National Radio
Astronomy Observatory

4.3.5 Correction for delay and antenna position offsets using gencal

The gencal task provides a means of specifying antenna-based calibration values manually. The values are put in designated tables and can be applied to the data on-the-fly in solving tasks and using applycal.

The gencal task has the inputs:

Current antenna-based gencal options (caltype) are:

• ’amp’ — amplitude correction
• ’ph’ — phase correction
• ’sbd’ — single-band delay (phase-frequency slope for each spw)
• ’mbd’ — multi-band delay (phase-frequency slope over all spw)
• ’antpos’ — ITRF antenna position corrections (automatic parameter lookup is supported)
• ’antposvla’ — VLA-centric antenna position corrections

The calibration specifications cannot be time-variable in the present implementation. Calibration values can be assigned to each spw, antenna and pol selection, where applicable. The list of calibration values specified in parameter must conform to the range of spectral windows, antennas, and polarizations specified in spw, antenna and pol, with the values specified in order of the specified polarizations (fastest), antennas, and spectral windows (slowest). If any of spw, antenna, or pol are left unspecified (empty strings), the values specified in parameter will be assumed applicable to all values of the unspecified data axes. The output caltable will otherwise assume nominal calibration values for unspecified spectral windows, antennas, and polarizations. Note that antenna position corrections formally do not have spectral-window or polarization dependence; such specifications should not be used with ’antpos’.

The same caltable can be specified for multiple runs of gencal, in which case the specified parameters will be incorporated cumulatively. E.g., amplitude parameters (caltype=’amp’) multiply and phase-like parameters (’ph’, ’sbd’,’mbd’,’antpos’) add. Parameters for ’amp’ and ’ph’ corrections can be incorporated into the same caltable (in separate runs), but each of the other types require their own unique caltable. A mechanism for specifying manual corrections via a text file will be provided in the future.

Two kinds of delay corrections are supported. For caltype=’sbd’, the specified delays (in nanoseconds) will be applied locally to each spectral window, referring the derived phase corrections to each spectral window’s reference frequency (where the phase correction will be zero). The phases in each spectral window will nominally be flattened, but any phase offsets between spectral windows will remain. (These can be corrected using caltype=’phase’, or via ordinary spectral-window-dependent phase calibration.) For caltype=’mbd’, the derived phase corrections are referred to zero frequency. This causes a correction that is coherent over many spectral windows. If the data are already coherent over many spectral windows and share a common multi-band delay (e.g., JVLA data), caltype=’mbd’ corrections will maintain this coherence and flatten the frequency-dependent phase. Using caltype=’sbd’ in this instance will introduce phase offsets among spectral windows that reflect the multi-band delay.

For antenna position corrections (caltype=’antpos’), the antenna position offsets are specified in the ITRF frame. If the antenna field is left empty, gencal will try to look up the appropriate antenna position offsets at the time of the observation from the JVLA baseline webpage http://www.vla.nrao.edu/astro/archive/baselines/.

For VLA position corrections in the VLA-centric frame, use caltype=’antposvla’, and gencal will rotate them to ITRF before storing them in the output caltable.

The sign and scale convention for gencal corrections (indeed for all CASA caltables) is such that the specified parameters (and as stored in caltables) are the factors that corrupt ideal data to yield the observed data. Thus, when applied to correct the data, their effective inverse will automatically be taken. I.e., amplitude factors will be divided into the data on correction. Phase-like parameters adopt the convention that the complex factor for the second antenna in the baseline is conjugated, and then both antenna factors are divided into the data on correction. (These conventions differ from AIPS in that multiplying correction factors are stored in AIPS calibration tables; however, the phase convention ends up being the same since AIPS conjugates the complex factor for the first antenna in the baseline.)

The following series of examples illustrate the use of gencal.

For the dataset ’data.ms’, the following sequence of gencal runs introduces, into a single caltable (’test.G’), (1) an antenna-based amplitude scale correction of 3.0 for all polarizations, antennas, and spectral windows, (2) phase corrections for all spectral windows and polarizations of 45 and 120 degrees to antennas EA03 and EA04, repectively, (3) phase corrections for all spectral windows of 63 and -34 in R (only) for antennas EA05 and EA06, respectively, and (4) phase corrections for all spectral windows of 14, -23, -130, and 145 degrees for antenna/polarizations EA09/R, EA09/L, EA10/R, and EA10/L, respectively:

In the following example, delay corrections in both polarizations will be adjusted for antenna EA09 by 14 nsec in spw 2 and -130 nsec in spw 3, and for antenna EA10 by -23 nsec in spw 2 and 145 nsec in spw 3:

In the following example, antenna position corrections in meters (in ITRF) for antenna EA09 (dBx=0.01, dBy=0.02, dBz=0.03) and for antenna EA10 (dBx=-0.03, dBy=-0.01, dBz=-0.02) are introduced. Note that three parameters are required for each antenna. The antenna offsets can be obtained for the ’JVLA/VLA Baseline Corrections’ web page: ttp://www.vla.nrao.edu/astro/arcive/baselines. The table given on this webpage has a format like:

If your observations fall in between the ’Antenna Moved’ and ’Put_In_’ dates of a given antenna, you may choose to apply the offsets in that table; the ’Put_In_’ time stamp marks the date where the more accurate solution was introduced in the data stream directly and no correction is required anymore. In gencal the offsets will be inserted as:

In the following example, antenna position corrections (in the traditional VLA-centric frame) will be introduced in meters for antenna EA09 (dBx=0.01, dBy=0.02, dBz=0.03) and for antenna EA10 (dBx=-0.03, dBy=-0.01, dBz=-0.02) These offsets will be rotated to the ITRF frame before storing them in the caltable.

More information about CASA may be found at the CASA web page

Copyright © 2010 Associated Universities Inc., Washington, D.C.

This code is available under the terms of the GNU General Public Lincense