The east-west nature of the Synthesis Telescope requires that an object be tracked for 12 sidereal hours to obtain full hour-angle coverage. We describe this as an observation, although the total duration may sometimes be less than 12 hours. Twelve such observations produce a complete set of visibilities, fully sampling the available baselines; we describe such a set of data as an observation set.
In normal use the telescope is scheduled in a 4-day cycle, during which 7 back-to-back 12-hour observations with interleaved calibrations are completed, leaving approximately 8 hours per cycle for moving antennas and maintenance. The cycle is repeated 12 times to gather data on all possible spacings for the 7 fields. Occasional equipment failure and additional maintenance can add an extra day per field; an observing rate of 46 to 52 fields per year can be achieved in this mode of operation.
Amplitude calibration and phase referencing are achieved by observing a set of strong, compact calibration sources which are known not to be time variable. The derivation of calibration coefficients uses standard antenna-based algorithms, with the complication that, since the Synthesis Telescope has a very wide field-of-view, it is necessary to account for nearby extraneous sources, which would otherwise cause undesirable hour-angle (and possibly time) dependencies. To this end, visibility models of the region surrounding each calibrator have been derived from observations, and are used to remove the effects of the extraneous sources. These models are updated periodically to allow for possible source variability, and are used routinely in deriving the calibration coefficients.
The most commonly used calibrators are listed in Table 5. The flux densities used for these sources are internally consistent and referenced to 3C 286. The flux densities adopted for the calibrators, based on measurements made in 1994, are consistent with the scale of Ott et al. (1994). The same sources are used to determine polarization calibration parameters. The stability of telescope gain and phase is such that it is only necessary to observe a calibrator prior to and following each 12-hour observation. Variations on shorter time-scales are corrected using self-calibration techniques.
Source | S408 (Jy) | S1420 (Jy) | p1420 (%) |
![]() ![]() |
---|---|---|---|---|
3C 48 | 38.9 | 15.7 | 0.6 | - |
3C 147 | 48.0 | 22.0 | <0.25 | - |
3C 286 | not used | 14.7 | 9.25 | 33.5 |
3C 295 | 54.0 | 22.1 | <0.25 | - |
Calibration of the spectrometer is difficult because the narrow channel
bandwidth leads to high noise levels in individual channels. Calibration is
achieved in a two-step process, by using calibration coefficients measured
in the continuum bands. At the point where IF signals enter the central
telescope building, a broadband noise signal can be injected into all IF
paths, at a level sufficient to produce a correlation coefficient of
,
adequate to give an accurate measurement of gain and phase in
each spectrometer channel in 15 minutes. The continuum calibration measures
antenna-based gain and phase which vary relatively slowly with frequency,
but may vary with time. The IF calibration measures mostly the
channel-to-channel amplitude and phase differences, which arise largely in
the filters which define the bandpass, just before the signals are
digitized; these are quite stable with time. This IF calibration is
performed once every 4 days.
In later stages of image processing, amplitude and phase corrections are
determined for individual antennas from self-calibration of Stokes Iimages. These corrections are also applied to polarimeter images. Under
some circumstances it is beneficial to apply them to Hi images
(e.g. when there is a strong continuum source in the field).
Weighting | Frequency | Beamwidth | First sidelobe | Second sidelobe | ||
(MHz) | (half power) | level | radius | level | radius | |
Untapered | 1420 | 49'' | -13% | 1' | 5.9% | 1.65' |
Untapered | 408 | 2.8' | -13% | 3.5' | 5.9% | 5.85' |
Gaussian* | 1420 | 58'' | -2% | 1' | 2.4% | 1.65' |
Gaussian* | 408 | 3.4' | -2% | 3.5' | 2.4% | 5.85' |
* A Gaussian taper falling to 20% at 144L (617.1 m). |
The antennas are capable of tracking down to an elevation of 12,
so
the observable sky from the observatory's terrestrial latitude of 49
nominally extends from declination -29
to +89
(the latter
imposed by the telescope drive system). However, full hour-angle coverage
is possible only for declinations north of +18
.
Sources north of
declination +54
are circumpolar, and there is an overlap in
hour-angle coverage at lower culmination, making it possible to track a
circumpolar source for 26 sidereal hours without interruption.
Non-sidereal tracking rates are available for observing solar-system
objects.
Since they are equatorially mounted, the antennas track only in hour angle,
although pointing corrections are made in both hour angle and declination.
A pointing model for each antenna accounts for zero offset and ellipticity
of the position encoders, misalignment of the polar axis, and
non-orthogonality of the declination and hour-angle axes. These parameters
are empirically determined from "nodding'' observations of selected
calibrators at many hour angles and declinations. A "nod'' consists of
measurements on-source and at 4 off-source positions situated 1
north, south, east, and west of the source, to which two-dimensional
Gaussians are fitted to obtain errors in hour angle and declination. The
rms pointing accuracy across the observable sky is 2.4' in RA and
3.1' in Dec.
For a given observation, the pointing corrections at the declination of interest are calculated from the pointing model for every 30 min of hour angle, and interpolated linearly to intermediate hour angles. The tracking accuracy is then determined by a feedback loop in the control system, which measures the deviation of the antenna position from the requested position; at present errors in antenna position exceeding 3'are corrected. Such corrections typically occur on timescales of 30 min.
Copyright The European Southern Observatory (ESO)