Monitoring of the Ultrasonic Wire Bonding Quality Based on System Impedance

A new method is putted forward to monitor online wire bonding quality through monitoring the change of system input electrical impedance. The signals of voltage and current of ultrasonic piezoelectric driver are transformed analytic signals by Hilbert transform. Using these analytic signals, the input electrical impedance is acquired. There are significant changes in the system input electrical impedance. Good correlation is found between the change in system input electrical impedance and the bonding intensity. A neural network module is constructed to classify and identify bonding quality. The results evaluated are uniform basically with the destructed shear test experiment. Bonding process drift towards unacceptable bond quality can be identified. The work will be used to develop a real-time monitoring system for ultrasonic wire bonding.