Design framework for inverter cascode transimpedance amplifier using Gm/ID based PSO applying design equations

Abstract

This paper presents a framework for the design of the inverter cascode (InvCas) transimpedance (TIA) for the optical receiver's front-end. The framework combines the particle swarm optimization (PSO) evolutionary algorithm with the gm/ID methodology to calculate the required circuit parameters needed to obtain the target specifications entered by the designer. The framework relies on the derived circuit equations to define the design space and evaluate the performance using the parameters calculated by the gm/ID and PSO. This technique speeds up the design process by avoiding the need for Spice simulation to assess the fitness of each particle during the optimization process. The proposed framework is implemented in a Matlab program utilizing the framework that takes target specifications as an input and provides the design parameters of the particle with the best circuit performance satisfying these specifications as an output. The framework is tested using the 130 nm CMOS technology to produce three different designs each targeting a different bandwidth and gain for different optical communications data rates while optimizing the power and noise performance of each design. All the evaluated designs achieved the required specifications and show superior performance compared with TIA state-of-the-art.