Simplified adjusting and simulation of a pseudo Gaussian function in voltage domain generated with FGMOS transistors on circuit simulation software exportable to a multi-domain platform

摘要

Simulation software is a fundamental tool that should be used when designing systems dealing with circuit interaction, for instance. Nevertheless, there exist some inconveniences that limit and make specific software hard to use in this task. On one side, it may require a large processing hardware capacity, and on the other, simulation time increases substantially if the model complexity and accuracy required is high [1]. This work presents the procedure for adjusting and simulating a programmable pseudo Gaussian function generator implemented with eight FGMOS transistors in an analog and mixed-signal circuit simulator (Pspice®) with the interaction of a multi-domain design and simulation platform based on dynamic systems (Simulink® of Matlab®). Using the results, the best value of charge in each one of the eight transistor's floating gates to generate a desired form of the Gaussian function can be known: This function can maintain its form thanks to the FGMOS features, using a Fowler-Nordheim charge injection. [2, 3] The main goal of this study is to show how an automated adjust using FGMOS transistors can be made in order to reach a desired and repeatable output. This is very useful because it can be used to program an artificial neuronal network.