Jerod F. Mason

Temperature dependent linear HEMT model extracted with multi-temperature optimization

Temperature dependent model is conventionally developed with 1) individual linear model extraction at each temperature and 2) temperature dependence extraction for each model parameter. A novel approach based on a multi-temperature optimization is proposed in this study. The new approach is faster and more accurate since the linear models at different temperatures are extracted simultaneously and inter-relatedly. It has been found, based on the model extracted, that there is a specific gate-source voltage of Vgs0, where the transconductance (Gm) keeps constant versus temperature. Gm increases (decreases) with temperature for Vgs<;Vgs0 (Vgs>;Vgs0). With increasing temperature, a substantial decrease in extrinsic inductance is also observed. Our findings are believed to be useful for designing HEMT amplifiers with less temperature dependence.

Analysis and modeling on linearity for multi-throw TX/RX switches

Analytical analysis on linearity of multi-throw switches is given that serves as design guide. A switch multi-gate switch model is developed that includes all important mechanism associated with linearity, such as CV below pinchoff, leakage between gate to drain/source and drain to source. A highly linear nine-throw switch for GSM/WCDMA applications is used to validate the model that shows an IMD3 of −105 dBm at the control voltages starting as low as 2.8/0 V. This high linearity performance was achieved using good control on CV flatness and leakages in processing.

Optimum Matching Technique for the Switch Design

The theoretical background and hardware implementation of optimal matching filters designed to minimize harmonic generation by RF switches and provide harmonic filtering of the PA generated harmonics is presented. The switch with the optimally matched filters shows 5-7 dB improvements in harmonics generated under high VSWR, leading to better than -39dBm harmonics at 35/33 dBm for all phases of VSWR=5:1 for low/high band GSM applications. This technique leads to a fully integrated solution offering both improved harmonics performance and reduced size.

Multi-device Optimization for Scalable DC HEMT Model with Self-Heating Effect

This paper presents a new approach for extracting a scalable DC HEMT model. Scaling rules with unknown coefficients are assumed for each size dependent parameter, and the model parameter of different devices is thus correlated. A scalable model can then be extracted by multi-device optimization. The optimization is carried out on five devices with different gate width. In this way, accurate scaling rules for each parameter and very good I-V fittings for each device have been achieved simultaneously.