On the control of micro Ball Grid Array (μBGA) production systems

摘要

In this paper, two novel control implementations for the production of Ball Grid Arrays (BGA) at the micro scale are presented. The first is a linear control algorithm; the second is an adaptive neural network. The linear control algorithm was developed to manage the key control parameters of the system and to reach the desired performance as fast as possible based on measurements of system's output. The neural network uses an adaptive control architecture for controlling the performance of the system. This architecture dynamically assigns weights to output variables, in current point time as well as in previous points in time, and uses them as control inputs to improve both production performance and stability. Brief theoretical background, experimental validation and a comparison of the two algorithms in terms of performance and stability are provided.