CALIBRATION INFORMATION

Traditional calibration of VLBI fringe amplitudes requires knowing, for each VLBI station, the on-source system temperature in units of Janskys (see the ``VLBA Observational Status Summary''). For instructions on applying the VLA calibration data in the NRAO AIPS package, see ``A Step-by-Step Recipe for VLBA Data Calibration in AIPS''.

For single-antenna VLBI, the system temperature $T_{sys}$ on source is monitored in degrees Kelvin and can readily be converted to Janskys by dividing by the nominal antenna gain in units of K Jy$^{-1}$ listed below. It may also be desirable to apply corrections for the position dependent gain, or ``gain curve'', of the antenna; see Section 11. No automatic procedure is available for measuring an antenna temperature $T_{ant}$ on source.

For phased-array VLBI, a nontraditional calibration quantity is monitored: the ratio $T_{ant}/T_{sys}$ on source. Once $T_{ant}/T_{sys}$ has been corrected for source-to-source flux density differences, it can be expressed as a system temperature in Janskys and used to correct gain variations like those modelled for the phased array in Figure 1, plus gain variations due to other effects (e.g., pointing errors or position-dependent gains of VLA antennas). If $N_{ant}$ antennas are in use, then $T_{ant}/T_{sys}$ can be estimated from the average of the $N_{ant}(N_{ant}-1) / 2$ correlation coefficients measured by the VLA, provided the source is unresolved by the instantaneous synthesized beam.

During VLBI programs, the VLA's Modcomp computers acquire and log the calibration information with program VLBLOG. These logged data are filtered by AOC staff with program YCAL, which writes ASCII VLBI calibration files that can be read by program ANTAB in the NRAO AIPS package. Some of the information provided is for you to read, and follows the comment indicator ``!'' so it will not be read by ANTAB.

YCAL delivers the VLA information described below.

Headers.

Headers at the beginning of each scan give various scan-fixed parameters plus weather station data at the scan start.

Calibration data.

During a scan, calibration data are usually given at 1-m intervals. These calibration data are derived from a single VLA record, usually of duration 10 s. YCAL does not do any time averaging. Time stamps are in IAT, the time used at the VLA. YCAL gives the value of IAT-UTC for conversion to UTC if required.

Single-antenna VLBI.

YCAL delivers columns of day number, IAT, $T_{sys}$, and the assumed receiver calibration temperature $T_{cal}$, with the last quantity being behind a comment indicator. Some of the $T_{cal}$ values are based on measurements at the standard VLBI frequencies using hot and cold loads; such measurements are probably accurate to about 10%. YCAL also delivers a nominal antenna gain of 0.071 K Jy$^{-1}$ at 0.33 GHz, 0.098 K Jy$^{-1}$ at 1.7 GHz, 0.123 K Jy$^{-1}$ at 5.0 GHz, 0.112 K Jy$^{-1}$ at 8.4 GHz, 0.103 K Jy$^{-1}$ at 15 GHz, 0.071 K Jy$^{-1}$ at 22 GHz, and 0.062 K Jy$^{-1}$ at 43 GHz. These nominal gains are based on nominal antenna efficiencies tabulated in the ``VLA Observational Status Summary''. Finally, since 2000 November YCAL delivers normalized power gain curves for whichever single VLA antenna was used for VLBI. These gain curves are flat at 5.0 GHz and lower frequencies but not at higher frequencies. All available gain curves at these higher frequencies are plotted and tabulated in Section 11. It is known that not all antennas have the nominal gain quoted for 43 GHz; these tables will thus be updated to reflect the true gains when available.

Phased-array VLBI.

YCAL delivers columns of day number; IAT; $T_{ant}/T_{sys}$ for the array; and the absolute value of the ratio of the real part of the array vector average ($REAL$) to the imaginary part of the array vector average. The last quantity is behind a comment indicator. YCAL assumes $T_{ant}/T_{sys} = (0.969 \times N_{ant} \times REAL) / 256$ for the array, where $REAL$ was described above; $N_{ant}$ is the number of antennas delivering any data, good or bad, to the vector average; and 0.969 is the duty cycle of the VLA waveguide transmission system.¹ Low absolute values for the ratio of the real and imaginary parts of the vector average can be used as a diagnostic on the quality of $T_{ant}/T_{sys}$, by indicating that the array is still phasing up during the first 1 or 2 m of a scan in mode VA, or that $T_{ant}/T_{sys}$ may be too noisy to be useful for calibration during a scan in mode VX. Another diagnostic is that during a scan in mode VA, once the array is phased, $T_{ant}/T_{sys}$ should remain stable unless the VLA operator removes or adds antennas.