After the passband solution has been derived and copied to the other sources in your track (e.g. phase calibrator, flux calibrator, science targets), you will want to calibrate the phase (and possibly amplitude) as it changes with time over the course of your observations. This is accomplished by self-calibrating on your gain calibrator, which was observed periodically over the course of your observations, and then using this solution for your science target(s). The necessary assumption is that the atmospheric variations are the same for the science target as for the gain calibrator, which of course depends on the angular distance between the two.
Here is an example of how phase transfer is accomplished:
1: set VIS = c0001.example.passband.miriad 2: uvcat vis=$VIS select="source(3C273),win(1),-auto" out=3C273.w1.passband.miriad 3: uvcat vis=$VIS select="source(science),win(1),-auto" out=science.w1.passband.miriad 4: selfcal vis=3C273.w1.passband.miriad refant=8 interval=5 options=phase 5: gpcopy vis=3C273.w1.passband.miriad out=science.w1.passband.miriad
In line 1, the passband calibrated data is defined as "c0001.example.passband.miriad".
In line 2, UVCAT is used to select the first window (the lower sideband of band 1) of the gain calibrator (3C273) and write out just this data in a file called "3C273.w1.passband.miriad".
In line 3, UVCAT is used to select the first window (the lower sideband of band 1) of the science target (science) and write out just this data in a file called "science.w1.passband.miriad".
In line 4, SELFCAL is used to derive the phase solution as a function of time using the gain calibrator. The reference antenna is defined as C8 (refant=8), which was chosen because it was in the center of the array and didn't have to be flagged earlier on in the data reduction process. The interval was set to five minutes (interval=5) because that is the amount of time spent on the gain calibrator in each calibrator-science target cycle. Finally, SELFCAL was only used to find the variation of phase with time, and not amplitude with time, by setting options=phase. Amplitude calibration will be discussed in a later section.
In line 5, the phase/time solution derived in line 4 is copied from 3C273 data to the science target.
The above example only calibrates one window of the calibrator and copies that solution to the same window of the science target. Of course, you will want to calibrate every window. For continuum observations, each window has 500 MHz bandwidth, and the easiest way to calibrate the data will be to derive a solution for each window of the gain calibrator and copy the solution to the same window of the science target. For most line observations, usually one or two bands will have narrow bandwidth (62 MHz or less) and the remaining band(s) will have 500 MHz bandwidth. In this case, it is probably best to derive the SELFCAL solution on the gain calibrator for the 500 MHz window and copy that to the other windows in the same sideband.
Here is an example of how to derive the phase/time solution for a narrow band dataset in which band 1 has 500 MHz bandwidth and bands 2 and 3 have 8 MHz bandwidth:
1: set VIS = c0001.example.passband.miriad 2: uvcat vis=$VIS select="source(3C273),win(1),-auto" out=3C273.w1_500.passband.miriad 3: uvcat vis=$VIS select="source(3C273),win(2),-auto" out=3C273.w2_8.passband.miriad 4: uvcat vis=$VIS select="source(3C273),win(3),-auto" out=3C273.w3_8.passband.miriad 5: uvcat vis=$VIS select="source(3C273),win(4),-auto" out=3C273.w4_500.passband.miriad 6: uvcat vis=$VIS select="source(3C273),win(5),-auto" out=3C273.w5_8.passband.miriad 7: uvcat vis=$VIS select="source(3C273),win(6),-auto" out=3C273.w6_8.passband.miriad 8: uvcat vis=$VIS select="source(science),win(1),-auto" out=science.w1_500.passband.miriad 9: uvcat vis=$VIS select="source(science),win(2),-auto" out=science.w2_8.passband.miriad 10: uvcat vis=$VIS select="source(science),win(3),-auto" out=science.w3_8.passband.miriad 11: uvcat vis=$VIS select="source(science),win(4),-auto" out=science.w4_500.passband.miriad 12: uvcat vis=$VIS select="source(science),win(5),-auto" out=science.w5_8.passband.miriad 13: uvcat vis=$VIS select="source(science),win(6),-auto" out=science.w6_8.passband.miriad 14: selfcal vis=3C273.w1.passband.miriad refant=8 interval=5 options=phase 15: selfcal vis=3C273.w4.passband.miriad refant=8 interval=5 options=phase 16: gpcopy vis=3C273.w1.passband.miriad out=science.w1_500.passband.miriad 17: gpcopy vis=3C273.w1.passband.miriad out=science.w2_8.passband.miriad 18: gpcopy vis=3C273.w1.passband.miriad out=science.w3_8.passband.miriad 19: gpcopy vis=3C273.w4.passband.miriad out=science.w4_500.passband.miriad 20: gpcopy vis=3C273.w4.passband.miriad out=science.w5_8.passband.miriad 21: gpcopy vis=3C273.w4.passband.miriad out=science.w6_8.passband.miriad 22: gpcopy vis=3C273.w1.passband.miriad out=3C273.w2_8.passband.miriad 23: gpcopy vis=3C273.w1.passband.miriad out=3C273.w3_8.passband.miriad 24: gpcopy vis=3C273.w4.passband.miriad out=3C273.w5_8.passband.miriad 25: gpcopy vis=3C273.w4.passband.miriad out=3C273.w6_8.passband.miriad
In lines 2-13 we split the passband calibrated data into 12 files, one for each source (the gain calibrator and the science target) and window (windows 1-6). In lines 14+15 we derive the phase/time solution for the gain calibrator (3C273) in the two wide (500 MHz) windows. In lines 16-18 we copy the solution from window 1 of the gain calibrator to windows 1-3 (the lower sideband windows) of the science target. In lines 19-21 we copy the solution from window 4 of the gain calibrator to windows 4-6 (the upper sideband windows) of the science target. Finally, in lines 22-25 we copy the solution from the wide bands of the gain calibrator to the narrow bands of the gain calibrator.