3 Observing strategy

The effective width of the CCD is 2060 pixels which corresponds to drift scans of width 23'. The selection of where to observe is constrained by the survey nature of the current project. Since specific objects are not the targets, the observing is carried out at evenly spaced declinations (every $0.25\hbox{$^\circ$ }$). This allows for a uniform overlap of 8' in declination between scans, which provides enough data to quantify the difference between two adjacent frames and calibrate the atmospheric fluctuations (see Sect. 4.3).

Each day preliminary calibrations are carried out which provide quality control information which is then fed into the observation selection programme. The basic principle of the selection programme is to maximize the lengths of the observations. The reason for doing this is to minimize the number of intervals between observations during which the telescope is moving to a new declination and the CCD is being read out and thus maximize the amount of time collecting data.

Various declination zones have different priorities, so that the primary survey area (-3 $\hbox{$^\circ$ }$ to +30 $\hbox{$^\circ$ }$ declination) is completed before other areas are observed. Additionally, a $20^\circ$ zone of avoidance around the Moon is used. There is also a minimum observation length of 20 min, otherwise the Tycho 2 preliminary calibration will probably fail due to too few standards being available. A typical drift scan lasts about an hour, although exposures up to 5 hours have been made.

Additional calibrations are carried out off-line at Cambridge in order to account for the CTE and atmospheric fluctuation problems. From these it is possible to identify further problem data frames that need to be reobserved. This information is then fed back to the selection programme after the calibrations have been visually inspected.