Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.atnf.csiro.au/computing/software/miriad/userguide/node96.html
Дата изменения: Unknown
Дата индексирования: Mon Apr 11 01:51:24 2016
Кодировка:

Поисковые слова: п п п п п п п п п п п п
An alternative approach to calibration


An alternative approach to calibration

This calibration approach is recommended when dealing with high frequency data from the old ATCA correlator, especially at 3mm wavelengths. The calibration at these wavelengths is more difficult because of weak calibrators, poorer system sensitivity and more challenging atmospheric conditions. The approach described here exploits the dual frequency bands that were available with the old correlator. The steps are as follows:

  1. The first step is not a step at all! Unlike the conventional approach, the approach described here does not split the data into separate single source, single frequency files before calibration. Rather the current approach assumes that you have a single Miriad dataset containing bandpass calibrator, phase calibrator, flux density calibrator and program source. We focus on data having two simultaneous frequency bands here, but the approach works equally well for one or many simultaneous frequency bands within the dataset.

  2. The first real step is akin to a bandpass calibration step. When dealing with two frequency bands, each with two polarisation products (XX and YY), it is generally appropriate to assume that the bandpasses of the bands and the phase offsets between bands and polarisations are constant with time.

    To determine the bandpasses and the phase offsets, you need to use mfcal on an observation of a strong continuum source. Typically this would be 1253-055, 1921-293 or 0537-441. Typical inputs are as follows:

    MFCAL
    in=vela.fixed.uv Input multi-source, dual-band dataset.
    select=source(1921-293) Select the strong continuum ``bandpass'' calibrator.

    The important output of this step is a bandpass calibration table which will contain the relevant phase offsets. The output also contains antenna gain calibration tables, but this will be overwritten below.

    In doing this step, you may well want to avoid using some edge channels in the bandpass calibration process. To achieve this, you can use the edge parameter of mfcal. The edge parameter sets the number of channels to drop for all bands equally. It does not allow you to set different number of channels to drop in the different bands. If the channels and bandwidth characteristics of the different bands differ significantly, the edge parameter probably does not give you sufficient flexibility. In this case, you could use uvflag to explicitly flag edge channels.

  3. The next step is to determine the antenna gain calibration, again using mfcal. To do this, you will want to select those data for the phase calibrator which will give the best overall sensitivity appropriate to your observation.

    Note in this calibration scheme, it is implicitly assumed that the antenna gain changes are common to both frequency bands. This is a reasonable approximation given that the frequency separation between the bands is at most 2.7 GHz and potentially much less. Consequently the fractional frequency separation is at most a few percent, and so the calibration change between bands can be assumed to be minimal.

Typical inputs for mfcal follow. This assumes a continuum source as the phase calibrator and does a joint solution for the two polarisations:

MFCAL
vis=vela.fixed.uv Input multi-source, dual-band dataset.
select=source(1622-297) Select phase calibrator.
options=nopassol Do not solve for bandpass.
stokes=i Do a joint solution for both polarisations.
To derive a separate solution for the two polarisation products, the stokes keyword would be left unset. Note the use of options=nopassol. This causes mfcal to not attempt to solve for the bandpass and phase offsets again, but rather to apply the previously determined bandpass and phase offset solutions.

Typical inputs when using an SiO maser would be as follows: Typical inputs to mfcal are as follows:

MFCAL
vis=vela.fixed.uv Input multi-source, dual-band dataset.
select=source(oceti) Select SiO secondary calibrator.
options=nopassol Do not solve for bandpass.
stokes=i Do a joint solution for both polarisations.
line=chan,20,45 Select the range of channels where the SiO
  signal is strong.

At this point, you have a dataset that has been bandpass calibrated, and whose complex gains are determined as a function of time. All the remaining steps can be completed the same way as in the conventional approach: continue on from step 6 in Section 12.3.

Miriad manager
2015-09-14