The very large array: Design and performance of a modern synthesis radio telescope

Abstract

Since its development in the 1960's, the technique of obtaining high-resolution radio images of astronomical objects using Fourier synthesis has advanced sufficiently so that today such images often provide better angular resolution than is obtainable with the largest optical telescopes. A synthesis array measures the Fourier transform of the observed brightness distribution by cross-correlating the signals from antennas separated by distances up to tens of kilometers. The antennas must be equipped with low-noise receiving systems and connected together by phase-stable transmission links. Wide-bandwidth digital correlators are used to perform the cross correlation. The data-reduction algorithms and computing system play a critical role in determining the quality of the images produced by the array. The Very Large Array (VLA) synthesis telescope, recently constructed in New Mexico, consists of twenty-seven 25-m-diameter antennas arranged in a Y-shaped array. Each arm of the Y is approximately 21 km long and the antetmas can be moved to various positions on the arms by a rail-mounted transporter. The antennas are equipped with cryogenically cooled receiving systems and are interconnected by low-loss, TE01-mode, large-diameter waveguide. The cross-correlation products for each of the 351 pair combinations of antennas are measured for 4 IF signals by a 50-MHz bandwidth digital correlator. In this paper we discuss the design of synthesis arrays in general, and describe the design and performance of the VLA in particular, under the seven headings: array geometry design, sensitivity considerations, phase stability requirements, signal transmission system, delay and correlator system, control system, and data-reduction requirments. In each section, we review the underlying instrumental requirements and provide details of how the VLA was designed to meet them. Recently developed data-reduction algorithms provide effective ways of correcting synthesis images for the effects of missing Fourier components and instrumental and atmospheric amplitude and phase errors. The power of these algorithms is demonstrated using actual VLA images.