Time domain analysis of low frequency noise in giant magneto-resistive recording heads

The low frequency noise observed in giant magneto-resistive recording heads was studied in the time domain. Time series data was collected on bottom spin-valve sensors while varying the bias current (I _b) about a nominal operating point. An optimal time delay of 5 ns was chosen based on a 80% reduction in mutual information. Two dimensional embedding of the data revealed a progression from one stable state for low I_b to two metastable states and finally a second stable state at high I_b. The metastable states manifest themselves as random telegraphic noise in the time series. The average time between noise spikes was obtained by calculating the time between consecutive points on a Poincare section and was estimated as 0.37 μs. The Renyi entropy was used to quantify the stability of the system both with and without an isolated pulse pattern