TABLE OF CONTENTS A variety of errors, both random and spatially correlated, are introduced across the array of imperfect components and signal distribution networks, and these reduce the precision of the array excitation. An array illumination, designed to produce very low sidelobes without errors, may result in only modest sidelobes in the presence of phase and amplitude errors. If the errors are due to tolerance limits on the individual devices, it is usually possible to ensure that the errors have zero mean, at least at the array center frequency of operation. For example, an error in the power divider network that results in a progressively increasing phase error across the array can often be compensated for by measuring the error and resetting the phase shifters to correct for the power divider error. If, however, the power divider error is due to transmission line length errors, then the phase shifter correction will only compensate at center frequency. More serious yet, if errors are correlated from element to element or across large sections of the array, then the resulting radiation pattern can have large, distinct sidelobes.
Usually, it is the intent of the designer to ensure that all correlated errors are removed, so that all that remains are the residual, uncorrelated phase and amplitude errors limited by the ultimate precision of the components. The remaining errors are treated as random, and the residual (average) sidelobe errors, peak sidelobe expectation, gain degradation, and beam pointing error are estimated by statistical procedures. Results of this...
