Physical modeling of GaAs MESFETs in an integrated CAD environment: from device technology to microwave circuit performance

Abstract

-A CAD environment leading from technology to performance evaluation by integrating process, device, and circuit simulation would be a valuable tool for the development of monolithic microwave circuits. The paper focuses on the linkage between a physical device simulator for smalland large-signal characterization, and CAD tools for both linear and nonlinear circuit analysis and design. Efficient techniques are presented for the physical dc and small-signal analysis of MESFETs; then, the problem of physical simulation in a circuit environment is discussed, and it is shown how such a simulation makes it possible to obtain small-signal models accounting for propagation and external parasitics. Finally, efficient solutions are proposed for physical large-signal simulation, based on deriving large-signal equivalent circuits from small-signal analyses under different bias conditions. The small-and large-signal characterizations thereby obtained allow physical simulation to be performed efficiently in a circuit environment. Examples and results are presented throughout the paper.

Physics-based electron device modelling and computer-aided MMIC design

“…In accordance with other case studies [44], good agreement is found with measured dc data, although the substrate residual donor and acceptor concentrations had to be reverse-modelled so as to match the threshold voltage exactly. The ac model had to include low-frequency dispersion effects due to substrate deep levels, without which the output conductance would have been underestimated by a factor 2.…”

Physics-based large-signal sensitivity analysis of microwave circuits using technological parametric sensitivity from multidimensional semiconductor device models

“…The shortterm memory requirement is commonly verified for "intrinsic" devices, i.e., devices not affected by dominating parasitic effects. Such a case occurs in physics-based models, wherein parasitics can be exactly deembedded [14]. Actually, a number of validation tests have shown good agreement between 2-D simulated LS responses and the corresponding values predicted according to (8).…”

“…In particular, according to the approach in [2], [14], the LS circuit analysis is carried out through circuit-oriented nonlinear models and their characteristic parameters are extracted from dc and small-signal ac physics-based device simulations. In this way, LS circuit analysis is performed with good computational efficiency, while the link between physical parameters and the corresponding circuit performance is still provided by an off-line, physics-based device sensitivity analysis.…”

“…Among these models, the nonlinear integral model (NIM) proposed in [13] is particularly well suited, since it only requires dc and small-signal data to be evaluated through the physical model; the accuracy of the NIM when compared to multidimensional simulation is excellent, as shown in [13]. Nevertheless, more conventional chargecontrol LS equivalent circuit models (which can be interpreted as limiting cases of the NIM) can also provide a very good link between physics-based and LS simulation, as discussed in [14].…”

An improved CAD procedure for oscillator circuit design using fast-approximation formulas describing the voltage-dependent non-linear characteristics of the intrinsic equivalent circuit elements of the active devices

“…The main reason for this is the severe voltage dependence [1,2] of the intrinsic transistor equivalent-circuit elements, which represent the physical properties of the semiconductor substrate. Therefore, in actual practice, a constant voltage regulator is used typically when assembling such an oscillator.…”