Phase Principle for Measuring Location or Spectral Shape of a Discrete Radio Source

Abstract

This paper describes a phase principle for measuring the location or the spectral shape of a discrete radio source. The phase principle is relatively simple to implement and leads to a measurement of location or spectral shape which is insensitive to receiver gain fluctuations. For measuring the location of a weak, discrete radio source, the theoretical accuracy is slightly better than the theoretical accuracy resulting from the Ryle interferometer. For measuring the spectral shape of a weak, discrete radio source, the theoretical accuracy is slightly better than the theoretical accuracy resulting from either the Ryle interferometer or the Dicke radiometer. Furthermore, the implementation of the phase principle doesn't require input switching. Also, the calibration curve associated with the phase principle is independent of changes in the average receiver gains of the two receivers.

Spectrum estimation and harmonic analysis

“…As a result of this effect, there is no obvious reason t o expect the statistics of ~( f ) to resemble those of d Z ( f ) . It should also be noted that, unlike y(f), the basic periodogram: P,(f) = I y(f) 12, is not a sufficient statistic for the data, which implies that the phase information abandoned in periodogram-based estimates is essential [237], [267] 1) The insufficiency of the periodogram is clearly inherited by any estimate based on or equivalent t o the periodogram. This obviously includes both smoothed periodograms (and it is irrelevant if the smoothing is done directly on the periodogram, on the log periodogram, or by fitting a spline or rational polynomial t o it), and, because the transform of the periodogram is the sample autocovariance function, autoregressions, moving-average representations, and other decompositions based on sample autocovariances.…”