---

---

First, ensure that your data sets have been reprocessed using a common CalDB.

Relative Astrometry and Observation Aspect Solution: is it good enough?

Chandra's absolute pointing accuracy is generally better than 0.4 arcsec. This can be improved in an absolute sense by cross-matching Chandra sources with other higher precision catalogs or in a relative sense by cross-matching one Chandra observation with another observation.

In particular, when combining observations for spatial analysis, it is crucial to address the problem of the possible systematic shift in the coordinate systems of the various event lists which could create unexpected effects and apparent shifts in the merged observations.

It is always advisable to check the astrometric alignment of observations before trying to combine them. If a shift is found then the reproject_aspect script (or the single tools) should be used to reduce the errors in the absolute or relative astrometry in a Chandra dataset. Its usage is explained in the Correcting Absolute Astrometry thread.

---

Merging the Reprocessed [and Corrected] Data

While merging different observations may seem the "best" solution, not only can there be technical problems (e.g. unmatched columns or incompatible grid sizes) but it can also lead to incorrect or misleading results. Two primary documents exist on the CIAO website with details about why to and why not to combine datasets:

• How & Why (or Why Not) to Combine Data
• Why combine data?

---