24/05/2013

Telescope pointing update

I have made progress towards resolving the telescope pointing issue on a number of fronts over the past week.
Plate Solving - The temporary solution
On the night of the 21st of May the initial version of the plate solving integration went live on the telescope system. The system is only used for Galaxy jobs, and does not work for SSBODY job types. The new code allows the controlserver to compare the actual coordinates of the image with those requested by the job. The controlserver can this use this information to reject jobs where the pointing is too bad.

Currently the rejected jobs show up as TF: Mount error in the RTD system, but we plan to implement a set of return codes for the plate solver shortly.
RTD showing many more Mount errors than usual
The plate solving integration also produces graphs showing the angular separation and relative angles of the pointing errors for each job completed and solved in a night. Analysis of this data indicates that it may be possible to reduce the pointing error by applying an offset to the pointing of the mount. At the moment this idea is in the planning phase as it will require significant changes to the way in which we interact with the mount.

Mount Repairs - The permanent solution
The information from the plate solving graphs continues to point towards a problem with the homing of the mount which results in a fixed error in Dec pointing throughout the night. The graphs are also showing that the pointing of the Dec axis is in general significantly less accurate than that of the Ra axis. As we will be disassembling the Dec drive to investigate the homing error we have decided to take the opportunity to replace the Dec worm drive as well. The result of this work should be noticeably better pointing when the system returns to operation at the end of the summer maintenance trip.

1 comment: