A bug was found in the makepi files in the CALDB (used in the xispi command to calculate the PI and grade values from PHA) version 20111227, which affects the XIS0 and XIS3 data taken with a window option in the period between 2011-12-29 and 2012-07. Data procesed with the CALDB version 20120703 or newer are corrected for this issue already. Users are requested to reprocess their own data following the recipe prescribed below. Should you have any questions, please contact xishelp_astro.isas.jaxa.jp.
Cause
For the data taken with the Normal clocking mode with a window option, an additional gain correction is made to match with the gain for the Normal clocking mode (no option) in the pipeline processing. The additional correction was mistakingly NOT made for the XIS0 and XIS3 data in the makepi version 20111227.
Affected data
The XIS0 and XIS3 data in the following observations were affected.
| # |
Object |
Obs data |
Seq number |
PI |
| 1 |
PERSEUS_1_4_WIN |
2012-02-08 |
106007020 |
TSUJIMOTO |
| 2 |
E0102-72_1_4_WIN |
2012-04-23 |
107003010 |
TSUJIMOTO |
| 3 |
PKS2155-304 |
2012-04-27 |
107010010 |
ISHIDA |
| 4 |
GX 304-1 |
2012-01-31 |
406060010 |
POTTSCHMIDT |
| 5 |
4U 1705-44 |
2012-03-27 |
406076010 |
DI SALVO |
| 6 |
AX J1846.8-0240 |
2012-04-01 |
407019010 |
MAKISHIMA |
| 7 |
4U 1630-47 |
2012-02-13 |
906008010 |
TOO BY NEILSON |
| 8 |
EXO 2030+375 |
2012-05-23 |
407089010 |
FERRIGNO |
| 9 |
3C273 |
2012-07-16 |
107013010 |
ISHIDA |
Effect
The gain difference (Normal clocking mode with 1/4 window option) - (Normal clocking mode with no option) is shown in the figure below using the Perseus cluster data taken with both setups. The same XIS0 and XIS3 data were processed using the two makepi versions for comparison. A similar effect is expected for data taken with the 1/8 window option, though we have not confirmed it yet.
At high energies, the gain of the Normal clocking mode with a window mode is different from that of the Normal clocking mode (no option) by up to 20-30 eV in the distributed data with the old makepi version 20111227. At low energies, the difference is not as large as high energies. These differences are corrected after the data are reprocessed with the recipe below using the new makepi version 20120527.
Recipe
We ask users to reprocess their own data with the following recipe.
) f1=ae${seqnum}xi${sensor}_*_uf.evt .. unprocessed event file found in ${seqnum}/xis/event_uf/
) f2=ae${seqnum}xi${sensor}_*_new.evt .. output file with the correction.
) f3=ae${seqnum}xi${sensor}_*.hk.gz ... hk file foound in ${seqnum}/xis/hk/
) f4=ae_xi${sensor}_makepi_20120527.fits.gz ... makepi file available below.
) xispi infile=${f1} outfile=${f2} hkfile=${f3} makepifile=${f4}