A scaleable, statistical SPICE Gummel-Poon model for SiGe HBTs

Abstract

A scaleable, statistical model has been developed for silicon germanium heterojunction transistors (SiGe HBTs), which are components of a commercially available BiCMOS technology for high-frequency applications. The SPICE Gummel-Poon (SGP) model parameters are scaled, and statistics added, using language features built into HSPICE, DC and AC fit is good over a wide range in emitter sizes, allowing an open-ended set of devices to be used with valid modeling capabilities. Features of IBM's HBT technology that contribute to the scaleability of the technology are discussed.