Gaussian. For the best data, we find the
resulting FWHM is close to 5.7. While there are many
measures of source smearing, we prefer this approach over measuring
radial profiles because there is no uncertainty relating to the
position of the source center; we are normally dealing with elliptical
rather than circular distributions; and visual inspection of the two
dimensional image serves as a check on severe abnormalities.
It has been our experience that when we are able to reduce the
FWHM of the PRF, the wings of the PRF are also reduced.
If the effective PRF is evaluated for each OBI separately, the wobble problem is manifest by a degraded PRF in one or more OBIs. Most OBIs contain only the initial acquisition of the guide stars, so when the PRF of a particular OBI is smeared, it is likely to be caused by the wobble error and the solution is to perform the phased "de-wobbling".
For those cases where each OBI has a relatively good PRF but the positions of each centroid have significant dispersion, the error cannot be attributed to the wobble. We use the term "misplaced OBI" to describe the situation in which a different aspect solution is found when the guide stars are reacquired. In the worst case, multiple aspect solutions can produce an image in which every source in the field has a companion displaced by anywhere from 10 to 30 arcsec or more. When the separation is less than 10 arcsec, the source can appear to have a tear drop shape (see Sect. 7.1.1) or an egg shape. However, depending on the number of different aspect solutions, almost any arbitrary distortion to the (circularly symmetric) ideal PRF is possible. The fix for these cases is simply to find the centroid for each OBI, and shift them before co-adding (e.g., see Morse et al. 1995).
Copyright The European Southern Observatory (ESO)