Neural Network Model for On-Chip Spiral Inductors in Packaging Processes

In this study, a novel artificial neural network (ANN) model that covers a wide operating frequency range and comprehensive design parameters is proposed. Compared to the original inductor, the inductor incorporating package layers shows variations in performance, mainly in the advance of self-resonant frequency (SRF), accompanied by a reduction in both the peak quality (Q) factor and the inductance (L) at the self-resonant frequency. Notably, excellent speed and accuracy are retained by ANN model even after incorporating the design parameter of package layer thickness.