Modal Characterization of Skynet Response to Electrical and Photon Stimulation

Modal analysis places some requirements that should be met during the experiment: (1) Close control of the timing, particularly the relative timing between different sensor records for each excitation shot. Serious problems were encountered in the analysis of the EWR data because insufficient care was taken in documenting changes in cabling. (2) Reliability and stability in data link gain. The mode comparison performed in this report is limited in accuracy by drifts in the gain and linearities of the fiber-optics data links. (3) Cleanliness of the electrical pulses. Modal analysis of electrical test data revealed a disturbance ~120 ns after the beginning of the excitation (i.e. mode fitting was unsuccessful if it attempted to span the time from the beginning past 120 ns). Subsequent measurements proved that the slope of the pulser voltage decay curve changed at that time, presumably because the pulser switch reopened. (4) Minimum high frequency noise. The best fits are achieved with minimum noise. The smoothing procedures used make a major improvement in the fit, but they are most effective if the noise is small to begin with. Averaging over repetitive pulses for electrical tests is particularly valuable. Unfortunately, photon tests do not allow such averaging. Our experience with modal decomposition by the revised Prony method indicates: (1) The eigenvalue/eigenvector approach to solving the Prony equation matrix removes some of the bias seen in the conventional approach, and provides multiple fits from which the best can be chosen.