Toward a new paradigm for final test of magnetic media for hard disc drives

• Published in: IEEE transactions on magnetics 1997-01, Vol.33 (1), p.962-967
• Main Author: Pressesky, J.
• Format: Article
• Language: English
• Subjects: Drives; Magnetic films; Magnetic heads; Magnetic levitation; Magnetic properties; Monitoring; Optical films; Pollution measurement; Testing; Transducers
• ISSN: 0018-9464; 1941-0069
• DOI: 10.1109/20.560139

This paper proposes a new approach to the final test of rigid disc magnetic media. The conventional approach to the final test of magnetic media is to construct an analog of the disc drive consisting of a spin stand and calibrated test heads mounted on air bearing sliders. The properties of the test head (both initial and time evolving) must be accounted for In interpreting the test results. The new approach outlined in this paper replaces the calibrated test heads with a combination of transducers that are maintained at a fixed and relatively large distance from the disc surface. Scanning optical interferometry is proposed as a method of monitoring film integrity. Film integrity is defined to be the absence of asperities with heights greater than a specified maximum and a distribution of flaws (film voids, delaminations, depressions) that meet the requirements of the disc drive manufacturer. It is proposed that the film magnetic properties be monitored directly through the characteristics of the remanent hysteresis loop, rather than inferred indirectly from the recording performance properties of the head-medium ensemble. The guiding principles behind this approach are twofold: 1. The test transducers should never contact the surface of the disc in any way. This requirement preserves the transducer characteristics over long periods of time and reduces or eliminates the need for transducer replacement or re-calibration and minimizes the opportunities for the transfer of contaminants to the disc surface. 2. The measured properties of the medium should be fundamental, and consequently universal. This will simplify the test process and eliminate unnecessary confusion. Thus one measures coercivity and coercive squareness, not overwrite and one measures the flaw size in /spl mu/m, not as the number of bits in error.