The generic technique of applying pseudorandom, discrete-level, periodic reactivity perturbation signals to measure the reactivity-to-power frequency response function was extended to the liquid-metal reactor, Experimental Breeder Reactor-II (EBR-II). This technique was developed in the late 1960s and applied in several reactor designs with extensive testing performed at the Molten Salt Reactor Experiment. Signals employed at EBR-II included the pseudorandom binary sequence, quadratic residue binary sequence, pseudorandom ternary sequence, and multifrequency binary sequence. For all the signals employed, the resultant reactor power perturbation was small enough to be acceptable for normal at-power operation and in-place irradiation experiments. The frequency response results are compared with the zero-power frequency response function, yielding a quantitative measure of the EBR-II reactivity feedback effects. The frequency response function results are in good agreement with rod-oscillator data and model predictions. The multifrequency binary sequence concentrated 64% of the total signal power into the four feedback frequencies associated with the predominant feedback time constants. The input signal quality, characterized by the autocorrelation function and power spectra, validated the automatic control rod drive system design and operation as an effective tool for frequency response determination over the range of frequencies where important system dynamic effects occur.