Masters Theses

Keywords and Phrases

Decoupling Capacitor; PDN

Abstract

"Decoupling capacitors (decaps) are used in power distribution network (PDN) design to act as a low impedance return path for noise and act as a local source of charge when required by integrated circuits (IC). For placement of decaps, which can number even to the hundreds, many algorithms have been proposed including genetic algorithms (GA), iterative algorithm, and machine learning methods.

One limitation of iterative decap placement algorithms (adding one decap at a time) is how the construction of the algorithm affects the form of the final solution. For example, if the reward function (used in GA and machine learning methods) for evaluating a solution was based on maximizing the number of points below the target impedance, then larger package size decaps may be placed first, as low impedance points are easier to bring below the target using fewer number of decaps. These decaps may also be placed in locations near the IC for the benefit of low loop inductance. The expectation, however, should be for small package decaps to be placed near the IC as they contribute less inductance due to their smaller geometry. In this work, we propose two GAs for minimizing the required number of decaps in PDN design, with the goal of minimizing the number of assumptions about the structure of the solution within the genetic operators.

For the second topic, for capacitors mounted to a board, there is a mutual coupling between the capacitor and the return plane that is stack-up dependent and so cannot be captured in one measurement/model. We propose a physics-based curve fitting scheme to interpolate inductances over many stack-ups in just two measurements"--Abstract, p. iv

Advisor(s)

Hwang, Chulsoon

Committee Member(s)

Beetner, Daryl G.
Kim, DongHyun (Bill)

Department(s)

Electrical and Computer Engineering

Degree Name

M.S. in Electrical and Computer Engineering

Publisher

Missouri University of Science and Technology

Publication Date

Spring 2023

Pagination

xii, 84 pages

Note about bibliography

Includes_bibliographical_references_(page 82)

Rights

© 2023 Jack Juang, All Rights Reserved

Document Type

Thesis - Open Access

File Type

text

Language

English

Thesis Number

T 12250

Electronic OCLC #

1423538018

Download

Share

 
COinS