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, band passes
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.
- 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
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,
and gpscal, are discussed in Chapter 15.