Cost-effective aperture arrays for SKA Phase 1: single or dual-band?

TL;DR

This paper compares the costs and performance implications of single versus dual-band aperture arrays for SKA1-low, highlighting cost similarities and the impact of beamforming choices on system flexibility and scientific capability.

Contribution

It provides a first-order cost analysis of single and dual-band SKA1-low implementations, exploring how design choices affect overall costs and scientific performance.

Findings

Costs are similar for single and dual-band arrays but have high uncertainty.

Digital processing costs are higher for single-band arrays.

Tile beamforming can reduce costs but limits flexibility.

Abstract

An important design decision for the first phase of the Square Kilometre Array is whether the low frequency component (SKA1-low) should be implemented as a single or dual-band aperture array; that is, using one or two antenna element designs to observe the 70-450 MHz frequency band. This memo uses an elementary parametric analysis to make a quantitative, first-order cost comparison of representative implementations of a single and dual-band system, chosen for comparable performance characteristics. A direct comparison of the SKA1-low station costs reveals that those costs are similar, although the uncertainties are high. The cost impact on the broader telescope system varies: the deployment and site preparation costs are higher for the dual-band array, but the digital signal processing costs are higher for the single-band array. This parametric analysis also shows that a first stage of analogue tile beamforming, as opposed to only station-level, all-digital beamforming, has the potential to significantly reduce the cost of the SKA1-low stations. However, tile beamforming can limit flexibility and performance, principally in terms of reducing accessible field of view. We examine the cost impacts in the context of scientific performance, for which the spacing and intra-station layout of the antenna elements are important derived parameters. We discuss the implications of the many possible intra-station signal transport and processing architectures and consider areas where future work could improve the accuracy of SKA1-low costing.