Neural network based prediction of stable equivalent series resistance in voltage regulator characterization

Mohd Hairi Mohd Zaman, M. Marzuki Mustafa, M. A. Hannan, Aini Hussain

Research output: Contribution to journalArticle

Abstract

© 2018 Institute of Advanced Engineering and Science. All rights reserved. High demand on voltage regulator (VR) currently requires VR manufacturers to improve their time-to-market, particularly for new product development. To fulfill the output stability requirement, VR manufacturers characterize the VR in terms of the equivalent series resistance (ESR) of the output capacitor because the ESR variation affects the VR output stability. The VR characterization outcome suggests a stable range of ESR, which is indicated in the ESR tunnel graph in the VR datasheet. However, current practice in industry manually characterizes VR, thereby increasing the manufacturing time and cost. Therefore, an efficient method based on multilayer neural network has been developed to obtain the ESR tunnel graph. The results show that this method able to reduce the VR characterization time by approximately 53% and achieved critical ESR prediction error less than 5%. This work demonstrated an efficient and effective approach for VR characterization in terms of ESR.
Original languageEnglish
Pages (from-to)134-142
Number of pages119
JournalBulletin of Electrical Engineering and Informatics
DOIs
Publication statusPublished - 01 Mar 2018

Fingerprint

voltage regulators
Voltage regulators
Regulator
Voltage
Neural Networks
Neural networks
Series
Prediction
predictions
Tunnel
tunnels
output
Output
Tunnels
Resistance
New Product Development
Multilayer Neural Network
product development
Multilayer neural networks
Prediction Error

All Science Journal Classification (ASJC) codes

  • Computer Science (miscellaneous)
  • Control and Systems Engineering
  • Information Systems
  • Instrumentation
  • Hardware and Architecture
  • Computer Networks and Communications
  • Control and Optimization
  • Electrical and Electronic Engineering

Cite this

@article{933e4befc64a42009a9da7dd5b41d2bf,
title = "Neural network based prediction of stable equivalent series resistance in voltage regulator characterization",
abstract = "{\circledC} 2018 Institute of Advanced Engineering and Science. All rights reserved. High demand on voltage regulator (VR) currently requires VR manufacturers to improve their time-to-market, particularly for new product development. To fulfill the output stability requirement, VR manufacturers characterize the VR in terms of the equivalent series resistance (ESR) of the output capacitor because the ESR variation affects the VR output stability. The VR characterization outcome suggests a stable range of ESR, which is indicated in the ESR tunnel graph in the VR datasheet. However, current practice in industry manually characterizes VR, thereby increasing the manufacturing time and cost. Therefore, an efficient method based on multilayer neural network has been developed to obtain the ESR tunnel graph. The results show that this method able to reduce the VR characterization time by approximately 53{\%} and achieved critical ESR prediction error less than 5{\%}. This work demonstrated an efficient and effective approach for VR characterization in terms of ESR.",
author = "Zaman, {Mohd Hairi Mohd} and Mustafa, {M. Marzuki} and Hannan, {M. A.} and Aini Hussain",
year = "2018",
month = "3",
day = "1",
doi = "10.11591/eei.v7i1.857",
language = "English",
pages = "134--142",
journal = "Bulletin of Electrical Engineering and Informatics",
issn = "2089-3191",
publisher = "Institute of Advanced Engineering and Science (IAES)",

}

Neural network based prediction of stable equivalent series resistance in voltage regulator characterization. / Zaman, Mohd Hairi Mohd; Mustafa, M. Marzuki; Hannan, M. A.; Hussain, Aini.

In: Bulletin of Electrical Engineering and Informatics, 01.03.2018, p. 134-142.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Neural network based prediction of stable equivalent series resistance in voltage regulator characterization

AU - Zaman, Mohd Hairi Mohd

AU - Mustafa, M. Marzuki

AU - Hannan, M. A.

AU - Hussain, Aini

PY - 2018/3/1

Y1 - 2018/3/1

N2 - © 2018 Institute of Advanced Engineering and Science. All rights reserved. High demand on voltage regulator (VR) currently requires VR manufacturers to improve their time-to-market, particularly for new product development. To fulfill the output stability requirement, VR manufacturers characterize the VR in terms of the equivalent series resistance (ESR) of the output capacitor because the ESR variation affects the VR output stability. The VR characterization outcome suggests a stable range of ESR, which is indicated in the ESR tunnel graph in the VR datasheet. However, current practice in industry manually characterizes VR, thereby increasing the manufacturing time and cost. Therefore, an efficient method based on multilayer neural network has been developed to obtain the ESR tunnel graph. The results show that this method able to reduce the VR characterization time by approximately 53% and achieved critical ESR prediction error less than 5%. This work demonstrated an efficient and effective approach for VR characterization in terms of ESR.

AB - © 2018 Institute of Advanced Engineering and Science. All rights reserved. High demand on voltage regulator (VR) currently requires VR manufacturers to improve their time-to-market, particularly for new product development. To fulfill the output stability requirement, VR manufacturers characterize the VR in terms of the equivalent series resistance (ESR) of the output capacitor because the ESR variation affects the VR output stability. The VR characterization outcome suggests a stable range of ESR, which is indicated in the ESR tunnel graph in the VR datasheet. However, current practice in industry manually characterizes VR, thereby increasing the manufacturing time and cost. Therefore, an efficient method based on multilayer neural network has been developed to obtain the ESR tunnel graph. The results show that this method able to reduce the VR characterization time by approximately 53% and achieved critical ESR prediction error less than 5%. This work demonstrated an efficient and effective approach for VR characterization in terms of ESR.

UR - https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85045645286&origin=inward

UR - https://www.scopus.com/inward/citedby.uri?partnerID=HzOxMe3b&scp=85045645286&origin=inward

U2 - 10.11591/eei.v7i1.857

DO - 10.11591/eei.v7i1.857

M3 - Article

SP - 134

EP - 142

JO - Bulletin of Electrical Engineering and Informatics

JF - Bulletin of Electrical Engineering and Informatics

SN - 2089-3191

ER -