Although the pointing is stable and well defined, there are some systematics that are unexplained. Specifically, there are systematic deviations in altitude which very in a well defined way as a function of altitude and azimuth. These effects are not very large (up to 10 arcsec in the most extreme case), but there seems to be a potential to reach an RMS pointing in the order of 1-2 arcsec which is of general interest, but in particular for spectroscopy as it could lead to significantly reduction in time to acquire a target on the slit of fiber. At the moment it is not certain that this is a specific behaviour of the telescope, but the deviations appear to be stable in which case any pointing model should in principle correct for this. It is possible that the pointing model is wrongly implemented, but that seems unlike. We believe the most likely alternative to be that the pointing model is wrongly calculated, or more accurately, does not provide an adequate description for the precise behaviour of the telescope. This is currently being investigated.
A separately point is that in general there does not seem to be any significant difference between a pointing model based on measurements made with an instrument as compared to one made with the guide probe at the optical axis.
We are planning to migrate the telescope pointing model routine from the TCS to the sequencer, allowing for fully automated pointing tests using more suitable stars from the GSC catalog.
Thomas Augusteijn 2010-05-27