i) Arc and Flat-Field Calibrations
In principle we expect the system to be stable over sustained periods, and indeed test data obtained during the commissioning runs show this to be generally the case. However, very occasionally relatively large variations have been seen that defies explanation. While this problem remains under-investigation, `over-calibration' would seem a wise procedure to adhere to (especially as the overheads in obtaining calibration data are small).
ii) Swing Plate Mechanism
The SP mechanism is stable at the level of our uncertainties when determining centroids (i.e., less than 0.2 arcsec / 40 microns).