|
|
|
Nancy Evans
2000-03-02
We have become aware of a number of problems which may effect the aspect solution in processed Chandra data. The various problems and information on how to determine whether your data are likely to be affected are listed below.
I. What are the problems?
A. Any ACIS-S data processed with R4CU4UPD4, R4CU4UPD5, R4CU4UPD6.1, and R4CU4UPD6.2, or processed from Dec 22, 1999 to Feb. 14, 2000 will have an 8" offset. The direction of the offset is primarily along the spacecraft Z axis.
B. Any HRC observation with non-standard FID lights, processed before Feb. 14, 2000 will have large offsets (up to 10"). The standard FID lights are 1, 3, 4. The direction of the offset is not easily predictable.
C. Any ACIS observation with non-standard FID lights, processed before Feb. 14, 2000 will have small offsets (up to 1"), because of FID position errors. Standard FID lights are 2, 4, 5.
D. Any ACIS-S observations processed with software version R4CU4UPD6.3 or later will have an offset that varies linearly from 0.0" at 1999 day 220 to 2.0 arcsec at 1999 day 310, and is steady at 2.0 arcsec after that. The CXC is working to produce a time-dependent set of boresight alignments which will remove this offset. This includes only those observations processed since Feb 15, 2000. This affect may also be present in earlier processed data where it is generally masked by the larger offset (see A above).
E. Effects B and C above refer to cases when the observation was made with nonstandard FID lights. In addition, occasionally, a FID light was lost by the processing. These observations also have aspect solutions based on nonstandard FID lights, and have the same aspect problems. If the FID light is lost throughout the whole observation, a single image at the wrong position is produced. If the FID light is lost in the middle of the observation, a double image will be produced.
F. Occasionally there are discrete jumps in the aspect solution of up to 0.5 arcsec. This will be seen in X-ray images as a double source. This is due to warm pixels in the ACA CCD and slight thermal drifts of the telescope structure. This can be checked by plotting the columns "dy" and "dz" in the aspect solution file (pcad*asol1.fits). It should be smooth and continuous. This problem will be fixed in the next release. (Time of writing: Mar. 2, 2000).
G. Observations which were processed before CXCDS version R4CU4 have unreliable absolute celestial location. The errors can be as great as 100 arcsec.
Summary:
The times and instruments for which items A thru E apply is summarized in the schematic table below. The vertical lines in the table indicate the beginning and end times of the relevant processing. That is, the second box (left to right) is for processing before Dec. 15, the third is between Dec. 15 and Feb. 15, etc.
Position Errors
Process. Launch Dec 15 Feb 15
Date 1999 2000
Instrument
---------------------------------------------------
| | |
ACIS-I | 1" N | 1" N | 0
| | |
---------------------------------------------------
| | |
ACIS-S | 1" N | 8" S,N | <= 2" S,N
| | |
---------------------------------------------------
| | |
HRC I,S | 10" N | 10" N | 0
| | |
-------------------------------------------- ------
N nonstandard FID lights
S standard FID lights
<= 2" : 0" before 1999 day 220, 2" after 1999 day 310, linearly
increasing in between
II. How can I tell if my data are affected?
The problems A thru E can depend on both the FID lights used and the date of processing. The items below tell you how to access the lists of affected data and also how to check the processing date directly.
A. Affected Data: Observers do not routinely have information about which FID lights were used for an observation, so we have prepared web pages where affected observations are listed.
The pages are self-explanatory, except for the 2 FID columns. The column ``FID light" gives the number of a non-standard FID light. An observation will sometimes show up twice in successive lines if 2 non-standard FID lights were used. The column ``FID count" gives the total number of FID lights used.
These pages also contain information about the date on which the observations were processed.
Finally, there is a table of observations for which a FID light was rejected during processing.
B. Processing dates: This can be checked directly in the header.
The version of processing software is listed in the header keyword ASCDSVER. This can be checked (for a file acisf00408N001_evt2.fits) with the command:
unix: dmlist acisf00408N001_evt2.fits head | grep ASCDSVER
which provides the output:
0007 ASCDSVER R4CU4UPD6.4 String ASCDS version numberR4CU4UPD6.4 is the software version.
The date of processing is in the keyword date. The command:
unix: dmlist acisf00408N001_evt2.fits head | grep DATE
returns the output:
0010 DATE 2000-02-15T11:26:14 String Date and time of file creation 0011 DATE-OBS 2000-02-05T14:34:36 String Date and time of observation start 0012 DATE-END 2000-02-05T15:38:33 String Date and time of observation stopThe keyword of interest is DATE, which contains the date of processing.
Last modified: 11/28/00
 |
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. |