Quantization effects on the performance of adaptive arrays

The authors examine the quantization problems in the array weights and in the signals which are present on the array elements. The two method generally used to achieve complex weights are considered. In the rectangular method, the signals are separated in quadrature components and then weighted with real quantities. In the polar method, an attenuator and a phasor with linear control are used. The rectangular solution is shown to be a better solution for the complex weights implementation. A precision of eight bits is sufficient to place a -40-dB null in the direction of a jammer with a two-element array. A precision of six bits is sufficient to compute the optimum weights.<>