Analyzing timing data
The text file containing timing error data should be plotted with some graphics program. In the simplest situations, the timing errors will be well approximated by a linear function of exposure number. The slope of this linear function (which may be determined using program regress or any other linear regression program) is equal to the average accumulated timing error per exposure.
In some cases, the timing error plot will show more complex behavior, which cannot be well approximated by a linear function. This is often caused by inaccuracies in the slew time models, equipment parameter values, or the duration of focus breaks assumed in the scheduler configuration file. For example, if the assumed duration of focus breaks is significantly different from the actual average duration of a focus break, the timing error plot may exhibit sharp discontinuities (jumps) at the exposure numbers which are adjacent to focus breaks. In such situations, improving the models and parameter values in the scheduler configuration file may lead to timing error plots which may be modeled by a linear function.
Adjusting the empirical timing
Note that adjustments of the timing correction may be done after any observing run (not just the first), in an attempt to reduce the timing errors. In particular, it is useful to analyze the timing errors after any changes in equipment configuration, in the parameters in the scheduler configuration file, or in the focus parameters in the MU script configuration file.