Analytical Inverter Delay Modeling Using Matlab's Curve Fitting Toolbox

TL;DR

This paper introduces an analytical inverter delay model using Matlab's Curve Fitting Toolbox, achieving high accuracy and efficiency for deep submicron CMOS technology by modeling complex process and load dependencies.

Contribution

The paper develops a new analytical delay model fitted with Matlab's Curve Fitting Toolbox, providing a fast and accurate alternative to SPICE simulations for various process parameters.

Findings

Average error less than 2% compared to SPICE simulations

Effective across multiple process nodes and load conditions

Suitable for statistical gate delay characterization

Abstract

This paper presents a new analytical propagation delay model for deep submicron CMOS inverters. The model is inspired by the key observation that the inverter delay is a complicated function of several process parameters as well as load capacitance. These relationships are considered by fitting functions for each parameter derived from the Curve Fitting Toolbox in Matlab. Compared to SPICE simulations based on the BSIM4 transistor model, the analytical delay model shows very good accuracy with an average error less than 2% over a wide range of process parameters and output loads. Hence, the proposed model can be efficiently used for different technology nodes as well as statistical gate delay characterisation.