An Adaptive Compensated Proportional – Derivative Controller Using Artificial Neural Networks for Hard Disk Drive Lapping Machines

Abstract

In this paper, an adaptive compensated proportional – derivative controller using the artificial neuralnetworks was presented for hard disk drive head lapping machines where standard deviation reduction ofthe final stripe height was determined to be a key performance. Hebbian learning technique was applied toadjust the gains. More than 10,000 heads from two different products that final stripe heights are different,were lapped using learning rate = 0.20, Kpc=13 and Kdc=16. As a result, 89.41 percent of the heads withstandard deviation less than 6 μΩ was obtained.

Keywords : Adaptive Compensated Proportional – Derivative Control (ACPD) / Hard DiskDrive Head / Deviation of Stripe Height / Hebbian learning