E.6 Calibration with the SysPower table

Having done a more careful job with your editing, it is now time to discard with EXTDEST the bandpass (BP) tables and all CL tables after the one written by VLANT. Discard all SN tables, but keep the highest numbered flag (FG) table.

Because of its wide dynamic range, the EVLA does not normalize its output visibilities. To calibrate gains it records the total power when the switched noise tube is on and when it is off. These data, taken in synchronism with the visibilities, are recorded in the SysPower table of the ASDM. The OBIT program BDFIn, available to AIPS users in the new verb BDF2AIPS, reads this table and creates an AIPS SY table. The columns of this table contain POWER DIF (Gain × (Pon - Poff), POWER SUM (Gain × (Pon + Poff), and POST GAIN (Gain) columns for right and left polarizations with values for each IF.

This table is accessible to AIPS users with a number of tasks. Begin with task PRTSY to view the table statistically over time on a per IF, per antenna basis or to view scan or source median averages of one or more of the SY table parameters. Then to examine its contents in more detail in various ways, use SNPLT with OPTYPEs ’PDIF’, ’PSUM’, ’PGN’, ’PON’, ’POFF’, ’PSYS’, ’PDGN’, or ’PSGN’. You could use OPTYPE = ’MULTto examine more than one of these at one time, comparing any oddities in e.g., Psum and Pdif. Note that PSYS is especially interesting since Psum∕Pdif2 * Tcal = Tsys, the system temperature. It should reflect changes in elevation and strength of the observed source, but should be immune to adjustments to the gain of the telescope. It determines data weights in TYAPL while ∘ ---------
  PdifiPdifj divides into the visibilities. You may use EDITA (4.4.2) to edit your uv data on the basis of the contents of the SY table. Editing may be based on Psum, Pdif, Pgain, Tsys, and on the differences between these parameters and a running median of these parameters. One may also edit the SY table itself with SNEDT; the same parameters are available. LISTR can even display the SY table Psum, Pdif, system temperatures, and gain factors with OPTYPE GAINand DPARM(1) set to 17, 18, 15, or 16, respectively.

More importantly, the SY table can be used to do an initial calibration of the visibility data. Use the display programs to decide if your SY table is fine as is or needs editing. The tasks TYSMO and TYAPL (4.1.1.1) may be used with EVLA data having an SY table. TYSMO flags SY samples on the basis of Pdif, Psum, Pgain, and Tsys and then smooths Psum, Pdif and Pgain to replace the flagged samples and/or reduce the noise. You may want to do this to remove outlying bad points and to reduce the jitter in these measurements. TYSMO even applies a flag table to the SY before its clipping and smoothing operations. Be sure to plot the results to make sure that the task did what you wanted. Then use TYAPL to remove a previously applied SY table (if any) and to apply the SY table you have prepared. The result should be data scaled nearly correctly in Jy and weights in 1Jy2 in all IFs. The CUTOFF option allows you to use obviously good values from the SY table while passing the data from antennas with poor SY values along unchanged. If the SY values from some polarizations or IFs are bad due to RFI while others are good, you may copy the good values to replace the bad values using task TYCOP. The wide-band, 3-bit mode of the EVLA has serious non-linearities in the Pdif measurements which are not yet fully understood. For such data, use OPTYPE=’PGN’ to apply only the post-detection gains to the data. This will remove abrupt jumps due to changes in those gains (which will be more common with 3-bit data). Note that TYSMO and TYAPL also require a table of the Tcal values which OBIT provides in an AIPS CD table. Amplitude calibrations are not applied to EVLA data weights until they have been made meaningful by TYAPL or REWAY. Set FLAGVER in TYAPL if you want to apply your flag table once and for all.

If you have observations of the Sun, do not use TYAPL; use SYSOL instead. This task does the usual TYAPL operation on non-Solar scans, but, for Solar scans, it must determine the average gain and weight factors on those antennas having solar Tcals and then apply those averages to all antennas not equipped with Solar Tcals. Use a well-edited SY table if possible. This Solar capability was made available in OBIT (version 567) and AIPS in mid-June 2017.

Now return to E.4 to repeat the bandpass and continuum calibrations with correctly scaled data with most of the RFI removed.