A NEW APPROACH TO INVESTIGATE LADDER-TYPE SAW FILTERS INCORPORATING PACKAGING EFFECTS

Ladder-type surface acoustic wave (SAW) filter is a commonly used radio frequency (RF) filter in many communication systems. The circuit structure of a ladder-type filter consists of several one-port SAW resonators on a piezoelectric substrate arranged in series and parallel alternatively. The shape and dimension of bonding pads, the lengths of bonding wire, and the internal connection paths between one-port SAW resonators may present a significant influence on the performance of the filter. Hence, how to incorporate these effects into the early stage of the filter design is an important issue. In this paper, neural network was employed to extract the lumped element parameters of bonding pads and internal connections. As an example, an RF ladder-type SAW filter used in globe positioning system (GPS) was examined. The result showed a good agreement with measurement data as well as the data obtained by applying full electromagnetic wave simulator.