Characterization of head stability induced signal fluctuation

An experimental procedure to measure head-instability-induced noise has been developed. The procedure uses a separate head to provide a noise-independent timing reference. The ordinary disk and head noises are also removed. The nonstationary autocorrelation function, which provides a look at noise in relation to the readback signal, is used. The autocorrelation function is calculated, and the noise power and the correlation at signal peaks are extracted for further study. It is found that the noise magnitude is largest near the signal peak, and the noise correlation is not -1 at positive and negative peaks