Alma Memo #419 the Y+ Long-baseline Configuration to Achieve High Resolution with Alma

The ALMA array configuration design has been split into a compact array, an extended array, and a “zoom” spiral array which aims to connect them. The currently accepted extended array is a 14 km ring around Chascón. Based on Fourier plane coverage, resolution, operations, engineering, and environmental arguments, we present an alternative to the extended configuration: a Y-shaped set of zooming configurations, which naturaly grows out of the spiral zoom configurations. This set of preliminary configurations indicates a savings of about 20 km in roads and cables, and 35 fewer pads compared to the current ring design. It has smoothly varying resolution, which permits the observer to tailor the resolution to his requirements. Reconfiguration is much smoother with the Y configuration than the 14 km diameter ring. The most extended Y configuration obtains the same resolution as the 14 km ring array. However, in order to achieve these advantages, together with the highest possible resolution for the ALMA interferometer, we need to explore the possibility of extending the Science Preserve by at least two kilometres to the west of its current western boundary. 1.A Brief History of “Y” and “O” Configurations This section is a bit pedantic, but it is instructive in that it shows us how we got to the current design with a 14 km ring array around Chascón, and how we have a 4.5 km triangle around the Conway spiral configuration. There is nothing particularly fundamental about these decisions. On the other hand, using a Y configuration for the longest baselines is a viable alternative, which is in keeping with much of the design philosophy behind the Conway spiral configurations. Many historical and even modern radio telescopes have the linear E-W configuration. The main advantage of the E-W configuration is simplicity. The most unfortunate aspects of an E-W array are the inability to perform snapshot observations and the inability to image low declination sources in two dimensions. Adding a N-S spur to the E-W track can fix both of these problems, and a Y can be thought of as a further modification. The experience at the VLA indicates that the ability to image far southern sources and to make snapshot observations has been greatly beneficial. However, there are problems with the VLA's configurations. Any Y arrangement of antennas will have more short baselines than long baselines, and the exponential distribution on antennas along each arm further exacerbates this situation. In addition, the regularity of the placement of the antennas at the VLA results in very large sidelobes in the snapshot point spread function (PSF). These two problems led Cornwell (1986) and Keto (1992) to write algorithms which sought to spread out the Fourier plane samples more. Cornwell's circular configuration and Keto's Reuleaux triangle configuration resulted in uniform (u,v) distributions with a maximum number of long baselines, and the ring became the leading configuration design for the MMA during the period 1986-1996. In the mid 1990's, some concern developed over the uniform (u,v) coverage which the ring arrays produced. Specifically, the uniform coverage resulted in very large near-in sidelobes in the PSF, and it was thought that a Gaussian coverage might be better, since we often aim to achieve a Gaussian PSF restoring beam (Holdaway, 1996a). Defenders of the uniform coverage held on for many years. The main argument for the ring arrays seemed to be that it afforded the highest resolution for any given maximum baseline length. Supporters of the filled-type configurations, which produced a more Gaussian coverage, responded by saying that if you need higher resolution, go to a more extended configuration. Of course, that argument breaks down for the most extended configuration, and it was conceded that a ring array had a place for the most extended configuration.