Determining Calibration Solutions

There are currently four main tasks which solve for calibration parameters -- each with their own strengths and weaknesses. These are:

• mfcal  : This determines antenna gains, bandpasses and delays given a point-source calibrator. Multi-frequency observations are supported. Although it can handle dual polarisation data, it assumes the calibrator is unpolarised, and that the polarisation leakages are 0. This assumption does not affect the bandpass or delay that it finds.

• gpcal  : This determines antenna gains and leakages given a point-source calibrator. The polarisation properties of the calibrator can be determined as part of the solution process. Gains and leakages are assumed to be independent of frequency.

• selfcal  : This determines a single set of antenna gains given a model of the visibility data. The antenna gains are assumed to be independent of all observing parameters other than time and antenna number. Consequently, its handling of multi-frequency and polarisation data is fairly simple. It can handle multiple pointings and sources ( e.g. mosaiced observations).

• gpscal  : This is useful when self-calibrating strongly polarised sources measured with a telescope with linear feeds ( e.g. the ATCA). Given models of Stokes I , Q , U , and V , as well as the leakage parameters, this solves for the appropriate antenna gains.

We will now describe mfcal and gpcal in some detail. This is largely for reference -- the following chapter will give a step-by-step approach to the calibration process. The self-calibration tasks, selfcal and gpscal , are discussed in Chapter 14.