|
|
|
Nov. 2008: A new time-dependent HRC-S gain map and LETG/HRC-S background filter have been developed. Users are strongly encouraged to use these, which will require reprocessing their data from Level 1.5 using contributed software. It is expected that the new gain map and filter will be incorporated into CIAO/CALDB sometime in 2009, along with application of the new gain map during Standard Processing.
By filtering on the pulse-height value pi as a function of
dispersed-photon wavelength,
the background rate in the HRC-S,
when used in conjunction with the LETG,
can be reduced by a factor of two to five
(depending on wavelength and the level of filtering)
by using a dmcopy command with a region filter,
as described below.
(Background in gratingless HRC-S data can also be reduced,
by about half, as explained
here.)
Note that this command will only work for CIAO 2.1
or later releases and is not yet part of automated pipeline processing.
Users who apply this filtering will need to rerun the last
processing step (tgextract)
to recreate their Level 2 pha2 file. Please see the
"Obtain Grating Spectra from LETG/HRC-S Data" analysis thread
for further information. Note that you do not need
to redo the GTI or status-bit filtering described in the thread--those
steps are only
required if you are reprocessing data that were
processed before July 2, 2001.
Three levels of filtering are available, corresponding to fractional X-ray losses of 0.5, 2, and roughly 5-10%, compared to data undergoing standard pipeline processing. The "Heavy" (5-10%) filter should not be used to create publishable spectra, as it is not well calibrated and removes up to ~15% of valid x-ray events at some wavelengths. See Filtering Errors for further information.
The filter can be run on either Level 1.5 (1a) or Level 2 event files, and is based on the relationship between event pulse height (adjusted for spatial gain variations over the detector and recorded in the pi data column) and the wavelength of 1st-order dispersed photons (the tg_lam value). The command is:
dmcopy "input.fits[events][pha=0:254,
(tg_lam,pi)=region(regionfile.fits)]" output.fits opt=all
where input.fits is the unfiltered file,
output.fits is the filtered file,
and regionfile.fits is one of the following files:
| region file | filename in CALDB 2.3 | X-ray loss (%) |
background reduction (%) |
|---|---|---|---|
| filt_light.fits | letgD1999-07-22pireg075_N0001.fits | 0.5±0.3 | 50-70 |
| filt_medium.fits | letgD1999-07-22pireg062_N0001.fits | 2.0±0.7 | 55-75 |
| filt_heavy.fits | letgD1999-07-22pireg046_N0001.fits | 5-15 | 60-80 |
The quoted X-ray losses apply for 1st-order wavelengths longer than 6 Å. The region filtering is designed to not affect data with shorter wavelengths (which includes all of 0th order) because of scattering and support structure diffraction from 0th order, which includes photons from all wavelengths. Instead, the pha=0:254 filter removes about 25% of the background within this near-0th-order region, with no X-ray losses. Further reduction of background around 0th order can be attained, but this requires some "manual intervention" and usually isn't worth the trouble since the S/N there is so high anyway.
In most cases, the medium level of filtering (2% X-ray loss) will provide the best improvement in signal-to-noise, but lighter filtering is nearly as good and with virtually no possibility of introducing spurious spectral features. Heavier filtering has very limited applications, and should never be used to create publishable spectra. We therefore recommend using the light filter as a conservative default.
As noted above, the quoted X-ray losses apply only for 1st-order photons. Losses for higher orders are correspondingly higher, and increase with order, but are still very small. For light filtering, the fractional loss is less than 10% for 8th order.
These pulse-height filters are based on a gain map which is binned on 1/2-tap (823 micron) scales. Apparent gain variations of more than 30% occur on even smaller scales, indicating that a more finely gridded gain map would allow tighter filtering. Some of these gain variations, however, seem to arise from skewed pulse-height distributions in certain (small) areas of the HRC-S. These matters are currently under investigation.
Last modified: 11/21/08
 |
The Chandra X-Ray
Center (CXC) is operated for NASA by the Smithsonian Astrophysical Observatory. 60 Garden Street, Cambridge, MA 02138 USA. Email: cxcweb@head.cfa.harvard.edu Smithsonian Institution, Copyright © 1998-2004. All rights reserved. |