Magnetic recording and reproduction of pulses

The most widely used techniques for recording digital information on a moving magnetic medium are return-to-zero (RZ) and non-return-to-zero (NRZ). Both techniques have some peculiar advantages and disadvantages. Although sophisticated coding methods may be utilized to increase information density, the density achievable by any method is determined by the basic resolution of the record and reproduce processes. An expression is derived for the output of a reproduce head when an ideally recorded pulse is passed over it. The output is a function of the head fringing field in the region occupied by the recorded medium, and is the sum of the outputs produced by the longitudinal and perpendicular components of magnetization. A novel technique is used to measure the relative magnitudes of the components in a typical saturated recording tape. The perpendicular component is 11 per cent of the longitudinal component and may be neglected for the practical case. From a combination of experimental and theoretical data the width and height of the reproduced pulse are computed for variable gap width, medium thickness, and head-to-medium spacing. The effect of a nonideal pulse with a finite recorded width is considered. The total output pulse width is shown to be the sum of the computed ideal reproduce pulse width and the width of the actual recorded pulse. From the curves presented, one may observe that only a slight increase in resolution can be achieved by utilizing very small reproduce head gaps. Data are presented on the measured initial magnetization characteristics of a typical oxide. The characteristic may be approximated by either an offset linear curve up to about 60 per cent of saturation or a fourth power curve up to about 30 per cent of saturation. Data are also presented to show the effect of previous magnetization upon the transfer characteristic. The record process is analyzed with a step-by-step technique utilizing measured data on the head field and oxide magnetization characteristics. It is shown that both the shape and location of the recorded pulse are functions of the medium magnetization characteristic, the record head gap width, the record current, the medium thickness, and the head-to-medium spacing. The effect of each of these variables is computed. The computed results are verified experimentally. It is shown that under a wide range of conditions the reproduce pulse width obtained from a given head is approximately five times the width of a pulse ideally recorded by the same head. It is further shown that when the spacing between head and medium is larger than the gap width, the resulting over-all reproduced pulse width is approximately seven times the head-to-medium spacing. Previous recording history of the medium has a significant effect upon the pulse location. Data presented indicate that the record current must be approximately twice the current required for medium saturation to make the pulse location error unmeasurable.