD1999-09-16osipN0007.fits

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 data.

Pipes/Tools affected:
SDP/CIAO Grating L1.5 processing tool tg_resolve_events

Threads affected:
"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 Fraction)

LOCATION: $CALDB/data/chandra/tel/hrma/cpf/rpsf/
FILENAME: hrmaD1996-12-20reefN0001.fits

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 runs 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 observation sourcelist.

- 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 3.4/4.0.1 applications.

Pipes/Tools affected:
ACIS and HRC Level 3 pipeline tool mkpsfsize, and the Level 3 S-Lang PSF size module

Threads affected:
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 Calibration 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 point (DATE-OBS).


Figure 1: A comparison of old and new T_GAIN 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 data for the aimpoint of ACIS-S3, with the BI chip CTI correctors applied.


Figure 2: A comparison of old and new T_GAIN 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 data for which the CTI correction is unavailable or has not been applied to the BI chips.


Figure 3: A comparison of old and new T_GAIN 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 Sorting and 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 essentially eliminates the difficulties with the S1 gain function, a more restrictive OSIP 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 ACIS-S1) for the BI chips, and S1 in particular.

Table 1: ACIS -110C T_GAIN and DET_GAIN file released March 31, 2008, CalDB 3.4.3.
(There are no FI chip T_GAIN data for this epoch.)

Table 2:
ACIS -110C OSIP files N0006 (current release) and N0007 (proposed).
(derived from
(derived from

Ad hoc adjustment to admit full order-sorting profile of HEG -1st order. See CalDB 2.26 release notes.
(derived from

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.

CalDB Version
acisD1999-09-16gainN0006.fits acisD1999-09-16t_gainN0005.fits
acisD1999-09-16osipN0006.fits 4b
acisD1999-09-16gainN0006.fits acisD1999-09-16t_gainN0005.fits acisD1999-09-16osipN0007.fits 4c

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 hrmaD1996-12-20hrci_ecf_N0002.fits. 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 location in the field-of-view. For the tasks described above a detailed knowledge 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 from the hrmaD1996-12-20hrci_ecf_N0002.fits file with a revised file structure (see description in the file's ICD Specification, explicitly linked at supporting/ecficd.pdf) Note that 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 HRMA PSF 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 of the energy bin (with the upper and lower ENERGY bounds ENERG_LO and ENERG_HI, respectively).

The calculation were carried out using rays produced by SAOsac and the 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 bounding 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.

The 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.

<Return to CalDB Release Notes>