Phased Array Feeds

The Parkes thirteen-beam receiver is an impressive implementation of the state-of-the-art of centimeter-wave array receivers. There is not much room for improvement within the constraints of independent feeds and good off-axis efficiency.

The next big advance in array feed efficiency requires that we abandon the use of independent feeds and that we fully sample the focal plane field with small elements. This opens the possibility of arbitrarily close beam spacing, which means about a 16-fold increase in sampling density, greater on-axis aperture and spillover efficiency, and maintenance of this efficiency at large beam offsets. Present signal processing capacity now makes prototypes of these arrays feasible, but it will be a number of years before these arrays approach the sensitivity of the best single-feed receivers.

At NRAO Richard Bradley, Kamaljeet Saini, and I are constructing a 19-element prototype feed array using broadband sinuous antennas designed for the 1 to 2 GHz range (Du Hamel and Scherer, 1993; Saini and Bradley, 1996). Our intention is to try the array on the 140-foot telescope later this year. The pre-amplifiers are uncooled, and we will use a switching matrix to sample the 171 element pairs, 4 at a time with an existing FFT spectrometer. Hence, the first prototype will be relatively insensitive.

The array bandwidth, limited by a minimum element spacing of half a wavelength and the appearance of a grating lobe at the high frequency limit, is expected to be about 1.4:1. An optimized element is now in hand, and linear array tests show that mutual coupling is manageable with the element separation required. The first objective for the prototype array is to verify our array theory from measurements of aperture efficiency, reflector beam shape, and spillover using different array correlation weights.

The next planned developments are a balanced HEMT amplifier for each element and experiments with cooling the amplifiers with a distributed cooling system or a small inexpensive cooler on each amplifier. Ohmic losses in the sinuous antennas may be significant so we need to develop a method for measuring very low antenna losses. Ray Escoffier has investigated both direct and correlation-based digital signal combining networks, and as money becomes available we will construct a correlator to measure all antenna pairs simultaneously.

The long range goal is to be able to build arrays of up to 100 elements that can form more than 50 simultaneous beams with aperture efficiencies and system temperatures as good as present dual hybrid mode waveguide feed receivers. This is 5 to 10 years away.

© Copyright Astronomical Society of Australia 1997