Abstract:
This research presents the approach to improve a fine lapping process of hard disk drive
(HDD) lapping machines by removing materials from each slider together with
controlling the strip height (SH) variation to minimum value. There are two algorithms
in this research. First, classification of initial SH pattern using Radial Basis Function
Neural Network for initial Kp prediction is developed. A design of experiment (DOE)
with factorial analysis by two-way analysis of variance (ANOVA) is adopted to obtain a
statistically information of input parameters. The result shows that the model needs a lot
of data for training to achieve a good classification significantly, moreover considering
another effect such as a machine, a product, or a row tool. Hence this algorithm is hard
to be implemented for a HDD lapping machine. The second algorithm is the force
compensation with PD controller. A finger force is adjusted for tracking improvement
by a compensation force. Tuning the proportion and derivative gain of PD controller is
experimented to yield the optimum value. The result shows the desirable value of the
final SH variations are obtained.