Abstract

A hardware prototype of vector spectrum analyzer, which is based on a randomized sampling strategy and associated alias-free digital signal processing (DSP) algorithms, has been fully designed and implemented. The instrument exploits a couple of identical independently programmable digital data acquisition channels, whose synchronization allows to sample the input signal at instants that can be very close along the time axis (tens of picoseconds), against a maximum single-channel average throughput rate of 250 ksamples per second (kSa/s). This architecture has shown to be suitable for the accurate implementation of the randomized periodic sampling with uniform jitter on which the instrument is based. In addition, the design solutions adopted for the sampling command generators, which exploit the interaction between a digital gate signal and a phase-controlled sinusoidal wave, together with the particular analytical definition proposed for the spectral estimators, have allowed to make practically negligible the quantum time according to which the sampling instants are generated. Experimental results are provided, which validate both the alias-free spectrum analysis technique and the capability of the architecture proposed of reliably implementing the analytical benefits deriving from the randomized sampling strategy adopted.