Condition-based maintenance of mechanical setup in aluminum wire bonding equipment by data mining

On the way to full production control, wire bonding equipment requires data-driven condition-based maintenance of the mechanical setup. In this paper, we identify aspects of regular mechanical equipment setups that severely affect bonding quality and equipment health in mass production. We show that mechanical equipment setups lower process stability in aluminum wire bonding. Typical faults in mechanical setup directly influence device quality, which is measured through the use of pull and shear tests. In this paper we demonstrate that data mining can detect mechanical setup faults by extracted features of monitored machine data. Condition-based maintenance enables the verification of mechanical setup conditions after manual adjustments and individual bonding events. This novel condition-based maintenance has the potential to partially substitute pull and shear tests, and hence avoid related equipment downtime as well as additional cost.