Theory of Random Spurs and Spectrum-Collapse from SDR Phase-Noise and Carrier-Phase Measurements

Abstract

The underlying processes causing random-spurs and spectrum-collapse for phase noise spectrum measurements are here investigated using low-cost SDR (Software-Defined-Radio) signal processing techniques. Suitable windowing and long-record-averaging for the SDR FFT, with and without phase-locking, provides a substantial cure. Metastability-theory is derived and found to explain long term persistence of spurs and collapse, and when they are likely to occur. For oscillator phase noise measurement, the dynamic-range is limited by the noise figure of the SDR ADC and not by its number of bits, provided that sufficient oversampling is used. Micro-hertz phase noise can be measured up to one hundred times faster by measuring carrier phase-difference against a free-running stable reference source suitably tuned to the source under test. A novel method of signal compression promises ~20 to 30 dB increase of dynamic range for phase-noise measurement of oscillators. The target > 150dbc at 0.lHz offset, using a low-cost SDR, has so far been achieved.