The Second Release of the SIS CTI (Charge Transfer Inefficiency) Calibration File.
As you know, the FTOOL SISPI and its associated calibration file sisph2pi.fits enable the PI column in SIS event files to be filled. The calibration file is also used by SISRMG to generate the PI response matrices available in the Calibration Database. By working with PI rather than PHA, users can correct for the gain changes induced by radiation damage, as well as combine data from different chips.
The previous version of sisph2pi.fits was released last summer. We have discovered two problems associated with extrapolating to more recent data:
- There was a bug in SISPI: it was not extrapolating the CTI trend.
- The calibration of the CTI trend was not accurate enough to support much extrapolation anyway.
To correct these two problems, a new version of sisph2pi.fits has just been released. It takes care of both problems because, with the new file, extrapolation is no longer needed to fill the PI columns in recent datasets. [Note, however, that the extrapolation bug in SISPI remains and will be corrected for the September 1995 release of FTOOLS.]
This new file is based on the efforts of various SIS team members. It embodies their best understanding of the instrumental Ni line data through November 1994 and the calibration observation of Cas A taken in July 1994. In terms of gain, it is estimated to be currently accurate to about 1 per cent for SIS0-chip1. The file comes with a nominal expiration date of February 1996.
Please note, however, that:
- As before, PI does not account for DFE or RDD effects.
- Although PI can correct for time-dependent gain, it cannot, by
itself, deal with time-dependent spectral resolution. For this,
time-dependent matrices are necessary and will be supported by the next
release of SISRMG (the response matrix generator).
- There remain some inter-chip gain calibration problems.
The new file is located at:
Koji Mukai and Charles Day, ASCA GOF
If you have any questions concerning ASCA, visit our Feedback form.