SKA-Low prototype confirms its sunflower antenna layout loses sensitivity at 125 MHz

The problem was the spiral's regularity. A Vogel layout is meant to reduce interference between neighboring antennas while keeping the station's collecting power ^[5]. It also repeats the same spacings across the whole station. The older AAVS2 station used a scattered, pseudo-random pattern, chosen to avoid unwanted signal artifacts, and it showed no such dip ^[3][10].

What I like about this experiment is the order of events. The prediction came first, from simulations run while AAVS3 was still being built ^[9]. The team, led by Shin'ichiro Asayama of the SKA Observatory, then ran the 256-antenna station both as a beamformer and as an imaging array. They pointed it at the sun, bright radio sources, the galactic plane and several pulsars ^[6][7]. A dip that shows up exactly where a model said it would, on a source as broad as the galactic plane, is hard to blame on a loose cable. The team concluded the loss is built into the geometry ^[11]. The AAVS2 comparison is between two separate prototype stations, and the phys.org summary does not say how deep the dip was.

The dip also falls in an awkward part of the band. Mutual coupling, the interference between antennas placed too close together, grows stronger below about 150 MHz ^[4]. The spiral was proposed to deal with that. The loss it introduced sits at 125 MHz, inside the same band ^[1].

Calibration produced a second, quieter result. The team built two schemes. One uses the sun as a known reference. The fuller one adds a sky-brightness model and detailed antenna response profiles, and it also works at night ^[8]. At 230 MHz, night measurements with the fuller scheme closely matched predictions. Daytime measurements missed by more, most likely because unusually high solar activity made the sun's brightness harder to predict, according to the team ^[12].

The thing this doesn't tell you is whether the replacement works on the sky. The "Perturbed Vogel" layout keeps antennas spaced to limit interference and removes the repeating pattern ^[13]. Testing it falls outside this study ^[14]. Simulation called the 125 MHz dip correctly before observation confirmed it ^[9][11]. I think that record justifies some confidence in the simulated fix. It is still a model result until a station built to it is pointed overhead at the galactic plane.

The choice has to hold at scale. Each station's antennas combine like a single large dish, so their arrangement shapes what the telescope can detect. The finished SKA-Low will link 512 stations across an area about 75 km wide ^[2], observing between 50 and 350 MHz from Western Australia ^[1]. The AAVS3 work, together with lessons from AAVS2, has shaped the observatory's testing process and supported the AA0.5 milestone, the first stations to come online ^[15].