The geometry of the array is illustrated in Figure 1. Ten of the telescopes (labeled 0 to 9) are on fixed mountings, 144 metres apart; the four (2 x 2) remaining dishes (labeled A, B, C, D) are movable along two railtracks. One, 300 m long, is adjacent to the fixed array and another, 180 m long, is 9 x 144 m eastwards. The movable dishes can be used at any position of the rail tracks (minimum distances 9A = 36 m, AB = 48 m, CD = 48 m), although not all are calibrated to mm accuracy, and it will usually be preferable to use one or more of the standard configurations described below. In the array, the baselines can extend from 36 m to 2.7 km.
of the WSRT array (Click here for a large version)
The antennas are equatorially mounted 25-m dishes (with f/D ratio of 0.35). This type of mounting ensures a fixed orientation of the receiving linear dipoles with respect to the sky. The maximum hour-angle is +/-6 hour at any declination; this limits observations of circumpolar sources. Table 1 summarizes the telescope parameters at the different frequencies that can be observed (see Sec. 5 Technical Information). The slew rate of the antennas is 16-18 deg/min on both axes.
|Wavelength||Aperture||Single antenna FWHM||Synthesized beam|
|(cm)||Efficiency||field of view (deg)||(arcsec)|
Table 1. Telescope parameters
The choice of the suitable array configurations depends strongly
on whether a number of 12h observations are requested (for a single field).
If this is the case, the best (u,v) sampling is obtained by observing
using the so called ``traditional'' configurations (see below), each 12h run taken
with a different RT9-RTA spacing.
A) Maxi-Short. Optimum imaging performance for very extended sources within a single track observation can be obtained by adopting a RT9-RTA separation of 36 m, together with an RTA-RTB separation of 54 m, while still employing a RTC-RTD separation of 72 m. The shortest baselines with this configuration (36, 54, 72 and 90) are identical to those obtained in the past with a traditional 4x12 hour coverage. While overall sidelobe levels are not as low, image fidelity in the reconstructions of very extended sources is comparable. The (9A,9B,9C,9D) distances to specify for this configuration are (36,90,1332,1404 m).
B) Traditional configurations. The ``traditional'' WSRT configurations are based on the principle that (u,v) sampling is acquired with a uniform radial sampling increment. This is motivated by the desire to achieve high dynamic range imaging over a limited angular field-of-view (FOV, we define the FOV as the area free of self confusion due to gratings lobes) with minimal deconvolution . Given the vast improvements in both software and hardware since the construction of the WSRT, it is not as much of an issue as it was in the past (and a number of new configurations have been designed, see Sec. 3.3). However, it is in many cases still a useful consideration.
Traditional configurations are those in which RT9 and RTA have a particular separation, for example 36, 54, 72 or 90 m (or less typically 36, 48, 60, 72, 84 or 96 m or even one of 36, 45, ...99 m), while the two pairs of movable telescopes, RTA and RTB, as well as RTC and RTD, are each kept at a fixed separation of 72 m. The RTC/D pair is then moved such that the separation RT9-RTC is (9◊144 m) plus the same 36, 54, 72 or 90 m.
Each of these configurations has the property that a regular sequence of baselines with a 72-m increment is obtained beginning with the shortest spacing (e.g. RT9-RTA=36 m or RT9-RTA=90 m) and proceeding to that length plus (37◊72) m. The synthesized beam which results from a single complete 12 hour observation, has particularly strong grating ellipses at intervals which correspond to the 72 m baseline sampling increment. These reach about 10% of the peak response initially and decline somewhat with radius. At 1420 MHz, these grating responses occur at multiples of 10 arcmin (East-West), effectively defining the useful field of view to less than 10 arcminutes when imaging extended sources.
As mentioned above, the usefulness of these configurations is when datasets taken with different RT9-RTA spacing (i.e. 36, 54, 72 or 90 m) are all combined together. If a FOV of l/72 m is sufficient, a single12-h run can be employed. If twice this FOV is required then any two configurations shifted by 36 m could be employed. If four times the FOV is required then the (36, 54, 72 or 90) series should be used, if six times the FOV is needed then the (36, 48, 60, 72, 84, 96) series, and in the extreme case of eight times the FOV, then the (36, 45, ...99) sequence.
Since both the sensitivity and the number of correlated baselines have increased dramatically in the past few years (due to the new receivers and correlator), it is much less often the case that multiple 12-hour tracks are required for WSRT imaging of a field. The much larger number of baselines due to fixed-fixed telescope pairs (with n◊144-m separation) that now are also correlated is always present and can be used to enhance the (u,v) coverage within the inner 1.5 km of the array. For example, even the traditional ``36 m'' configuration now provides a denser sampling of the inner (u,v) plane and with it's two shortest baselines (36 and 72 m) by itself provides some of the imaging performance of a 2◊12 hour observation of the past.
When multiple 12 hour tracks are required for sensitivity reasons, it is probably still appropriate to use a combination of the traditional configurations listed above to enlarge the unconfused FOV.
C) 2x48. The inner UV plane can be sampled with a uniform 48 m increment within a single track by adopting a 48 m separation for both 9A and AB. Uniform outer plane sampling with a 72 m increment requires a CD separation of 72 m. The three shortest spacings (48, 72 and 96 m) provide somewhat less sensitivity for very extended sources than the Maxi-Short option, but average sidelobe levels are somewhat reduced due to the more uniform sampling. The (9A,9B,9C,9D) distances to specify for this configuration are (48,96,1332,1404 m).
D) Mini-Short. Since shadowing (see §) can lead to severe data loss when observing at declinations below about 30∫, it is often preferable to choose a configuration optimized to minimize this effect. If the object of study is known to be compact, then such a configuration has the added advantage of minimizing data loss through interference and confusion by extended emission. By adopting a 96-m separation for 9A, AB and CD, shadowing is all but eliminated for all declinations (as can be seen in Figure 2). This choice also provides a uniform 48 m sampling interval for most of the inner UV plane, providing good sidelobe properties. The (9A, 9B, 9C, 9D) distances to specify for this configuration are (96, 192, 1332, 1428 m).
E) 2x96. If shadowing needs to be avoided, but the source declination is not too extreme (consult Figure 2) it is possible to achieve the uniform 48 m sampling of the inner UV plane but also retain the uniform 72 m sampling of the outer UV plane by adopting a 96 m separation for 9A and AB, while retaining a CD separation of 72 m. The (9A, 9B, 9C, 9D) distances to specify for this configuration are (96, 192, 1332, 1404 m).
At declinations below about +40°, shadowing will begin to affect short baseline data at large positive or negative hour angles. Thus a configuration should be selected such as to minimize this difficulty if possible. Moreover, a proper flagging of the data should be done (the exported data are not flagged for shadowing). In Figure 2 the shadowing limits for the WSRT are shown for a given spacing of the 9A baseline.
|Shadowing limits for the WSRT. For a given antenna spacing the line encloses the range of hour angles affected by shadowing.|