Contents 1. Introduction
Serial beamforming networks associated with linear array antennas are known to introduce phase dispersion due to unbalanced electrical path lengths. This results in frequency-controlled beam steering, sometimes used in radar applications [1]. This is a typical characteristic of traveling-wave and leakywave antennas. On the other hand, wideband beamforming networks associated with linear array antennas, such as parallel feeding networks, introduce a beam squint. This is due to the fact that the distance between radiating elements normalized to the wavelength varies with the frequency. Consequently, all beams deviate toward the boresight of the antenna as the frequency increases. A combination of these two phenomena could be explored to cancel the beam squint with frequency in serial multiple beamforming network designs, such as Blass [2] and Nolen [3] matrices. To do so, the beamforming network has to produce a true-time delay, compensating the time delay observed in radiation when the beam is tilted. This characteristic is naturally achieved in quasi-optic beamformers, as the wave is distributed over the antenna's aperture in the radiation mode. Well-known solutions with linear array antenna apertures include the Ruze lens [4], the Rotman lens [5], and, more generally, all bootlace lenses. Other quasi-optic solutions - such as reflector antennas, pillbox antennas [6], and the Luneberg lens [7], [8], which are characterized by a continuous antenna aperture - also display the same property. In guided-wave beamformers, this property requires more attention to achieve, particularly in the case of serial beamforming networks.
