An efficient neural network approach for nanoscale FinFET modelling and circuit simulation

Abstract

SUMMARYThe present paper demonstrates the suitability of artificial neural network (ANN) for modelling of a FinFET in nano-circuit simulation. The FinFET used in this work is designed using careful engineering of source-drain extension, which simultaneously improves maximum frequency of oscillation f max because of lower gate to drain capacitance, and intrinsic gain A V0 ¼ g m =g ds , due to lower output conductance g ds . The framework for the ANN-based FinFET model is a common source equivalent circuit, where the dependence of intrinsic capacitances, resistances and dc drain current I d on drain-source V ds and gate-source V gs is derived by a simple two-layered neural network architecture. All extrinsic components of the FinFET model are treated as bias independent. The model was implemented in a circuit simulator and verified by its ability to generate accurate response to excitations not used during training. The model was used to design a low-noise amplifier. At low power (J ds $10 mA/mm) improvement was observed in both third-order-intercept IIP 3 ($10 dBm) and intrinsic gain A V0 ($20 dB), compared to a comparable bulk MOSFET with similar effective channel length. This is attributed to higher ratio of first-order to thirdorder derivative of I d with respect to gate voltage and lower g ds in FinFET compared to bulk MOSFET.

Modelling and control of six‐phase induction motor servo‐driven continuously variable transmission system using blend modified recurrent Gegenbauer orthogonal polynomial neural network control system and amended artificial bee colony optimization

“…For achieving _ R 1 ≤0, the adaptation law _ Ω, the recompensed controller u c , and estimation law _ b σ can be designed (18), (27), and (28) are substituted into Eqn (26), and if Eqn (18) with I a = 0 is used, then Eqn (26) can be represented as…”

“…The use of amended artificial bee colony optimization yielded two optimal learning rates for the parameters, which helped improve convergence. Finally, comparison of the experimental results of the present study with those of previous studies demonstrated the high control performance of the proposed control scheme.Some of the principal advantages by using artificial neural networks are good learning ability and good performance for the tasks of system identifications and controls [15][16][17][18]. However, one of the major drawbacks in the artificial neural networks is to need large number of iterations and computationally intensive times for its training.…”

Cost‐effective GRNN‐based modeling of microwave transistors with a reduced number of measurements

“…Thus, criteria for the multi-output are obtained by summing criteria calculated for each component output using the formulae (4)- (6). Furthermore, RME is also used as given by Equation 3.…”

Nonlinear modeling of LDMOS transistors for high‐power FM transmitters

“…The main intrinsic element is the voltage-dependent current generator, I dsr , which models both the DC and dynamic I/V characteristics of the device; its description has to take into account the thermal phenomena in order to accurately predict the highfrequency device behavior. The remaining elements describe the strictly dynamic nonlinear effects, which are referred to as the capacitive core of the model, for which a non-quasi-static implementation has been considered [10,11].…”