A digital correlator for low-frequency signals

A digital correlator operating by polarity-coincidence is described in this thesis. Because the comparison levels for polarity determination are random variables with rectangular probability—density distribution, the output of the correlator is directly proportional to the correlation function of the input signals for all classes of signals. The computer calculates 32 uniformly distributed points of the correlogram simultaneously and is capable of operating as a computer of average transients or for calculating probability-density and cumulative distribution functions. The computer memory is provided by a 32 x 32 word-select memory core array. A theoretical study of the computer operation as a correlator is given. Two modes of sampling are considered; (l) the symmetrical—sampling mode which requires a shift-register delay line and samples taken at interval, ΔT , where ΔT is the spacing between points of the correlogram, and (2) the asymmetrical—sampling mode where delay is obtained by sampling the two input channels at different rates. Results of this study indicate the superiority of the symmetrical sampling system with regard to sampling fluctuations. A comparison is made between the UBC correlator and a conventional correlator on the basis of sampling fluctuations, with results dependent on the type of signal input and the method of comparison. It is shown that under a certain condition the UBC correlator is superior to the conventional device and in all cases the availability of the symmetrical sampling mode wholly or partially offsets the disadvantage of a high sampling-noise level which is inherently present because of the quantization process.Applied Science, Faculty ofElectrical and Computer Engineering, Department ofGraduat