EM Harp simulations of SKA-Low layouts

The purpose of this feature is to provide EM simulations of SKA-LOW station beams in support of the final choice of layout. This is needed urgently by infrastructure contractors to inform the final selection of station positions and rotations.

Criteria for a good station layout are:

1. Low sidelobe levels in the station beams. In particular, this requires a lack of a repeating pattern in the layout (which tends to lead to significant grating responses).
2. Spectral smoothness. This is affected by the presence of mutual coupling between antennas.  Resonances caused by interactions between neighbouring antennas in a station are particularly problematic.
3. Minimal instrumental polarization.

In order to make a rigorous choice between different layout choices, EM simulations are required. These predict station beams (which can in turn be used in array-level simulations with tools such as OSKAR) and embedded element patterns (needed for calibration).  To address criterion 2, simulations at fine frequency intervals are required (preferably around 1 MHz). 

The types of layout remaining under consideration are:

• Vogel (spiral)
• Pseudo-random

The Vogel layout was originally proposed to mitigate the effects of resonances at 55 and 78 MHz between close pairs of antennas, which are quite marked in the pseudo-random layout.

The great advantage of using HARP for this work is that the calculations can be split into two steps (for each frequency):

• antenna dependent (using FEKO)
• layout-dependent (using a fast, in-house solver).

Once the first step is complete, evaluations of different layouts are very fast. The method has been validated against simulations produced by the Curtin group (which use only FEKO).

The first step is currently under way for the SKALA4.1 antenna over the full frequency range of SKA-LOW (50-350 MHz) at 1 MHz intervals. This step is estimated to be complete ~end June.

The investigations covered by this feature, in order of priority, are:

• Investigation of potential modifications to the Vogel layout to reduce the effects of a coupling effect around 125 MHz leading to a drop in gain near the zenith and increased far-sidelobe levels. This effect appears to result from a residual regularity in the distribution of baselines which is not present in the pseudo-random configuration.  If this investigation is successful, then the revised layout can be tested on AAVS3.
• Production of a full set of station beams and embedded element patterns for the AAVS3 (Vogel) layout and any modifications as a result of the preceding step.  Check against the criteria 1-3 to make sure that no significant problems have been missed as a result of coarse frequency sampling in earlier simulations.
• Production of a full set of station beams and embedded element patterns for the AAVS2 station (one realisation of a pseudo-random layout). Again, the results will be compared with observations.

The outputs are a report describing the methods and results and a full set of station beams and EEP's (format tbd).