The STIS CCD mirror projects the center of the standard aperture onto the CCD detector near row 517. The operational and calibration software assumes that all CCD spectroscopic gratings will project the target at this same location in the cross-dispersion direction, apart from small differences of 1 - 2 pixels due to the non-repeatability of the STIS grating mechanism. However, in reality the alignment of many of the CCD gratings deviates noticeably from this. Most of these offsets are sufficiently small that they have no operational consequences, and they do not affect the calibration because the WAVECAL task in CALSTIS corrects for such cross-dispersion zero-point offsets at the same time that it corrects for offsets in the dipsersion direction. However, two settings have had unusually large offsets in the cross-dispersion direction since STIS was launched. The G430M 5216 centered the target near row 532, while the G750M 6094 was centered near row 490. When used with small sub-arrays, these offsets could cause part of the target spectrum to fall outside the sub-array potentially ruining the science. In some cases, GOs had to specify extra offsets in their phase 2 proposal to compensate for these misalignments.
Corrected alignments were installed in HST on March 4, 2010 and tested using internal exposures taken as part of the CAL/STIS program 12078. Operational workarounds when using small subarrays with these central wavelength settings are no longer required.