Unless my expectations are wrong, I noticed that the results while using excludeBaselines are not as expected.
Running PolConvert in an EVN observation, using a single scan to calculate the corrections, and with antennas A, B, C, D, E,..., where A and B recorded linear polarization.
Because of unrelated reasons (A and B did not observe the same all subbands), a combined approach of using in the same run the reference antenna C to solve for the two linear antennas at the same time was not producing optimal solutions for the subbands that did not overlap between A and B.
To solve it, I excluded the baseline between both of them :
excludeBaseline = [['A', 'B'], ['B', 'A']]
(note I first tried with only one list, with equal results)
With this, the results were still showing the issues coming from the non-overlapping subbands between A and B.
However, if I run polConvert first on A (excluding antenna B), and in a second iteration on B, the results are optimal.
In my opinion, meaning that B still had an effect on A when computing the conversion, even if the baseline was excluded.
Unless my expectations are wrong, I noticed that the results while using
excludeBaselinesare not as expected.Running PolConvert in an EVN observation, using a single scan to calculate the corrections, and with antennas A, B, C, D, E,..., where A and B recorded linear polarization.
Because of unrelated reasons (A and B did not observe the same all subbands), a combined approach of using in the same run the reference antenna C to solve for the two linear antennas at the same time was not producing optimal solutions for the subbands that did not overlap between A and B.
To solve it, I excluded the baseline between both of them :
(note I first tried with only one list, with equal results)
With this, the results were still showing the issues coming from the non-overlapping subbands between A and B.
However, if I run
polConvertfirst on A (excluding antenna B), and in a second iteration on B, the results are optimal.In my opinion, meaning that B still had an effect on A when computing the conversion, even if the baseline was excluded.