Since the introduction of the IBM 305 RAMAC system in 1956, performance and storage capacities of hard disk drives have improved tremendously. To reliably read and write data, the slider must follow the data stored on the magnetic disk closely enough while maintaining a near constant spacing. Currently, the spacing between the recording and the magnetic disk – the flying height – is on the order of 1-2 nm during reading and writing. At such low spacings, intermittent contacts are inevitable, giving rise to wear and degradation of the head-disk interface. Flying heights of 1-2 nm are achieved using thermal flying height control (TFC) technology. TFC recording heads, or TFC sliders, feature thin-film resistive heater elements near the read and write element. Actuating the heater element heats up the nearby material. The material expands due to the heat which causes the slider to (thermally) protrude towards the disk at the location of the read and write element. An increase in heater power increases this protrusion, thus locally reducing the slider flying height. In this dissertation, we focus on experimental investigations of the interface between a TFC slider and a magnetic disk from both a tribological and controls point of view.
First, contact and temperature rise between thermal flying height control (TFC) sliders and magnetic disks are studied. Head-disk contact is established by gradually increasing the power input to the resistive heater element of a TFC slider. Laser Doppler vibrometry is employed for studying the dynamics of the vertical gimbal velocity. The gimbal is part of the suspension which the slider is attached to. The temperature rise upon head-disk contact is estimated from the resistance change at the read element via auxiliary calibration measurements.
Next, wear of TFC sliders is studied. Head wear was determined by measuring the change in the heater touch-down power before and after wear testing. The touch- down power denotes the power input to the heater of a TFC slider at which the onset of slider-disk contact occurs. After wear testing, selected heads were examined using scan- ning electron microscopy to identify regions of wear on the write shields. Furthermore, atomic force microscopy images of worn and unworn recording heads were acquired to determine changes in surface roughness. The effect of bonded fraction of the lubricant, relative humidity, and temperature on head wear is investigated.
In addition, we study head wear as a function of relative humidity and DC bias voltage applied across the head-disk interface. Wear tests were performed at <8 %, 30 % and 52 % relative humidity. Similarly, head wear was determined by measuring the change in the heater touch-down power before and after 10 minute wear tests. After wear testing, selected recording heads were examined using atomic force microscopy to identify regions of wear or deposit formation on the slider surface.
A new approach for predicting the touch-down power of a TFC slider is pre- sented. The method utilizes the thermal contact sensor to sense head-disk proximity prior to actual head-disk contact. Impeding contact is predicted based on a change in dynamics from heater input to sensor output with decreasing flying height. The dynamics between the heater and the contact sensor are identified from experimental step response measurements using the step-based realization algorithm. The effect of step height is investigated and it is shown that the coefficients of the transfer function can be used to predict the onset of head-disk contact without having to perform a complete touch-down measurement.
Finally, an algorithm is proposed for minimizing low-frequency variations in flying height of thermal flying height control (TFC) sliders. The method utilize the re- sistive heater element of a TFC sliders for spacing adjustment and the embedded thermal contact sensor for estimating changes in flying height. Data based modeling is carried out to identify the dynamics of the thermal actuator. The optimal feedforward heater input profile is calculate via convex optimization techniques. The feedforward approach was verified experimentally on a spin-stand tester.