Citation

Abstract

This report continues the analysis of the cross-correlation procedure used in long baseline radio interferometry begun in Technical Report 32-1526, Vol. VIL, pp. 3750. It is assumed that the radio signal is generated by a very distant, completely incoherent, extended source. For both digital and analog recording systems, the normalized cross-correlation function is derived in terms of noise temperature, fringe visibility, and bandpass overlap. For very strong point sources and accurate model delays, it is shown that the digital cross-correlation function becomes a sawtooth time function whose extrema and zero crossings agree with the sinusoidal cross-correlation function produced by an analog system. For weak sources, such as those common to most very long baseline interferometry measurements, the digital cross-correlation function is identical to the normalized analog cross-correlation function, except for a loss of 2/z in amplitude. General signal/noise (S/N) expressions are derived for both the digital and the analog cross-correlation functions. For a very strong point source, the S/N ratio in a digital system can be infinitely better than the S/N ratio in an analog system at time points of maximum correlation. However, at points of weak correlation, the digital S/N ratio is 2/z smaller than the analog value. In the case of small correlated amplitude, the digital system produces a S/N ratio that is uniformly 2/7 worse than the analog system ratio.

Details

Volume
VIII
Published
April 15, 1972
Pages
29–38
File Size
938.6 KB