energy harvesting for implantable medical devices€¦ · energy conditioning for implantable...

Energy Conditioning for Implantable Medical Devices

A Multiple Input Redundant System

Ana Marta Carpinteiro de Barros Borges

Thesis to obtain the Master of Science Degree in

Electronics Engineering

Supervisor: Prof. Jose Antonio Beltran Gerald

Examination Committee

Chairperson: Prof. Pedro Miguel Pinto RamosSupervisor: Prof. Jose Antonio Beltran Gerald

Member of the Committee: Prof.a Maria Beatriz Mendes Batalha Vieira Vieira Borges

October 2018

Abstract

Powering up an Implantable Medical Device (IMD) by taking advantage of energy harvesting devices,

which convert energy collected from human body activities into electrical energy, has been increasingly

an alternative to fixed density and lifetime batteries, that may represent several drawbacks for patients.

Piezoelectric and electrostatic generators gather the best results in terms of output power generated and

reliability. In order to combine these energy harvesters’ characteristics, there are several approaches

that have multiple energy harvesting devices as input sources. Despite of IMDs are usually low power

devices, the energy generated by energy harvesters is not enough to power them. Therefore, it is

needed to boost the generated voltage, using voltage elevation circuits for this purpose.

In this work, a system capable of processing harvest energies to power up an implantable medical

device that, being very simple, automatically guarantees the existence of a working input power source,

was developed. It was proved that this system, besides providing some voltage elevation, is capable

of readjusting the input source, maintaining the minimum output voltage required. Also, a substantial

revision of the literature the energy harvesting state of the art was performed, in order to gather in just

one document, the relevant latest information spread by a varied literature (medical, physics, electrical

and technological publications).

Keywords

Energy Harvesting, Implantable Medical Devices, Boost DC-DC Converter, Multiple Input Sources Sys-

tem.

i

Resumo

Aproveitar a energia gerada pelo corpo humano esta a surgir progressivamente como uma alterna-

tiva ao recurso de baterias para alimentar dispositivos medicos implantaveis, cuja utilizacao pode provo-

car algumas complicacoes para os pacientes. A conversao da energia proveniente do corpo humano, e

da sua atividade fısica, em energia electrica utilizavel e realizada atraves de dispositivos implantados no

corpo humano para este efeito, sendo os geradores piezoeletricos ou electroestaticos alguns exemplos.

De forma a tirar partido das vantagens dos varios geradores disponıveis, propoe-se neste trabalho uma

solucao que combina varias entradas, sendo capaz de recolher energia de varias fontes, criando-se

assim um sistema redundante que garante sempre uma fonte funcional quando outra falha. No entanto,

apesar dos dispositivos implantaveis consumirem baixa potencia, a energia gerada pelos coletores de

energia nao e suficiente para os alimentar. Desta forma, e necessaria a implementacao de circuitos ca-

pazes de elevar a tensao. O sistema proposto permite alguma elevacao de tensao, alem da garantia de

continuidade de fornecimento de energia ao implante. Como objetivo secundario, foi tambem realizada

uma revisao abrangente sobre o tema de recolha de energia, reunindo num so documento informacao

relevante e atual sobre o tema, suportada por uma literatura variada.

Palavras Chave

Recolha de Energia, Dispositivos Medicos Implantaveis, Conversores DC-DC Elevadores de Tensao,

Sistema com Multiplas Fontes de Entrada.

iii

Contents

1 Introduction 1

1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

1.2 Goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

1.3 Novelties of the Thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

1.4 Document Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2 State of the Art 7

2.1 Energy Harvesting for Powering Implantable Medical Devices . . . . . . . . . . . . . . . . 9

2.1.1 Independent Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

2.1.1.A Biofuel Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

2.1.1.B Thermoelectric Generators . . . . . . . . . . . . . . . . . . . . . . . . . . 10

2.1.1.C Electromagnetic Generators . . . . . . . . . . . . . . . . . . . . . . . . . 12

2.1.1.D Electrostatic Generators . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

2.1.1.E Piezoelectric Generators . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

2.1.1.F Triboelectric Nanogenerators . . . . . . . . . . . . . . . . . . . . . . . . . 19

2.1.1.G Photovoltaic Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

2.1.2 Discussion and Proposed Approach . . . . . . . . . . . . . . . . . . . . . . . . . . 21

2.2 DC-DC Converters for Voltage Elevation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

2.2.1 Single Source DC-DC Converters . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

2.2.2 Multiple Source DC-DC Converters . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

3 Proposed Solution 27

3.1 Working Principle Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

3.1.1 Single Converter Working Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

3.1.2 Multiple Converters Working Principle . . . . . . . . . . . . . . . . . . . . . . . . . 34

3.1.2.A Continuous-Continuous Conduction Mode . . . . . . . . . . . . . . . . . 36

3.1.2.B Discontinuous-Discontinuous Conduction Mode . . . . . . . . . . . . . . 42

3.1.2.C Continuous-Discontinuous Conduction Mode . . . . . . . . . . . . . . . . 48

3.1.2.D Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

v

3.2 System Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

3.2.1 Load Impedance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

3.2.2 Inductor Value . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

3.2.3 Output Capacitor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

3.2.4 Input Voltage Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

3.2.5 Schottky Diodes and MOSFET . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

3.2.6 Duty Cycle Generator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

3.2.7 Zener Diode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

3.2.8 Dimensioned Circuit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

4 Circuit Performance in Case of Failure 63

4.1 Two Converters Failure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

4.1.1 Load Impedance Variation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

4.2 Three Converters Failure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70

5 Conclusions 75

Bibliography 79

A Appendix - Input Voltage Sources Generators 85

B Appendix - Duty Cycle Generator 89

vi

List of Figures

1.1 Diversity of implantable medical device applications [3]. . . . . . . . . . . . . . . . . . . . 4

2.1 Biofuel cell conceptual view [4]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

2.2 Conceptual view of thermoelectricity [4]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

2.3 Implanted TEG for thermal energy harvesting and the modal of a tissue [8]. . . . . . . . . 12

2.4 Artificial accommodation system within an eye powered by a TEG [9]. . . . . . . . . . . . 12

2.5 Conceptual view of an EMG for powering IMD [4]. . . . . . . . . . . . . . . . . . . . . . . 13

2.6 EMG structure for knee prosthesis: a) permanent magnets location; and b) coils location.

[14,15]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

2.7 Implantable EMG in a hip prosthesis [17]. . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

2.8 Conceptual view of ESG for harvesting energy from human body motion [4]. . . . . . . . . 15

2.9 Prototype of a MEMS based electrostatic generator [23]. . . . . . . . . . . . . . . . . . . . 15

2.10 Conceptual view of piezoelectric generators for harvesting energy from human body mo-

tion [4]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

2.11 Knee implant with three piezoelectric stacks [25]. . . . . . . . . . . . . . . . . . . . . . . . 17

2.12 Conceptual block diagram of an implanted piezoelectric generator from muscles contrac-

tions [29]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

2.13 Principle of operation and applications of a p-HEI device [30]. . . . . . . . . . . . . . . . . 18

2.14 Implantable energy harvester that uses blood pressure variations [33]. . . . . . . . . . . . 19

2.15 Working mechanism of TENG, using contact separation [37]. . . . . . . . . . . . . . . . . 19

2.16 Working mechanism of TENG, using contact sliding between two materials [37]. . . . . . 19

2.17 Solar energy harvester [39]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

2.18 Block diagram of the hybrid inductive capacitive converter [45]. . . . . . . . . . . . . . . . 24

2.19 Multiple input converter topology [47]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

2.20 Double input PWM DC-DC converter [48]. . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

2.21 Structure of multiple input source converter [49]. . . . . . . . . . . . . . . . . . . . . . . . 25

2.22 Multiple input step-up converter [50]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

vii

3.1 Block diagram of the proposed circuit. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

3.2 Block diagram of the entire system. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

3.3 Proposed Topology. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

3.4 Boost converter schematics [51]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

3.5 Continuous operation mode [51]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

3.6 Determination of the output voltage ripple [51]. . . . . . . . . . . . . . . . . . . . . . . . . 32

3.7 Discontinuous operation mode [51]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

3.8 Multiple parallel DC-DC boost converters sharing a same load. . . . . . . . . . . . . . . . 35

3.9 Simulations results for Scenario 1, when VI1 > VI2 and both converters in Continuous

Conduction Mode (CCM). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

3.10 Simulation results for Scenario 2, when VI1 = VI2 and both converters in CCM. . . . . . . 37

3.11 Simulation results for Scenario 3, when VI1 < VI2 and both converters in CCM. . . . . . . 38

3.12 Steady state simulation results of the converters set with the sizing done in Table 3.5. . . 42

3.13 Simulations results for Scenario 1, when VI1 > VI2 with both converters in Discontinuous

Condustion Mode (DCM). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

3.14 Simulation results for Scenario 2, when VI1 = VI2 with both converters in DCM. . . . . . . 43

3.15 Simulation results for Scenario 3, when VI1 < VI2 with both converters in DCM. . . . . . . 44


3.17 Simulations results for Scenario 1, when VI1 > VI2 with one converter in CCM and the

other in DCM. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

3.18 Simulation results for Scenario 2, when VI1 = VI2 with one converter in CCM and the

other in DCM. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

3.19 Simulation results for Scenario 3, when VI1 < VI2 with one converter in CCM and the

other in DCM. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50


3.21 Zener diode I-V characteristics [60]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

3.22 Multiple converters topology sized circuit. . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

3.23 Sized circuit behaviour. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

3.24 Time diagrams of converter 1, 2 and converters’ set. . . . . . . . . . . . . . . . . . . . . . 60

4.1 Proposed circuit with piecewise linear voltage sources as input sources for testing purpose. 66

4.2 Time diagram of output voltage, VO, and inductors current when a failure occurs in a two

converters’ system. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

4.3 Proposed circuit for impedance variation for testing purpose. . . . . . . . . . . . . . . . . 68

4.4 Time diagram of output voltage, VO and inductors and output current when decreasing

the impedance in a two converters’ system. . . . . . . . . . . . . . . . . . . . . . . . . . . 69

viii

4.5 Time diagram of output voltage, VO and inductors and output current when increasing the

impedance in a two converters’ system. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

4.6 Proposed circuit with piecewise linear voltage sources as input sources for testing pur-

pose in a three converters example. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

4.7 Sized circuit behaviour. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

4.8 Time diagram of output voltage, VO, and inductors current when a failure occurs in a three

converters’ system. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

A.1 Mechanical system of inertial energy harvester with viscous damping transduction [32]. . 87

A.2 Variable-capacitance-type electrostatic generating system [22]. . . . . . . . . . . . . . . . 88

B.1 LMC555 in Variable Duty Cycle Oscillator Configuration [59]. . . . . . . . . . . . . . . . . 91

B.2 LMC555 Configuration for 50% Duty Cycle [59]. . . . . . . . . . . . . . . . . . . . . . . . . 92

B.3 LMC 555 with 50% duty cycle. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93



ix

List of Tables

2.1 Comparison of independent system approaches for harvesting energy to power IMD. . . . 22

3.1 Descriptive equations of a single converter working principle. . . . . . . . . . . . . . . . . 34

3.2 Parameters for Continuous Conduction Mode. . . . . . . . . . . . . . . . . . . . . . . . . . 36

3.3 Converters Sizing for Continuous Conduction Mode, at start (when working individually). . 37

3.4 Simulation results analysis for scenario 1, 2 and 3 with both converters in CCM and work-

ing simultaneously. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

3.5 Parameters for CCM-CCM converters sizing example. . . . . . . . . . . . . . . . . . . . . 40

3.6 Parameters for Discontinuous Conduction Mode. . . . . . . . . . . . . . . . . . . . . . . . 42

3.7 Converters Sizing for Discontinuous Conduction Mode, at start (when working alone). . . 43

3.8 Simulation results analysis for scenario 1, 2 and 3 with both converters in discontinuous

conduction mode and working simultaneously. . . . . . . . . . . . . . . . . . . . . . . . . 44

3.9 Parameters for DCM-DCM converters sizing example. . . . . . . . . . . . . . . . . . . . . 46

3.10 Parameters for Continuous-Discontinuous Conduction Mode. . . . . . . . . . . . . . . . . 49

3.11 Converters Sizing for Continuous-Discontinuous Conduction Mode, at start (when working

individually). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

3.12 Simulation results analysis for scenario 1, 2 and 3 with converters in continuous-discontinuous

conduction mode, with converters working simultaneously. . . . . . . . . . . . . . . . . . . 50

3.13 Parameters for continuous-discontinuous conduction mode converters sizing example. . . 52

3.14 Vitatron pacemaker battery relevant parameters. . . . . . . . . . . . . . . . . . . . . . . . 55

3.15 Converters’ sizing parameters for CCM operation. . . . . . . . . . . . . . . . . . . . . . . 56

3.16 Parameters for CCM-CCM converters sizing example. . . . . . . . . . . . . . . . . . . . . 59

4.1 Parameters for a three converters sizing example. . . . . . . . . . . . . . . . . . . . . . . 72

xi

Acronyms

BFC Biofuel Cells

CCM Continuous Conduction Mode

DC-DC Direct Current to Direct Current

DC Direct Current

DCM Discontinuous Condustion Mode

EBC Enzematic Biofuel Cell

EMG Electromagnetic Generators

ESG Electrostatic Generators

IMD Implantable Medical Device

MEMS Microelectromechanical Systems

MFC Microbial Fuel Cell

MI Multiple Input

MOSFET Metal Oxide Semiconductor Field Effect Transistor

p-HEI Piezoelectricity Driven Hot-Injection Injectors

PDMS Polydimethylsiloxane

PET Polyethylene Terephthalate

PG Piezoelectric Generators

PVSC Pulsating Voltage Source Cell

PWM Pulse Width Modulation

xiii

PZT Lead Zirconate Titanate

RF Radiofrequency

TEG Thermoelectric Generators

TENG Triboelectric Nanogenerators

VCE Variable-Capacitance-Type Electrostatic

xiv

1Introduction

Contents

1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

1.2 Goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

1.3 Novelties of the Thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

1.4 Document Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

1

Introduction

1.1 Motivation

The growing need of powering electronic devices with everlasting batteries has made energy har-

vesting a subject with increasingly interest of study. There are many ongoing researches to discover

new techniques for supplying electronic devices that have limited lifetime power [1]. Energy harvesting

consists of scavenging energy from a source and then converting the harvested energy into electrical

energy capable of powering an usually small and low power device.

IMDs enable the monitoring of human body in real time and their application are increasingly expand-

ing, as illustrated in Figure 1.1. To power these devices, batteries have been the primary source, which

may be quite inconvenient. The batteries used for these medical devices have a considerable size and

limited lifetime, which means that they need to be maintained and replaced and this has repercussions

for patients (expensive and time consuming) [2]. As a solution, harvesting energy from the body or the

environment can be an alternative approach to the battery-powered medical devices. These techniques

minimize the need of a patient to undergo repeated surgeries to replace the batteries and for noninva-

sive devices, it would reduce the amount of wires used, which may lead to several complications, such

as skin infections.

Although most of implantable medical devices have low power requirements, the output power gen-

erated from energy harvesting systems may not be sufficient to efficiently power IMDs, in spite of using

only one or multiple energy harvesters as input power sources. Consequently, as the output voltage

generated is in the µV or mV order, it must be increased using voltage elevation circuits, such as Direct

Current to Direct Current (DC-DC) converters, and eventually gather the contribution of several sources.

However, when using these generators, their continuous working status cannot always be assured.

Which means that, if one input source fails and there’s no other source able to replace it, the whole

system will probably fail as well, compromising the functioning of the IMD or even leading to its failure.

3

Figure 1.1: Diversity of implantable medical device applications [3].

1.2 Goals

The main goal of this project is to develop a system capable of permanently assuring the power of

an implantable medical device, using more than one energy harvesting device for collecting energy from

human body activities.

To achieve the main objective, several goals have to be established, such as the performance of

a survey of the energy harvesting methods that are presently used in biomedical applications, as well

as the execution of a comparative analysis to infer which methods gather the best characteristics to

be implemented in the context of this project’s system. This survey is also intended to be a tutorial

study in what concerns the energy scavenging. After having the input energy sources, it is necessary

to implement a voltage elevation circuitry, since the output voltage generated by the energy harvesters

is not enough to power an implantable medical device, and it is also needed to dimension a system

capable of guarantee the existence of one working input source.

1.3 Novelties of the Thesis

The work presented in this Thesis includes new scientific results that, as far as the author knows,

have never been published in literature. The proposed work intends to be an autonomous system that

boosts the input voltage and automatically assures the continuous delivery of energy to an IMD. Within

the several existing systems, many are only intended to behave as voltage elevators and the ones that

allow a selective input source choice, require a microprocessor controlled decision, which demands a

lot more hardware and certainly results in a higher power consumption.

4

1.4 Document Structure

This document is organized in five different chapters:

• Chapter 1 is the introduction, where it is presented a framework of the project and in what it

consists.

• In chapter 2 it is presented the state of the art, where are presented several energy harvesting

techniques presently used in IMD’s context, as well as a discussion to infer which approach is more

suitable for this work and the proposed approach. A review of already existent voltage elevation

circuits and the selection of a circuit to apply in the proposed solution is also done.

• In chapter 3 it is described and explained the proposed solution, as well as an analysis of its

working principle and some simulations results in order to select the most adequate operation

mode for this project goals.

• In chapter 4 it is presented some simulation results regarding the circuit performance.

• Finally, in chapter 5, it is gathered the conclusions drawn from the performed work and presented

some follow up work for system improvement.

• Additional work is presented in appendix A, where the choice of the input voltage source generator

is supported and explained.

• In appendix B, it is explained in detail how the duty cycle is generated and a simulation of the

generator device performance is done.

5

2State of the Art

Contents

2.1 Energy Harvesting for Powering Implantable Medical Devices . . . . . . . . . . . . . 9

2.2 DC-DC Converters for Voltage Elevation . . . . . . . . . . . . . . . . . . . . . . . . . . 23

7

State of the Art

2.1 Energy Harvesting for Powering Implantable Medical Devices

There are several sources of energy to harvest, since the human body, which represents an excellent

source of energy produced through normal actions and physical activities, to environmental sources,

such as solar and infrared energy. The process of energy harvesting requires, apart from an energy

source, a small device capable of converting the energy that has been scavenging into electrical energy.

The systems that can implement energy harvesting methods are divided in two types: independent

systems, as presented in subsection 2.1.1, which don’t need an external unit to produce power and

systems with external unit, which obtain the power from an external unit, through energy transfer, such

as Radiofrequency (RF) transmission, inductive coupling and ultrasonic transducers. However, the use

of energy transmission systems requires an external unit which has several disadvantages, such as

the distance dependence between the transducer and the implant, side effects caused by transmission

techniques and large size. These disadvantages make the use of this type of systems improper in the

context of this project, so the following analysis will be focused on independent systems only.

2.1.1 Independent Systems

2.1.1.A Biofuel Cells

Biofuel Cells (BFC) are devices that convert biochemical into electrical energy and they are based

on electrochemical reactions. Fuel cells have two electrodes, an anode and a cathode, where chemical

reactions occur, oxidations and reduction, respectively, and an electrolyte that allows protons to move

between the two sides of the fuel cell. At the anode, a catalyst causes the fuel to undergo oxidation

reactions which will generate protons and electrons. The protons released from oxidation reaction flow

to the cathode through the electrolyte and at the same time electrons are drawn from the anode to the

cathode through an external circuit, originating an electrical current [3]. An illustrative schematic of a

biofuel cell conceptual view is presented in Figure 2.1.

9

Figure 2.1: Biofuel cell conceptual view [4].

Fuel cells can be classified based on its type of catalyst used: i) Microbial Fuel Cell (MFC) or ii)

Enzematic Biofuel Cell (EBC). In the context of IMD, MFC were first implemented in 1960s, when cell-

free enzyme-based fuel cells were used for implantable artificial heart. EBCs using glucose as fuel and

oxygen as oxidizer started to be investigated for IMD applications and since then, several approaches

have been proposed. For example, Mano et al. [5] proposed a miniaturized biofuel cell intended for

blood stream implementation which is able to produce 2.4 µW with output voltage of 0.52V. This device

employs glucose oxidation at the anode and oxygen at the cathode.

Although harvesting energy through BFC has advantages, such as the biological compatibility be-

tween them and the human body and the moderate operation conditions for the chemical reactions there

are some challenges to solve. The difficult to maintain the biocatalyst over a long period, the fact that

microwatt level of biofuel cells limits it’s use and the possibility of damaging the device or harming the

patients due to unavoidable biofuel are issues that need to be solved [3].

2.1.1.B Thermoelectric Generators

Heat is one of the possible sources of energy from the human body that can be harvested for power-

ing implantable medical devices. Thermoelectric sensors are based in the Seebeck effect and capable

of converting thermal into usable electrical energy. This effect consists of creating a voltage from tem-

perature gradient between hot and cold junctions, which is relatively small in the human body [6], as

shown in Figure 2.2. The efficiency of Thermoelectric Generators (TEG) is proportional to the Carnot

efficiency [7],

ηc = 1− TcTh

=Th − TcTh

, (2.1)

where ηc is the Carnot efficiency, Th and Tc represents hot and cold temperatures in Kelvin, respec-

tively.

10

Figure 2.2: Conceptual view of thermoelectricity [4].

TEG are normally made of semiconductor material, and the most common ones are Bismuth Telluride

(Bi2Te3) and polycrystalline silicon-germanium (poly-SiGe) film. Each thermoelectric module is usually

formed by n-doped and p-doped semiconductor thermocouple and it is placed electrically in series and

thermally in parallel.

The mechanism of these devices employs fabric and thermocouples and the output voltage is defined

as

V = nα∆TTEG, (2.2)

where n is the number of thermocouples, α is the Seebeck coefficient of the thermoelectric material

and ∆TTEG is the temperature difference between the hot and cold junctions. From the equation, the

dependence of temperature gradient in the two junctions in the TEG performance is clear.

In 2007, Yang et al. [8] exploited the thermal gradient between inner body and skin surface as source

for thermal energy harvesting with an implanted TEG and simulated it use in different depths and config-

urations. An illustration of the implanted TEG location is shown in Figure 2.3. The authors also propose

and discuss several approaches for increasing the energy generation, such as the intentional cooling

and heating of the skin surface. For their in vitro experiment, a TEG was implanted in a rabbit’s ab-

domen and the temperature difference achieved was 1.3 C with the TEG voltage of 5 mV. When the

rabbit skin surface was cooled with an ice water bag, the temperature difference increased up to 5.5 C

and the TEG voltage increased to 25 mV.

11

Figure 2.3: Implanted TEG for thermal energy harvesting and the modal of a tissue [8].

Nagel et al. [9] investigated the possibility of powering an artificial accommodation system within

an eye using a TEG, as shown is Figure 2.4. The characteristics of high-performance thermoelectric

materials and the temperature distribution within the human eye were considered to estimate power

generated by a TEG as part of the accommodation system. The achieved results were from 4.6 µW in

the worst case up to 24.4 µW in the best case.

Figure 2.4: Artificial accommodation system within an eye powered by a TEG [9].

TEG are only able to produce sufficient power for IMD which require low microwatt power, but they

have unlimited lifetime, since human body is an unlimited heat-energy source. In order to increase output

powers, several thermocouples have to be cascaded in a proper way, but there are several inconvenient

issues driven this approach, such as size, reliability and biocompatibility [6].

2.1.1.C Electromagnetic Generators

Human body is a great source of kinetic energy created through motions, such as walking, running

and related physical activities. Electromagnetic energy scavenging through Electromagnetic Generators

(EMG) can power implantable medical devices. These devices harvest energy based on the Faraday-

Neumann-Lenz law, which states that relative motion between a permanent magnet and a coil produces

a time-variable magnetic flux, generating a voltage, as shown in Figure 2.5. There are two approaches

for achieving this: i) relative motion is used while the generating system is fixed and ii) rigid body motion

is used weight the inertia force of a weight on the generator [10]. The most common power generation

12

approaches are through the relative movement of the magnet and coil or due to changes in the magnetic

field. Therefore, the amount of energy generated depends on the magnetic field strength, relative motion

velocity and the number coil turns.

Figure 2.5: Conceptual view of an EMG for powering IMD [4].

Electromagnetic techniques are suitable for powering IMD due to the low-frequency and irregular

movements of humans. Harvesting energy from repeated heart muscle contractions at a range of fre-

quency between 0.5 and 1 Hz, a power of 40 to 500 µW was achieved [11, 12]. Roberts et al. [13]

investigated an electromagnetic generator using Microelectromechanical Systems (MEMS) technology

to enhance the power for pacemaker batteries in clinical trials.

Luciano et al. [14, 15] developed a miniaturized electromagnetic generator which can be implanted

in a human knee prosthesis, shown in Figure 2.6. The permanent magnet is installed in the femur

and the coil is on top of tibia. A current is induced in the coil by the magnets and coil relative movement

which happens when the knee flexes. This system, with a power conditioning circuit, produces an output

energy of about 22.1 µJ in 7 s for a gait frequency of 1.02 Hz emulated with an electric motor.

Figure 2.6: EMG structure for knee prosthesis: a) permanent magnets location; and b) coils location. [14,15].

Nasiri et al. [16] investigated the use of a linear permanent magnet generator implanted at the ab-

dominal muscles. The linear generator consists of two layers of permanent magnets and one layer of

coils. The abdomen moves with a frequency of 0.3 Hz during breathing, producing a power of about 1.1

mW through an electromagnetic generator with a volume of 16 cm3. It harvested 9 mJ from 22.5 s of

walking motion.

13

Morais et al. [17] reported a nonlinear electromagnetic generator implantable in a hip prosthesis.

The generator consists of two teflon tube with one or two external coils and a magnet attached to a

spring, as shown in Figure 2.7. The device harvest energy from the human gait to power a telemetry

system inserted in a smart hip prosthesis implant for early detection of loosening and implant failure.

With a power management circuit, the device is able to collect the energy needed to power the telemetry

system (1.9 mJ) for 9.2 s after charging for 34.8 s from a walking speed of 1.3 Hz.

Figure 2.7: Implantable EMG in a hip prosthesis [17].

2.1.1.D Electrostatic Generators

Electrostatic induction can also be a source for harvesting energy from the human body which is

achieved using Electrostatic Generators (ESG). Electrostatic induction consists in the redistribution of

electrical charge in a material under the influence of nearby objects that have electric forces and it

is based on the electrostatic potential energy as consequence of conservative Coulomb forces [18].

ESG exploit mechanical motion and induce electrical energy as result of the effect of an electric field

in the transducer’s moving parts. Commonly, ESG consist of two conductive plates relatively mobile

and electrically isolated via air, vacuum or a dielectric insulator (capacitor) [19]. An illustration of the

conceptual view of electrostatic energy transfer is shown in Figure 2.8. The human body motion will

incite the movement or vibration of the movable electrode which will result in a distance variation of the

capacitor’s plates.

14

Figure 2.8: Conceptual view of ESG for harvesting energy from human body motion [4].

Harvesting energy with ESG can be done through two different approaches: i) constant charge mode

and ii) constant voltage mode. In both, capacitor plates are charged with an external battery. According

to ESG’s actuation direction, they can be categorized in three types: i) in-plane gap closing; ii) out-of-

plane closing and iii) in-plane overlap. The first is the one that offers highest output power, compared to

the others techniques [20].

Tashiro et al. [21] developed an electrostatic generator that harvests energy from ventricular wall

motion to power a cardiac pacemaker and the system includes a accelerometer placed on a dog’s heart

to assure that the same amount of acceleration sensed from the heart drives the generator. The average

power recorded for 2 hours at 180 heartbeats per minute was 36 µW. In another approach, Tashiro et

al. [22] exploit heartbeat for powering a cardiac pacemaker and they obtained 58 µW.

Miao et al. [23] developed an electrostatic generation system using MEMS based technology, a non-

resonant MEMS based electrostatic generator and it was obtained for a movement of 0.1 m/s, at an

operation frequency of 30 Hz, an average power of 80 µW. A prototype of the electrostatic generator is

shown in Figure 2.9.

Figure 2.9: Prototype of a MEMS based electrostatic generator [23].

These devices are good candidates in the context of IMD due to the good integration capabilities

with microelectronic circuits and the fact that ESG can be built by silicon micromachining fabrication

techniques, which allow the use of MEMS and enable the miniaturized size, making them suitable for

implantable applications [22, 23]. However, the high output impedance and voltage of ESG make them

less suitable for power supply devices and the amount of energy produced is low. The major disadvan-

tage of this method is the need to initially charge the capacitor with an additional voltage source [4].

15

2.1.1.E Piezoelectric Generators

Another possible approach to harvest energy from human body motion is through Piezoelectric Gen-

erators (PG). These generators are made with piezoelectric materials which can generate power via

piezoelectricity, being the Lead Zirconate Titanate (PZT) the most widely used piezoelectric ceramic

material. Piezoelectricity was discovered in 1880 by bothers Jacque and Pierre Currie and consists of

subjecting certain crystals to a mechanical strain to cause an electrical polarization, which is propor-

tional to the applied strain [2]. A conceptual view of piezoelectric generators is shown in Figure 2.10.

Figure 2.10: Conceptual view of piezoelectric generators for harvesting energy from human body motion [4].

The physical phenomenon associated to this devices is based on the fact that the electrical charge

accumulated in a certain material will be induced when the material is subjected to mechanical trans-

formation, which means that piezoelectric transducers can exploit the mechanical energy produced by

human body motions. There are two categories to classify human body motions: i) discontinuous motion

and ii) continuous motion. The first refers to motions like walking, running or hand movement and the

last refers to motions like human breathing, heartbeats or blood flow.

The use of discontinuous motion has been studied and the devices that require this type of motions

make use of piezoelectric energy harvesting through a piezoelectric transducer placed in moving loca-

tions of the human body. For example, the knee is commonly used for placing piezoelectric transducers,

as it is exposed to a force up to three times higher than the body weigh [24]. Platt et al. [24, 25] exploit

the use of three piezoelectric elements inside of orthopedic implants to harvest energy from human body

motion, as shown in Figure 2.11. The relative motion between the femoral and bearing surfaces allows

the knee to function and the tibial tray supports a low-friction polyethylene bearing surface. When the

femoral component applies some force, this force will be applied to the bearing surface and to the three

piezoelectric stacks. A 900 N force in applied over one piezoelectric stack placed inside the prototype in

a laboratory setup, it generates up to 1.6 mW of power and a total of 4.8 mW for the three stacks, whose

dimensions is 10 x 10 x 20 mm.

16

Figure 2.11: Knee implant with three piezoelectric stacks [25].

Almouahed et al. [26,27] developed a more sophisticated knee implant. This device uses four smaller

piezoelectric elements of dimension 10 x 10 x 4 mm. The power generated for a single element with a

resistive load of 50 kΩ is about 1 mW, higher than the previous approach, and the total power generated

is about 4 mW. Cheng et al. [28] exploit PZT ceramics with dimensions of 5 x 5 x 18 mm with an

associated circuit, applicable to orthopedic implants. The collected power with also four piezoelectric

elements is about 1 mW.

Lewandowski et al. [29] reported an implantable piezoelectric generator device attached in series

with a muscle tendon to harvest energy from the muscles. Muscles contractions are electrically stimu-

lated and exert force on the piezoelectric generator, producing a charge. The charge is collected in an

energy storage circuitry and used to power the simulator and other loads. A conceptual block diagram

is presented in Figure 2.12. The system targets individual with extensive paralysis, where the electri-

cally simulated muscle wouldn’t interfere with natural muscle contraction or activities. A small PZT stack

prototype (5 x 5 x 18 mm) is able to generate 80 µW of power under a force application of 250 N.

Zhou et al. [30] presented the use of Piezoelectricity Driven Hot-Injection Injectors (p-HEI) as a

method for self-powered biomechanical health, where the use of batteries or remote powering is consid-

ered to be impractical. As shown in Figure 2.13, the work principle of a generic p-HEI device consists

in a piezoelectric transducer that harvests energy from mechanical strain variations to generate high-

energy electrons in the channel of a Metal Oxide Semiconductor Field Effect Transistor (MOSFET) [31].

When the energy of these electrons exceeds the energy barrier of the silicon, the electrons surmount

the barrier and get trapped onto a floating-gate. Due to the floating-gate be electrically isolated, the

injected electrons remain trapped for a long period of time. As the piezoelectric transducer is excited,

more electrons are injected and the total amount of charge stored increase with the duration and mag-

nitude of the mechanical excitation. The proposed injector can operate at 10 nW. This result makes the

injector suitable for designing structural health monitoring sensors that can be embedded and implanted

inside structures, although it is a lower order of magnitude.

17

Figure 2.12: Conceptual block diagram of an implanted piezoelectric generator from muscles contractions [29].

Figure 2.13: Principle of operation and applications of a p-HEI device [30].

There are several examples of piezoelectric systems that use continuous human body motion, such

as the system reported in [32], which harvests energy from heartbeats for powering cardiac implant

devices. The output power depends on the heartbeats’ acceleration spectrum and the achievable power

level is about 100 µW for a system with dimensions 15 x 7 x 5 mm.

In 2014, Deterre et al. [33] developed a device to power a leadless pacemaker based on a microspiral-

shaped piezoelectric energy harvester that collects energy from ordinary blood pressure variations in the

cardiac environment, as shown in Figure 2.14. This device enables direct blood pressure harvesting and

enables a high efficiency of energy transfer to a transducer operating in quasi-static mode (for the best

design, it was obtained 3 µJ/cm3/heartbeat and a transduction efficiency of 5.7x10−3 at 1.5 Hz) and

hence adaptable and unaffected by frequency heartbeat frequency changes.

In the context of IMD, piezoelectric harvesters are a good possibility, but their location in the human

body is limited due to the fact that it is required significant movements to produce power. The biggest

challenges of these devices are small size and biocompatibility issues.

18

Figure 2.14: Implantable energy harvester that uses blood pressure variations [33].

2.1.1.F Triboelectric Nanogenerators

Triboelectric Nanogenerators (TENG) are emerging as a solution for harvesting energy from the hu-

man body. These generators are made of two materials that use triboelectrification effect, which consists

in two dissimilar materials come into contact and one becomes electrically charged after contact with

the other material [34]. Almost all materials suffer from this effect, both natural or synthetic, from metals

to polymers. TENG systems operate by employing the coupling effects between triboelectrification and

electrostatic induction, which is a result of either contact separation, as shown in Figure 2.15, or relative

sliding between the two materials, as shown in Figure 2.16 [34, 35]. The power generation is a conse-

quence of triboelectric effect that induces charges on the surface of materials and results in electron

flow between the electrodes [36].

Figure 2.15: Working mechanism of TENG, using contact separation [37].

Figure 2.16: Working mechanism of TENG, using contact sliding between two materials [37].

19

Although the application of these devices is, so far, unknown for the implantable medical devices

context, there are some applications that make use of it. For example, Hou et al. [36] reported a TENG

based on the cycle contact-separation between a Polydimethylsiloxane (PDMS) film and a polyethylene

Polyethylene Terephthalate (PET) film, for harvesting footfall energy. The maximum output voltage and

current density reached up to 220 V and 40 µA.

2.1.1.G Photovoltaic Cells

The implant devices described above represent independent systems, capable of harvesting energy

from the human body for medical implanted devices. However, it is possible for a implantable medical

device to collect energy from outside the human body without having to use an external unit to achieve

it. Outdoor solar energy provides the highest power density among ambient energy sources, but human

body tissue reduces the light penetration [38]. A solar energy harvester for IMD make possible the

collection of solar energy in a certain wavelength range, where the optical absorption is small. An

implanted subcutaneous solar cell, as the one shown in Figure 2.17, is able to harvest power in the

order of microwatts in bright ambient conditions [39].

Figure 2.17: Solar energy harvester [39].

In 2017, Chen et al. [40] developed a single-chip solar energy harvesting system for a subdermal

implant application. For this application, the output power expected is in the order of µW and the key

challenge is to achieve high energy efficiency at ultra-low power levels in a small volume. The goal of

having a complete highly efficient energy harvesting system with a high output voltage as well as an

ultra-compact form is achieved through the single-chip solution. The incoming solar energy is harvested

through a solar cell and it provides power the other building blocks and to the load. This system uses a 3-

stage integrated pump with on-chip photodiodes that improve the efficiency to 3.5 times, when compared

with the conventional stacked photodiode approach. For the 1.54 mm2 system area, under an incident

power of 1.22 mW/cm2 from a halogen light source, a harvested power of 2.58 µW was obtained. As the

system consumed 523 nW, the resulted power delivered was 1.65 µW at 64% charge pump efficiency.

20

2.1.2 Discussion and Proposed Approach

There are several alternatives to avoid the use of limited lifetime and power density batteries in the

context of implantable medical devices. Some of these alternatives have been presented above and all

of them take advantage of harvesting energy either from human body or the surrounding environment.

Harvesting energy for powering implants minimizes the need of a patient to undergo repeated surgeries

to replace batteries and it reduces the amount of wires used that may lead to several complications.

Some of the requirements of energy harvesters are the generation of sufficient output power to power

the implanted devices, the minimum size possible and the biological compatibility. Based on these re-

quirements, a comparison study between the various methods for harvesting energy using independent

systems has been done.

Independent systems represent energy harvesting devices that don’t require an external unit to gen-

erate power, they collect energy from human body motion or chemical reactions present in the human

body, for example. This fact already represents an advantage over systems that require an external unit,

since the devices must be as simple as possible to be implanted in the human body without depriving

patients from doing ordinary activities.Table 2.1 compares the several harvesting techniques, generated

power, the size of the devices and its advantages and disadvantages. From the analysis of the sev-

eral approaches with respect to the parameters referred, it is possible to conclude that piezoelectric

harvesters are the most promising in terms of generated power, but their location in the human body

is limited, since these devices require significant force to generate power. Thermoelectric harvesters

have unlimited lifetime but they have size issues. Electromagnetic generators produce lower power than

piezoelectric device but don’t require large forces, making them suitable for any body part. Electrostatic

devices are also good candidates to power IMD due to the good integration capabilities with microelec-

tronic circuits which enable the miniaturized size. However, the need of an additional power source to

initially charge the capacitor is the major disadvantage of these devices. Biofuel cells have some advan-

tages in the context of IMDs, such as biocompatibility, but the difficult to maintain the biocatalysts over a

long period or the unavoidable biofuel which can damaging the device, represent some issues that need

to be solved. Finally, despite of subcutaneous photovoltaic cells having an unlimited power source, the

output power is very low due to the biological tissue attenuation.

21

Tabl

e2.

1:C

ompa

rison

ofin

depe

nden

tsys

tem

appr

oach

esfo

rhar

vest

ing

ener

gyto

pow

erIM

D.

App

roac

hG

ener

ated

Pow

erS

ize

Aut

hor,

Ref

eren

ceA

dvan

tage

sD

isad

vant

ages

Bio

fuel

Cel

l2.

4µ

W-

Man

o,[5

]B

iolo

gica

lcom

patib

ility

Life

time

Out

putp

ower

Ther

moe

lect

ricG

ener

ator

s5

mV

upto

25m

V20

x20

x3

mm

Yang

,[8]

Unl

imite

dlif

etim

eR

elia

bilit

yB

ioco

mpa

tibili

tyis

sues

4.6µ

Wup

to24

.4µ

W58

.9m

m2

Nag

el,[

9]

Ele

ctro

mag

netic

Gen

erat

ors

22.1µ

J-

Luci

ano,

[14,

15]

Unl

imite

dlo

catio

nsC

ompl

exfa

bric

atio

nte

chno

logi

es1.

1m

W27

.8cm

2N

asiri

,[16

]1.

9m

J3.

76cm

2M

orai

s,[1

7]

Ele

ctro

stat

icG

ener

ator

s36

µW

50x

30x

30m

mTa

shiro

,[21

]H

igh

outp

utpo

wer

Add

ition

alpo

wer

sour

ceH

igh

outp

utim

peda

nce

58µ

W-

Tash

iro,[

22]

80µ

W-

Mia

o,[2

3]

Pie

zoel

ectr

icG

ener

ator

s

4.8

mW

10x

10x

20m

mP

latt,

[24,

25]

Hig

hou

tput

pow

erN

oad

ditio

nalv

olta

geso

urce

Lim

ited

impl

anta

ble

loca

tions

Bio

com

patib

ility

issu

es

4m

W10

x10

x4

mm

Alm

ouah

ed,[

26,2

7]1

mW

5x

5x

18m

mC

hen,

[41]

80µ

W5

x5

x18

mm

Lew

ando

wsk

i,[2

9]10

nW-

Zhou

,[30

]10

0µ

W17

x7

x5

mm

Det

erre

,[32

,33]

Sub

cuta

neou

sP

hoto

volta

icC

ell

2.58

µW

1.54

mm

2C

hen,

[40]

Unl

imite

dpo

wer

sour

ceLo

wou

tput

pow

er

*Not

e:Tr

iboe

lect

ricG

ener

ator

sar

eno

tinc

lude

dsi

nce

ther

ear

eno

know

nap

plic

atio

nsin

the

cont

exto

fIM

D.

22

Comparing the several methods presented, it is possible to conclude that piezoelectric and elec-

trostatic generators gather the best characteristics among all approaches for the implantable medical

devices context. Piezoelectric generators present the best results in terms of generated output power

and electrostatic generators, besides the good results of generated output power, their localization is

almost limitless, since they do not need large amounts of force.

In order to take advantage of both output power generators, combining their advantages instead of

dispensing one for the other, it has been considered the possibility of combining several approaches,

resulting in a system with multiple energy harvesters as input power sources.

2.2 DC-DC Converters for Voltage Elevation

Although most of implantable medical devices have low power requirements, the output power gen-

erated from energy harvesting systems may not be sufficient to efficiently power IMD, in spite of using

only one or multiple energy harvesters as input power sources. Consequently, as the output voltage

generated is in the µV or mV order, it must be increased using voltage elevation circuits, such as DC-DC

converters.

2.2.1 Single Source DC-DC Converters

There are three topologies of DC-DC converters for boosting voltage: i) inductive converters, which

use inductive elements (e.g. boost converters); ii) capacitive converters, which use capacitive elements

(e.g. charge pump) and iii) hybrid converters, using both inductive and capacitive architecture.

Concerning inductive converters, there are several approaches based on the use of boost converters.

Bourgoine [42] reported a boost converter capable of converting an input voltage of 250 mV into a range

of values between 1.25 V and 5.25 V, being the range controlled by a current divider placed in circuit

output. Ramadass et al. [43] developed a boost converter for a TEG, using a capacitor to store the

energy and a controlling system to obtain a regulated output voltage of 1.8 V. This circuit needs an

external voltage to power the converter, since the TEG output voltage is too low.

A capacitive architecture example is the power management circuit for a TEG, developed by Doms

et al. [44]. This circuit contains a charge pump and a maximum of 8 stages and is able to convert an

input voltage of 0.6 V with a maximum total efficiency of 70%, but it requires a start-up voltage of 2 V.

Regarding hybrid converters, Richelli et al. [45] proposed a hybrid inductive and capacitive archi-

tecture, which can boost an input voltage of 200 mV to a output voltage of 1.2 V with a current of 120

µA delivered to the load and a maximum efficiency of 36%. This hybrid converter is constituted by two

boost converters which elevate the input voltage up to 0.8 V and a two-stage charge pump that boost

23

the voltage up to 1.2 V, as shown in Figure 2.18. Fontela [46] improved the same circuit and enabled a

output voltage of 2.4 V.

Figure 2.18: Block diagram of the hybrid inductive capacitive converter [45].

2.2.2 Multiple Source DC-DC Converters

Collecting power from multiple energy sources can increase the reliability of the system and enables

the combination of advantages of different sources with different voltage and current characteristics for

optimal energy and economic use. Multiple Input (MI) converters have been proposed and they can

implement two different configurations: i) combining various input energy sources in parallel and ii)

connecting the input voltage sources in series, to supply power simultaneously.

Khaligh et al. [47] proposed a multiple input hybrid energy conversion topology, as shown in Figure

2.19. This converter combines the various input power sources in parallel, can operate bidirectionally

and is capable of operating in buck, buck-boost and boost mode separately.

Figure 2.19: Multiple input converter topology [47].

24

Chen et al. [48] developed a double input Pulse Width Modulation (PWM) buck DC-DC converter,

as shown in Figure 2.20. This converter has four different operation modes, depending on the status of

the power switches and by applying the PWM control scheme, the converter can draw power from two

voltage sources individually or simultaneously. When the converter is operating with both input sources

transferring power to the load and if one of them is disconnected, the other can continue to deliver

power to the load normally, which represents a major advantage comparing to the next converters’

configuration.

Figure 2.20: Double input PWM DC-DC converter [48].

In order to supply power simultaneously from multiple input power sources, some topologies imple-

ment a series configuration of the input sources. Kumar et al. [49] proposed a non-isolated multiple

source DC-DC converter, as shown in Figure 2.21. In this topology, the input sources are connected

in series through power switches. Each series connected source and the corresponding power switch

forms a Pulsating Voltage Source Cell (PVSC) and enables inherent bypass circuitry for other sources.

This converter, as the one reported in [47], can also supply the load with the energy sources connected

simultaneously or individually and possibilities bidirectional power flow with buck, buck-boost and boost

operation modes.

Figure 2.21: Structure of multiple input source converter [49].

Deihimi et al. [50] reported a multiple input step-up DC-DC converter to interface multiple energy

25

sources of different output characteristics with a common load, as shown in Figure 2.22. This topology

provides higher voltage gains by increasing the number of inputs. Input energy sources are controlled

simultaneously to supply the load while adjusting the output voltage at the desired level.

Figure 2.22: Multiple input step-up converter [50].

Depending on the final application of the converter, series or parallel input power sources configu-

ration may be more or less suitable. Although parallel configuration can have the disadvantage of only

one input source contribute at a time and consequently the output voltage is not as high as with a se-

ries configuration with simultaneously contributions, it has the advantage of when an input source is

disconnected the system is able to continue working by using another input source.

26

3Proposed Solution

Contents

3.1 Working Principle Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

3.2 System Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

27

According to the goals of this work, for creating a redundant system capable of harvesting energy to

power up an implantable medical device that always guarantees the existence of a working input power

source, there are some aspects that have to be considered. First, the goal of the DC-DC converter is

input voltage boosting, so a boost DC-DC converter must be used. Secondly, the system must be as

simple as possible, once it will be implanted in human body, so it has been adopted a simple approach

that consists in using a system with multiple input power sources with boost DC-DC converters without

using additional control systems. This way, it is possible to boost each input contribution at a time to

achieve a certain output voltage and when an input source is disconnected, other input source starts

automatically working to guarantee the required output voltage. The block diagram of the proposed

circuit for two input power sources is represented in Figure 3.1.

Figure 3.1: Block diagram of the proposed circuit.

This system is intended to be set between the first voltage elevation circuit, if needed, and the IMD,

as shown in Figure 3.2. Since the energy harvesters are able to generate only µW or mW of output

power, it is needed to boost it before applying it to the proposed redundant system. This way, it is correct

to say that this system is seen by the IMD as its battery.

Figure 3.2: Block diagram of the entire system.

3.1 Working Principle Analysis

The proposed circuit topology in this thesis consists in multiple parallel boost DC-DC converters

sharing the same load, as shown in Figure 3.3 for a two converters case.

29

For the simplicity of the analysis, only two converters will be considered and the goal is to predict the

simultaneous functioning, considering the different operation modes for both converters.

Figure 3.3: Proposed Topology.

However, for best understanding the multiple converter functioning, a single converter steady state

analysis will be presented first.

3.1.1 Single Converter Working Principle

Considering firstly the individual working principle for one converter, this is an ordinary boost DC-DC

converter and as represented in Figure 3.4, it’s circuit is constituted by a coil L, a switch S, a diode

D1 and a capacitor C. The switch is periodically opened and closed, being controlled with a switching

period of

fsw =1

T(3.1)

where T represents the period and fsw represents the switching frequency.

Figure 3.4: Boost converter schematics [51].

The duty cycle, D, represents the time fraction the switch is closed and the output is high during

one complete working period, and it can take values between 0 and 1. The output voltage of this circuit

is nearly constant since the time constant of the resistance and the capacitor is much higher than the

30

switching period,

RC >> T. (3.2)

When the switch is closed, the current will flow through the inductor in clockwise direction and will

create a magnetic field. Therefore, the inductor will store some energy and in this moment the diode

is not conducting. When the switch opens, the current will flow to the diode and the magnetic field

previously created will be eliminated to maintain the current towards the load.

There are two operating modes for boost converters: i) CCM and ii) DCM. The circuit operates in

CCM if iL does not reach zero during all the switching period and it operates in DCM if iL is zero during

part of the switching period [51].

Considering the switch and the diode as ideal when the circuit is operating in the continuous mode

and the switch is closed, as shown in Figure 3.5, the voltage drop at the inductor, in steady state, is

vL = VI , (3.3)

which results in

∆i(1)L =

VILDT. (3.4)

Figure 3.5: Continuous operation mode [51].

When the switch is open, D1 conducts, which will ensure the inductor current continuity, leading to

vL = VI − VO, (3.5)

resulting, in steady state, in

∆i(2)L =

VI − VOL

(1−D)T. (3.6)

31

So, the average value of the input current, II will be equal to the average value of the current that

flows through the inductor, which is

IL =1

2

VILDDT +

1

2

VI − VOL

(1−D)DT = II (3.7)

Equaling equation 3.4 and equation 3.6 results the output voltage in steady state, which is

VO =VI

1−D. (3.8)

According to [51] and considering that the voltage drop at the switch and at the diode are VS and VD,

respectively, the efficiency is

η = 1− VSVID − VD

VO. (3.9)

The output voltage ripple can be calculated knowing that during the interval DT the diode is turned

off and the capacitor discharges through the resistance, as shown in Figure 3.6,

∆VO ≈1

CIODT =

1

C

VORDT. (3.10)

Figure 3.6: Determination of the output voltage ripple [51].

The circuit operates in continuous mode while the average current at the coil doesn’t reach zero,

which means, while IL is

IL >∆iL

2(3.11)

and the condition to operate in this mode is

L

R>D(1−D)2

2fsw. (3.12)

32

If the average inductor current is half of the inductor peak-to-peak ripple,

IL =∆iL

2, (3.13)

occurs the boundary mode of continuous conduction. In this mode the inductor current is

ILB =∆iL

2=

1

2

VILDT. (3.14)

When the circuit is operating in discontinuous mode, in steady state, as shown in Figure 3.7, the

voltage drop at the inductor is given by equation 3.3, when the switch is closed, and when the switch

opens is given by equation 3.5. The inductor current variation during DT is

∆i(1)L =

VILDT (3.15)

and during D0T is

∆i(2)L =

VI − VOL

D0T. (3.16)

Figure 3.7: Discontinuous operation mode [51].

Considering in steady state

VL = VIDT + (VI − VO)D0T = 0, (3.17)

33

it is possible to determine the output voltage, which is

VO =D +D0

D0VI . (3.18)

The input current average value, II that is equal to the average current that flows through the inductor

is

IL =∆IL

2=

∆iL2

=1

2

VILDT (D +D0) = II . (3.19)

The value of D0 is given by

D +D0

DD20

=RT

2L⇒ D0 =

L

DRT± 2

√( L

DRT

)2+

2L

RT. (3.20)

Summarizing, the equations that describe the working principle for a single boost DC-DC converter

are presented in Table 3.1.

Table 3.1: Descriptive equations of a single converter working principle.

Operation ModeCondition D0 Output Voltage

Continuous Conduction ModeL

R>D(1−D)2

2fsw- VO =

VI1−D

Discontinuous Conduction ModeL

R<D(1−D)2

2fswD0 =

L

DRT± 2

√( L

DRT

)2+

2L

RTVO =

D +D0

D0VI

3.1.2 Multiple Converters Working Principle

The main goal of this project is to guarantee, using an architecture as simple as possible and prefer-

ably without a control system, a certain voltage amount in order to power up a implantable medical

device. The supply is intended to be through the use of multiple boost DC-DC converters that have

sensors in their inputs and which will readjust every time a sensor is disconnected or fails. This way, it is

needed to assure that the converter’s working principle is verified when multiple parallel converters are

displayed and also that the voltage amount is maintained.

Since the system have multiple input power sources, each source has one DC-DC boost converter,

as shown in Figure 3.8.

34

Figure 3.8: Multiple parallel DC-DC boost converters sharing a same load.

The general circuit working principle is very similar to the previous described boost DC-DC converter

working principle. For this analysis, two input power sources are considered, VI1 and VI2 , also both

diodes, D1 and D2, and switches, S1 and S2, which are considered as ideal and the resistance RL

represents the load.

For studying the steady state behavior of the two converters working simultaneously and considering

that each one, when working alone, has two operation modes (continuous and discontinuous conduction

mode), it is considered the existence of four combinations, based on the converter’s individual working

principle, namely:

• Both converters in continuous conduction mode, CCM;

• Both converters in discontinuous conduction mode, DCM;

• Upper converter in continuous conduction mode and lower converter in discontinuous conduction

mode, CCM-DCM;

• Upper converter in discontinuous conduction mode and lower converter in continuous conduction

mode, DCM-CCM.

Once the position of the converters will not interfere with behavior of the set, the last two items can be

arranged in only one combination.

Considering the converters sizing, for each mentioned combination, there are the following scenarios:

1. VI1 > VI2 :

35

(a) VO1> VO2

;

(b) VO1 = VO2 ;

(c) VO1< VO2

.

2. VI1 = VI2 :

(a) VO1 > VO2 ;

(b) VO1= VO2

;

(c) VO1< VO2

.

3. VI1 < VI2 :

(a) VO1> VO2

;

(b) VO1 = VO2 ;

(c) VO1< VO2

.

In order to be able to study the behavior of the converters’ set, the previous mentioned sizing was

be simulated and a steady state analysis of the simulation results will be done. Also, a prevision of the

general working principle of the converters’ set will be held through the simulation results.

3.1.2.A Continuous-Continuous Conduction Mode

For the existence of this operation mode, it is needed to guarantee that continuous conduction mode

is verified when individually sizing the converters. This way, considering Equation 3.12, a set of possible

parameters is presented in Table 3.2.

Table 3.2: Parameters for Continuous Conduction Mode.

Commuting Frequency, fsw 10 kHzCapacitance, C 500 µF

Load, RL 10 Ω

Taking into account the converters’ individual sizing resulting scenarios for this operation mode, some

possible parameters are presented in Table 3.3, considering Equations 3.8 and 3.12. It was considered

that the switches’ duty cycle were synchronized instead of complementary, since there is no difference

once the diodes don’t conduct in reverse bias and VO > ∀VI , so one input’s the inductor current will

never flow backwards and will never interfere with another input.

36

Table 3.3: Converters Sizing for Continuous Conduction Mode, at start (when working individually).

Scenario 1VI1 > VI2

Scenario 2VI1 = VI2

Scenario 3VI1 < VI2

a) b) c) a) b) c) a) b) c)

Converter 1

Input Voltage, VI1 0.7 V 0.7 V 0.7 V 0.7 V 0.7 V 0.7 V 0.35 V 0.35 V 0.35 VDuty Cycle, D1 50% 50% 50% 50% 50% 30% 85% 75% 50%Inductance, L1 0.36 mH 0.36 mH 0.36 mH

Inductor Average Current, IL 48.6 mA 48.6 mA 48.6 mA 48.6 mA 48.6 mA 29.2 mA 42.4 mA 37.5 mA 25 mAOutput Voltage, VO1

1.4 V 1.4 V 1.4 V 1.4 V 1.4 V 1 V 2.3 V 1.4 V 0.7 V

Converter 2

Input Voltage, VI2 0.35 V 0.35 V 0.35 V 0.7 V 0.7 V 0.7 V 0.7 V 0.7 V 0.7 VDuty Cycle, D2 50% 75% 85% 30% 50% 50% 50% 50% 50%Inductance, L2 0.36 mH 0.36 mH 0.36 mH

Inductor Average Current, IL 25 mA 37.5 mA 42.5 mA 29.2 mA 48.6 mA 48.6 mA 48.6 mA 48.6 mA 48.6 mAOutput Voltage, VO2

0.7 V 1.4 V 2.3 V 1 V 1.4 V 1.4 V 1.4 V 1.4 V 1.4 V

The corresponding steady state simulations for Scenario 1, 2 and 3 with both converters in CCM,

working simultaneously, are shown in Figure 3.9, 3.10 and 3.11, respectively.

(a) Simulation results when VO1> VO2

. (b) Simulation results when VO1= VO2

. (c) Simulation results whenVO1 < VO2 .

Figure 3.9: Simulations results for Scenario 1, when VI1 > VI2 and both converters in CCM.



. (c) Simulation results whenVO1

< VO2.

Figure 3.10: Simulation results for Scenario 2, when VI1 = VI2 and both converters in CCM.

37




< VO2.

Figure 3.11: Simulation results for Scenario 3, when VI1 < VI2 and both converters in CCM.

From the simulation results, it is possible to verify the steady state behavior of the set, as it is pre-

sented in Table 3.4

Table 3.4: Simulation results analysis for scenario 1, 2 and 3 with both converters in CCM and working simultane-ously.

Scenario 1VI1 > VI2

Scenario 2VI1 = VI2

Scenario 3VI1 < VI2

a) b) c) a) b) c) a) b) c)

Converter 1Continuous Conduction Mode x(1) x(2) x(4) x(5) x(7)

Continuous Conduction Mode Limit x(8)Discontinuous Conduction Mode x(3) x(6) x(9)

Converter 2Continuous Conduction Mode x(3) x(5) x(6) x(8) x(9)

Continuous Conduction Mode Limit x(2)Discontinuous Conduction Mode x(1) x(4) x(7)

Higher Inductor Average Current, II IL1IL1

IL2IL1

IL1=IL2

I l2 IL1IL2

IL2

Set Output Voltage, VO VO1VO1

VO2VO1

VO1=VO2

VO2VO1

VO2VO2

Analyzing the simulation results and Table 3.4 it is possible to conclude that:

• When VO1> VO2

, for ∀ VI1 and ∀VI2 , the converter corresponding to VI1 is the one that enforce the

set output voltage, VO. The converter corresponding to VI2 operates in discontinuous conduction

mode, as shown in situations (1), (4) and (7).

• When VO1= VO2

, the converter corresponding to VI2 operates in the limit of continuous conduction

mode, for VI1 > VI2 (situation (2)). Situation (2) is the symmetric of (8).

– When VI1 = VI2 and VO1= VO2

, both converters contribute with 50% for the output voltage,

as presented in situation (5).

• Situations (3), (6) and (9) are the symmetric of (1), (4) and (7), respectively.

38

• The converter that enforces the output voltage is the master and it’s the one that, when working

alone, generates the higher output current and has higher inductor average current.

– The slave converter will contribute with the minimum amount of current.

Taking into account the final application of this work, it is intended to guarantee a certain amount of

VO, so the converters need to be sized to generate the same VO1 and VO2 . This way, the scenarios that

are more relevant for this analysis are the b) ones, where VO1= VO2

.

Next an analytic analysis will be performed to explain the above results.

Considering that both converters are working simultaneously and share the same load, the neces-

sary power to supply the load will be divided through both converters,

VI1II1 + VI2II2 = VOIO. (3.21)

The output voltage, VO, is supposed to be constant, which implies that IO has to be constant, and

consequently

VI1

(II1 +

VI2II2VI1

)=V 2O

R(3.22)

also has to be constant. This means that in the presence of a current, II2 , II1 drops apart fromVI2VI1

II2

and it is possible to conclude that is the total of the average currents that remains constant.

However, if VI1 >> VI2 , from Equation 3.21 results that VI1II1 must be constant, which means that

the current regime in L1 slightly decreases its average value and in L2, as D is the same, results a

smaller VO2. This happens due to decreases until discontinuous operating mode, where the value of

(1−D) is such that

VI2∆iL2

2(D +D0) =

V 2O

R, (3.23)

∆iL2 =VI2L2

DT. (3.24)

If VI1 drops its value, VO decreases, according to VO =VI1

1−D , until reaching the previous VO2value.

In this moment, IL2starts commanding and VO remains in VO2

.

39

For a situation where VI2 >> VI1 , the working principle would be equivalent to the previously de-

scribed one, but instead with IL2commanding in the first place.

Assuming that both converters are sized to generate the same output voltage individually, it is already

known that one converter will behavior as master and the other as slave (except for situation (9)). This

way, making use of Equation 3.21, knowing that IO =VOR

, and Equations 3.7 for the upper converter and

3.14 for the lower converter, it is possible to describe the converters simultaneous behavior in steady

state,

V 2O

R= VI1II1 + VI2II2

II1 =1

2

VI1L1

D1D1T +1

2

VI1 − VOL1

(1−D1)(1−D1)T + IOff =1

2

VI1L1

D1T + IOff

II2 =1

2

∆iL2

=1

2

VI2L2

D2T

VO =VI1

1−D1

⇔ (3.25)

VO =VI1

1−D1

II1 =VI1

2D2RLTD12 − 2VI2

2D2RLTD1 + VI22D2RLT − 2VI2

2L2

4VI1L2RLD1 − 2VI1L2RL − 2VI1L2RLD22

IOff =

(VI12D2RLTD1

2 − 2VI22D2RLTD1 + VI2

2D2RLT − 2VI22L2)L1 + VI1

2L2RLTD13

− 2VI12L2RLTD1

2 + VI12L2RLTD1

(4VI1L2RLD1 − 2VI1L2RL − 2VI1L2RLD22)L1

II2 =1

2

VI2L2

D2T

.

(3.26)

For exemplifying System of Equations 3.26, some values can be assigned, as presented in Table

3.5.

Table 3.5: Parameters for CCM-CCM converters sizing example.

Upper Converter Lower Converter

Input Voltage, VI VI1=0.7 V VI2 =0.35 VDuty Cycle, D D1=50% D2=75%Inductance, L L1=0.36 mH L2=0.36 mH

Inductor Average Inductance, IL IL1=48.6 mA IL2

=37.5 mA


Load, RL 10 Ω

40

Knowing that the condition for a boost converter to operate in continuous conduction mode is pre-

sented in Equation 3.12, it is possible to conclude that, if they are working alone, the converters are

operating in this mode, since the condition is verified for the upper converter

0.36× 10−3

10>

0.5(1− 0.5)2

2× 10× 103⇔ 3.6× 10−5 > 6.3× 10−6 (3.27)

and for the lower one

0.36× 10−3

10>

0.75(1− 0.75)2

2× 10× 103⇔ 3.6× 10−5 > 2.3× 10−6. (3.28)

The output power of each converter is also possible to determinate. For the upper converter, the

output voltage is given by

VO1 =VI1

1−D1=

0.7

0.5= 1.4 V. (3.29)

For the lower converter, the output voltage is given by

VO2=

VI21−D2

=0.35

0.75= 1.4 V. (3.30)

Considering now the steady state behavior of the set with both converters working simultaneously,

since the upper converter is the one that has the higher IL, it will be the master and will enforce the

output voltage of the set, VO = VO1= 1.4 V. The lower converter will behave as a slave and contribute

with the minimum output current. The converters are sized to generate the same amount of voltage, as

presented in Table 3.4, so the lower converter will be at the limit of continuous conduction mode.

Taking the solutions found in the system of equations 3.26, for the parameters of this example, the

values obtained are VO = 1.4 VII1 = 262 mAIOff = 213 mAII2 = 36.4 mA

. (3.31)

These values confirm the conclusions mentioned before and the simulation results that illustrate

these values is presented in Figure 3.12.

41

Figure 3.12: Steady state simulation results of the converters set with the sizing done in Table 3.5.

3.1.2.B Discontinuous-Discontinuous Conduction Mode

For the steady state analysis of this operation mode, it is needed to guarantee that discontinuous

conduction mode is verified when individually sizing the converters. This way, considering Equation

3.12, the parameters presented in Table 3.6 are considered.

Table 3.6: Parameters for Discontinuous Conduction Mode.


Load, RL 100 Ω


possible parameters are presented in Table 3.7, considering Equations 3.12,3.18 and 3.20.

42

Table 3.7: Converters Sizing for Discontinuous Conduction Mode, at start (when working alone).

Scenario 1VI1 > VI2

Scenario 2VI1 = VI2

Scenario 3VI1 < VI2

a) b) c) a) b) c) a) b) c)

Converter 1

Input Voltage, VI1 0.7 V 0.7 V 0.7 V 0.7 V 0.7 V 0.7 V 0.35 V 0.35 V 0.35 VDutty Cycle, D1 50% 21% 11% 50% 50% 29.7% 62% 62% 62%

D01 35% 49% 76% 35% 35% 41.6% 33% 33% 33%Inductance, L1 0.36 mH 0.36 mH 0.36 mH

Inductor Average Current, IL 41.3 mA 14.3 mA 9.3 mA 41.3 mA 41.3 mA 20.6 mA 20.6 mA 20.6 mA 20.6 mAOutput Voltage, VO1

1.7 V 1 V 0.8 V 1.7 V 1.7 V 1.2 V 1 V 1 V 1 V

Converter 2

Input Voltage, VI2 0.35 V 0.35 V 0.35 V 0.7 V 0.7 V 0.7 V 0.7 V 0.7 V 0.7 VDuty Cycle, D2 62% 62% 62% 29.7% 50% 50% 11% 21% 50%

D02 33% 33% 33% 50% 35% 35% 76% 49% 35%Inductance, L2 0.36 mH 0.36 mH 0.36 mH

Inductor Average Current, IL 28.6 mA 28.6 mA 28.6 mA 20.6 mA 41.3 mA 41.3 mA 9.3 mA 14.3 mA 41.3 mAOutput Voltage, VO2

1 V 1 V 1 V 1.2 V 1.7 V 1.7 V 0.8 V 1 V 1.7 V

The corresponding steady state simulations for Scenario 1, 2 and 3, when both converters are sized

to operate in DCM and working simultaneously, are shown in Figure 3.13, 3.14 and 3.15, respectively.




Figure 3.13: Simulations results for Scenario 1, when VI1 > VI2 with both converters in DCM.




< VO2.

Figure 3.14: Simulation results for Scenario 2, when VI1 = VI2 with both converters in DCM.

43




< VO2.

Figure 3.15: Simulation results for Scenario 3, when VI1 < VI2 with both converters in DCM.

From the simulation results, it is possible to verify the steady state behavior of the set, as it is pre-

sented in Table 3.8.

Table 3.8: Simulation results analysis for scenario 1, 2 and 3 with both converters in discontinuous conductionmode and working simultaneously.

Scenario 1VI1 > VI2

Scenario 2VI1 = VI2

Scenario 3VI1 < VI2

a) b) c) a) b) c) a) b) c)

Converter 1Continuous Conduction Mode

Continuous Conduction Mode LimitDiscontinuous Conduction Mode x(1) x(2) x(3) x(4) x(5) x(6) x(7) x(8) x(9)



Higher Inductor Average Current, II II1 II2 II2 II1 II1=II2 II2 II1 II1 II2

Analyzing the simulation results and Table 3.8 it is possible to conclude that when converters are

working simultaneously in steady state:

• The output voltage of the set, VO depends on both converters contribution, so a typical master-

slave topology cannot be applied in this case.

– The output voltage of the set, VO, is higher than any single converter output voltage VO1 or

VO2

– When any converter is disconnected, the value of the output voltage of the set will go down.

– The converter that has the higher inductor average current contributes with more current, and

consecutively, contributes more to the set’s output voltage.

44

• When VO1> VO2

, for ∀ VI1 and ∀ VI2 , ∆iL1> ∆iL2

, as demonstrated in situations (1), (4) and (7).

Symmetrically, when VO1< VO2

, for ∀ VI1 and ∀ VI2 , ∆iL1< ∆iL2

, as presented in situations (3),

(6) and (9). Also, when VO1 > VO2 , the converter corresponding to VI2 operates in discontinuous

conduction mode.

• When VO1= VO2

and VI1 > VI2 , ∆iL1< ∆iL2

, as shown in situation (2). Symmetrically, when

VO1= VO2

and VI1 < VI2 , ∆iL1> ∆iL2

, as presented in situation (8). If VO1= VO2

and VI1 = VI2 ,

∆iL1 = ∆iL1 , as situation (5) shows.

Again, taking into account the final application of this work, it is intended to guarantee a certain

amount of VO, so the converters need to be sized to generate the same VO1 and VO2 . This way, the

scenarios that are more relevant for this analysis are the b) ones, where VO1= VO2

.

Next an analytic analysis will be performed to explain the above conclusions.

Considering that both converters are working simultaneously and share the same load, assuming

that both converters are dimensioned to generate the same output voltage individually, it is already

known that both converters will contribute to the output current. This way, making use of Equation 3.21,

knowing that IO = VO

R and Equations 3.19 and 3.17, applied for each converter, it is possible to describe

the converters simultaneous steady state behavior and to predict the set’s output voltage,

V 2O

R= VI1II1 + VI2II2

II1 =1

2

VI1L1

D1T (D1 +DO1)

II2 =1

2

VI2L2

D2T (D2 +DO2)

⇔ (3.32)

VO =

2

√VI1

2RTD1DO1

L1+VI1

2RTD12

L1+VI2

2D2DO2RT

L2+VI2

2D22RT

L2

2√

2

II1 =(VI1D1DO1 + VI1D1

2)T

2L1

II2 =(VI2D2DO2

+ VI2D22)T

2L2

. (3.33)

For exemplifying this, some values can be assigned, as presented in Table 3.9, considering Equations

45

3.18 and 3.20 and that both converters are dimensioned for generate the same amount of output voltage

(it was considered VO1= VO2

= 1 V).

Table 3.9: Parameters for DCM-DCM converters sizing example.


Input Voltage, VI VI1=0.7 V VI2 =0.35 VDuty Cycle, D D1=21% D2=62%

D0 D01=49% D02=33%Inductance, L L1=0.36 mH L2=0.36 mH


=28.6 mA


Load, RL 100 Ω

Knowing that the condition for a boost converter to operate in discontinuous conduction mode is the

opposite of what is presented in Equation 3.12, it is possible to conclude that,in steady state, if they are

working alone, the converters are operating in discontinuous conduction mode, since the condition isn’t

verified either for the upper converter

0.36× 10−3

100≯

0.21(1− 0.21)2

2× 10× 103⇔ 3.6× 10−6 ≯ 6.6× 10−5 (3.34)

or for the lower one

0.36× 10−3

100≯

0.62(1− 0.62)2

2× 10× 103⇔ 3.6× 10−6 ≯ 4.5× 10−5. (3.35)

The output voltage of each converter is also possible to determinate. For the upper converter, the


VO1 =D1 +D01

D1VI1 =

0.21 + 0.49

0.490.7 = 1V. (3.36)


VO2 =D2 +D02

D2VI2 =

0.62 + 0.33

0.330.35 = 1V. (3.37)

Considering now the steady state behavior of the set with both converters working simultaneously

and taking the solutions found in the system of equations 3.33, for the parameters of this example, the

46

values obtained are

47

VO = 1.415 VII1 = 14.29 mAII2 = 28.63 mA

. (3.38)

As concluded before, both converters contribute to the output voltage of the set which results in a higher

value than VO1and VO2

. The lower converter will contribute more, since it has a higher II2 .

The steady state simulation results that illustrate these conclusions are presented in Figure 3.16.


3.1.2.C Continuous-Discontinuous Conduction Mode

For the steady state analysis of this operation mode, it is intended to dimension one converter to op-

erate in continuous conduction mode and the other in discontinuous conduction mode, when individually

sizing the converters. It doesn’t matter the order of the converter, since their operation is independent of

their positioning, so it will be considering that the upper converter (converter 1) will be operating in CCM

and the lower one (converter 2) in DCM. This way, considering Equation 3.12, the parameters presented

in Table 3.10 are considered.

48

Table 3.10: Parameters for Continuous-Discontinuous Conduction Mode.


Load, RL 10 Ω


possible parameters are presented in Table 3.11, considering Equations 3.8, 3.12, 3.18 and 3.20.

Table 3.11: Converters Sizing for Continuous-Discontinuous Conduction Mode, at start (when working individually).

Scenario 1VI1 > VI2

Scenario 2VI1 = VI2

Scenario 3VI1 < VI2

a) b) c) a) b) c) a) b) c)

Converter 1

Input Voltage, VI1 0.7 V 0.7 V 0.7 V 0.5 V 0.5 V 0.5 V 0.35 V 0.35 V 0.35 VDuty Cycle, D1 50% 50% 30% 75% 64% 50% 82,5% 75% 50%Inductance, L1 0.36 mH 0.36 mH 0.36 mH

Inductor Average Current, IL125 mA 37.5 mA 42.5 mA 29.2 mA 48.6 mA 48.6 mA 48.6 mA 48.6 mA 48.6 mA

Output Voltage, VO1 1.4 V 1.4 V 1 V 2 V 1.4 V 1 V 2 V 1.4 V 0.7 V

Converter 2

Input Voltage, VI2 0.35 V 0.5 V 0.5 V 0.5 V 0.5 V 0.5 V 0.5 V 0.5 V 0.5 VDuty Cycle, D2 30% 60% 60% 60% 60% 60% 60% 60% 60%

D02 15% 33% 33% 33% 33% 33% 33% 33% 33%Inductance, L2 0.036 mH 0.036 mH 0.036 mH

Inductor Average Current, IL265.6 mA 387.5 mA 387.5 mA 387.5 mA 387.5 mA 387.5 mA 387.5 mA 387.5 mA 387.5 mA

Output Voltage, VO21 V 1.4 V 1.4 V 1.4 V 1.4 V 1.4 V 1.4 V 1.4 V 1.4 V

The corresponding steady state simulations for Scenario 1, 2 and 3, when converters are operating

in CCM and DCM and working simultaneously, are shown in Figure 3.17, 3.18 and 3.19, respectively.




< VO2.

Figure 3.17: Simulations results for Scenario 1, when VI1 > VI2 with one converter in CCM and the other in DCM.

49




< VO2.

Figure 3.18: Simulation results for Scenario 2, when VI1 = VI2 with one converter in CCM and the other in DCM.




Figure 3.19: Simulation results for Scenario 3, when VI1 < VI2 with one converter in CCM and the other in DCM.

From the steady state simulation results, it is possible to verify the behavior of the set, as it is pre-

sented in Table 3.12.

Table 3.12: Simulation results analysis for scenario 1, 2 and 3 with converters in continuous-discontinuous conduc-tion mode, with converters working simultaneously.

Scenario 1VI1 > VI2

Scenario 2VI1 = VI2

Scenario 3VI1 < VI2

a) b) c) a) b) c) a) b) c)

Converter 1Continuous Conduction Mode x(1) x(4) x(7)

Continuous Conduction Mode LimitDiscontinuous Conduction Mode x(2) x(3) x(5) x(6) x(8) x(9)



Higher Inductor Average Current, II II1 II2 II2 II1 II2 II2 II1 II2 II2Set’s Output Voltage, VO VO1

VO1= VO2

VO2VO1

VO1= VO2

VO2VO1

VO1= VO2

VO2

50

Analyzing the steady state simulation results and Table 3.12 it is possible to conclude that:

• When VO1> VO2

, the set maintains the individual behavior (one converter operating in CCM

and other operating in DCM), as illustrated in situations (1), (4) and (7). In this case, as verified

in subsection 3.1.2.A, the converter 2 is the one that commands the set’s output voltage and a

master slave topology can be applied.

• When VO1≤ VO2

the converter 1 starts to have a behave like it is operating in discontinuous

conduction mode and the behavior of the converter’s set is the same as presented in subsection

3.1.2.B, as illustrated in situations (2), (3), (5), (6), (8) and (9).

Next an analytic analysis will be performed to explain the above results.

From the analysis of Table 3.12, it is possible to conclude that when both converters are working

simultaneously and share the same load it results in two different set’s converter behaviors: i) converter

1 in CCM and converter 2 in DCM (which output voltage will be the same as the converter 1’s individual

output voltage) and ii) converter 1 and 2 in DCM.

In behavior i), for describing and predict the converter set’s behavior it is needed to make use of

Equation 3.21, knowing that IO = VO

R and Equations 3.7, 3.17 and 3.8, resulting in

V 2O

R= VI1II1 + VI2II2

II1 =1

2

VI1L1

D1D1T +1

2

VI1 − VOL1

(1−D1)(1−D1)T + IOff =1

2

VI1L1

D1T + IOff

II2 =(VI2D2DO2

+ VI2D22)T

2L2

VO =VI1

1−D1

⇔ (3.39)

51

II1 =

(VI12D2

2 + VI12D2D02)RTD1

2 − (2VI22D2 + 2VI2

2D2D02)RTD1

+ (VI22D2

2 + VI22D2D02)RT − 2VI2

2L2

4VI1L2RLD1 − 2VI1L2RL − 2VI1L2RLD22

IOff = II1 ++VI1

2L2RLTD13 − 2VI1

2L2RLTD12 + VI1

2L2RLTD1

(4VI1L2RLD1 − 2VI1L2RL − 2VI1L2RLD22)L1

II2 =(VI2D2

2 + VI2D02D2)T

2L2

VO =VI1

1−D1

. (3.40)

In behavior ii), making use of Equation 3.21, knowing that IO = VO

R and Equations 3.17 and 3.19,

applied for each converter, as presented in System of Equations 3.32 and 3.33, it is possible to describe

the converters simultaneous steady state behavior and to predict the set’s output voltage.

Again, taking into account the final application of this work, it is intended to guarantee a certain

amount of VO, so the converters need to be dimensioned to generate the same VO1and VO2

. This way,

the scenarios that are more relevant for this analysis are the b) ones, where VO1= VO2

. Analyzing again

the Table 3.12, it is possible to conclude that the behavior of the set in this scenario corresponds to the

above mentioned behavior ii), so System of Equations solution 3.33 must be considered.

For exemplifying this, some values can be assigned, considering Equations 3.8, 3.18 and 3.20 and

that both converters are dimensioned for generate the same amount of output voltage (it was considered

VO1= VO2

= 1.4 V) and also considering (1−D1) = D01 , the values are presented in Table 3.13.

Table 3.13: Parameters for continuous-discontinuous conduction mode converters sizing example.


Input Voltage, VI VI1=0.7 V VI2 =0.5 VDuty Cycle, D D1=50% D2=60%

D0 1−D1=50% D02=33%Inductance, L L1=0.36 mH L2=0.036 mH


=387.5 mA


Load, RL 10 Ω

Knowing that the condition for a boost converter to operate in discontinuous conduction mode is the

opposite of what is presented in Equation 3.12, it is possible to conclude that, if they are working alone,

52

the lower converter is operating in discontinuous conduction mode, since the condition isn’t verified

0.036× 10−3

10≯

0.6(1− 0.6)2

2× 10× 103⇔ 3.6× 10−6 ≯ 4.8× 10−6, (3.41)

but the upper one is working on CCM,

0.36× 10−3

10>

0.5(1− 0.5)2

2× 10× 103⇔ 3.6× 10−5 > 6.25× 10−6. (3.42)

It is also possible to determine the output power of each converter. For the upper converter, the


VO1=

VI11−D1

=0.7

0.5= 1.4 V. (3.43)


VO2=D2 +D02

D2VI2 =

0.60 + 0.33

0.330.5 = 1.4 V. (3.44)

Considering now the steady state behavior of the set with both converters working simultaneously

and taking the solutions found in the systems of equations 3.33, for the parameters of this example, the

values obtained are VO = 1.509 VII1 = 48.6 mAII2 = 387.5 mA

. (3.45)

As concluded before, this behavior is very similar to the discontinuous-discontinuous conduction

mode, so both converters contribute to the output voltage of the set which results in a higher value than

VO1and VO2

, but the lower converter will contribute more, since it has a higher II2 .

The steady state simulation results that illustrate these conclusions are presented in Figure 3.20.

3.1.2.D Conclusions

From the previous converter behavior steady state analysis, it is possible to conclude that, as the

main goal of this work is to generate a certain amount of voltage to power up an IMD, the set of convert-

53


ers sizing has to guarantee the required amount specially when the system has to readjust to solve any

input source failure.

As previously concluded, within the three analyzed topologies, only the CCM-CCM topology main-

tains the output voltage when there are several input sources contributing, contrarily to either DCM-DCM

or CCM-DCM, which output voltage depends on the converters and its operation mode. This means that

if one source fails the output voltage may either increase or decrease its value, which is not the desired

result.

Since the more input sources are contributing to the output voltage, the least will be the necessary

delivered current from each converter to maintain the required output voltage. This way, the converters’

operation mode may change if its needed inductor current is too low. However, when this operation mode

alteration occurs, as it has already been seen, the output voltage may have a different value, besides

the desired one. So, it is needed to guarantee that the converters are always operating in CCM and

considering equation 3.12, it is necessary to set an inductor value that ensures the desired behavior for

the employed duty cycle values, once that the operating frequency and the load value are fixed.

3.2 System Architecture

Until now, the set of converters has been dimensioned for standard parameters in order to understand

its general steady state behaviour. As presented previously, this project intends to be a redundant

54

system for the human body’s energy harvesting technique to power an IMD, namely a pacemaker. This

means that, in order to study the behaviour as realistically as possible, the system has to be correctly

sized in terms of internal components, external load and output power.

3.2.1 Load Impedance

Taking as example the Vitatron E10 S Single Chamber Pacemaker System and its specification [52],

it is possible to gather some relevant information about its battery for circuit sizing, as presented in Table

3.14.

Table 3.14: Vitatron pacemaker battery relevant parameters.

Battery Voltage Impedance Reference Average Projected Capacity Longevity

2.8 V 500 Ω 0.92 Ah 10.4 years

From literature [53], [54] and [55], the impedance of a pacemaker varies from, at least, 200 to 4000

Ω and pacemaker batteries must be designed to cover this value. In the beginning of life, batteries are

projected to less than 1 kΩ impedance and over the years the impedance tends to get higher and the

current drain lower until it is time to renew it. The capacity of the battery is the estimated amount of

current that can be delivered to the load over time. For this example, a 0.92 Ah capacity means that with

a load of 500 Ω, the battery is capable of deliver 0.92 A per hour. As the longevity is presented as being

10.4 years, from

Time =Q

I, (3.46)

where Q represents the battery capacity, in Ah, and I represents the current drain, in A, by replacing

values it is possible to determine that it is delivered to the pacemaker a current drain of about 10 µA.

Also, from [56] the power consumption value of a pacemaker is between 10-40 µW.

55

3.2.2 Inductor Value

Regarding the converters’ set behaviour, it is already known from the conclusions of Section 3.1.2

that to guarantee the necessary voltage amount (in this case, at least, 2.8 V) when readjusting the input

sources, the system has to be dimensioned to operate in CCM for all the converters’ duty cycle. This

way, for a two converters’ example, taking Equation 3.12 and considering a duty cycle of 25% and 50%,

it is possible to determine the minimum value inductance for which the system behaves in CCM,

Lmin >Dmin(1−Dmin)2

2fswR = 0.18 mH. (3.47)

However, as concluded on Section 3.1.2.D, in order to guarantee the converters behaviour when multiple

converters are operating, the value of the inductance has to be higher than the CCM limit. This way, a

value of 0.68 mH was considered,

3.2.3 Output Capacitor

Concerning the output capacitor, from Equation 3.10 it is possible to determine the minimum value

of its capacitance and considering 50 mV as ∆VO,

Cmin =IOMAX

DMAX

fsw∆VO= 0.36 µF. (3.48)

The parameters for the CCM operation system sizing are presented in Table 3.15.

Table 3.15: Converters’ sizing parameters for CCM operation.

Duty cycle, D1 Duty cycle, D2 Load, RL Switching Frequency, fsw Inductance Value, L Output capacitor, Cout

25% 50% 500 Ω 200 kHz 0.68 mH 0.36 µF

3.2.4 Input Voltage Sources

Regarding the input voltage, taking Equation 3.8, to achieve this value, knowing that the converters

have to be operating at CCM and with a considered 3 V output voltage goal, the input voltages for

converters 1 and 2, if they were working alone, have to be 2.25 V and 1.5 V, respectively.

56

For the simultaneous operation, System of Equations 3.25 and its solution, System of Equations

3.39, have to be considered. Replacing the corresponding values, the obtained solution is

VO = 3 VII1 = 5.21 mAIOff = 0.621 mAII2 = 2.76 mA

. (3.49)

The previous results conclude that the input voltage values considered are valid to use for obtaining an

output voltage of 3 V.

The input voltage is intended to be collected from two different energy harvesting methods with

eventually their corresponding voltage elevator circuit. The chosen methods were the piezoelectric and

electrostatic energy harvesters which are able to generate a output voltage of 1.58 V and 2.28 V, re-

spectively. A more detailed analysis of these techniques is presented on Appendix A .

3.2.5 Schottky Diodes and MOSFET

Since it is intended to simulate the system as realistic as possible, the components that have been

considered as ideals until now, namely switches and diodes, have to be substituted for real components.

This way, concerning the diode selection some parameters have to be taken into account, such as the

reverse voltage, current rating and diode forward voltage. In order to get a voltage drop as low as

possible, a Schottky diode has been chosen, namely the 1n5817 one. As referenced in the datasheet

[57], this diode provides a 0.21 V forward voltage at 80 mA and with on resistance of 0.82 Ω, which

results a diode threshold voltage of 0.144 V. The diodes’ non ideality will result in a lower output voltage

value comparing to the projected one. Regarding the switches, they can be substituted by MOSFETs,

namely the Si8424DB one. This MOSFET has an anti-parallel body diode which may allow current to

flow unintentionally and has a low gate-source voltage value (1.2 V) and a drain-source voltage value up

to 8 V, as reported in the datasheet [58].

3.2.6 Duty Cycle Generator

In order to generate the different duty cycles, it is intended to implement a LMC555 CMOS Timer [59]

in each converter in order to generate the supposed duty cycle values. A more detailed explanation of

this device is done in Appendix B.

57

3.2.7 Zener Diode

As the principle of this project is to set a fixed Direct Current (DC) output voltage with value of 3 V, it

is needed to force down any higher peak voltage to maintain this value. To do that, a zener diode with a

4 V zener voltage is introduced and placed in anti-parallel with the load.

The zener diode is designed to have low and specified reverse breakdown voltage, which uses

reverse voltage when applied to it. When a zener diode is biased in forward direction (from anode to

cathode), it behaves like a normal diode, conducting the current. When reverse biased (from cathode to

anode), unlike usual diodes that block any flow of current, the zener diode starts to conduct in reverse

direction. The among of reverse voltage that drops through the zener diode increases to a maximum

value and once achieved, the voltage remains constant over a wide range of reverse currents [60]. The

voltage point at which the voltage across the zener diode becomes stable is called the “zener voltage”,

VZ , as shown in Figure 3.21.

Figure 3.21: Zener diode I-V characteristics [60].

A possible zener diode to implement is the 1n4731A [61].

Due to the increase of current that flows trough the zener diode in inverse conduction mode, a 0.3 Ω

resistor has been placed in series with the capacitor to decrease the amount of current.

58

3.2.8 Dimensioned Circuit

The dimensioned circuit is presented in Figure 3.22 and the correspondent components values are

presented in Table 3.16.

Figure 3.22: Multiple converters topology sized circuit.

Table 3.16: Parameters for CCM-CCM converters sizing example.


Input Voltage, V I VI1=2.25 V VI2 =1.5 VDuty Cycle, D D1=25% D2=50%Inductance, L L1=0.68 mH L2=0.68 mH

Inductor Average Inductance, IL IL1=5.75 mA IL2=2.57 mASchottky Diode Threshold Voltage, VDth

VD1Th=0.144 V VD2Th

=0.144 V

Commuting Frequency, fsw 200 kHzCapacitance, C 0.36 µF

Capacitor Resistor, RC 0.3 ΩLoad, RL 500 Ω

Zener Diode Voltage, VDZ4 V

The correspondent converters set behaviour has been simulated and the results are shown in Figure

3.23. Also, some resulting CCM boost converter typical waveforms are presented in Figure 3.24.

59

Figure 3.23: Sized circuit behaviour.

(a) Time diagrams of voltage at the switch, VS1 , and currentat inductor, IL1

, diode, ID1, and switch, IS1

, of converter 1.(b) Time diagrams of voltage at the switch, VS2 , and currentat inductor, IL2

, diode, ID2, and switch, IS2

, of converter 2.

(c) Time diagrams of current at the capacitor, IC , and in theoutput, IO as well as the output voltage, VO , of the set ofconverters.

Figure 3.24: Time diagrams of converter 1, 2 and converters’ set.

60

As expected, the behaviour corresponds to one reported in conclusions of Section 3.1.2: the higher

average inductor current source takes the command of the output voltage, behaving as the master and

the slave is at the limit of his CCM operation. Another conclusion that must be taken into account is that

the previously presented circuit sizing has been succeeded, at least, in respect to the output voltage,

which is less than the projected value (3 V), due to the 0.144 V voltage drop at the diode.

Concerning the true output power of the converters’ set, it is possible to conclude that it is about 16.4

mW, since

P =V 2

RL=

2.862

500= 16.4mW. (3.50)

61

4Circuit Performance in Case of Failure

Contents

4.1 Two Converters Failure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

4.2 Three Converters Failure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70

63

So far, the proposed circuit has been dimensioned in order to obtain a certain amount of voltage

when the input sources are working. As the main goal of this work is to design a security circuit, capable

of readjusting its input sources in case of a failure, it is needed to simulate their failures in order to

confirm the desired behaviour.

This chapter presents the previously sized circuit performance tests through simulation results per-

formed using a simulation software, called PSIM, provided and developed by Powersim.

4.1 Two Converters Failure

Concerning the proposed system, when all human body motion harvesters are working and supplying

the chosen implantable medical device for this analysis, a pacemaker, it has already been seen that the

behaviour of the system can be easily described making use of System of Equations 3.26. The worrying

situation is the system’s response in case of any input source failure. Ideally, the system should be

capable to readjust its input in such a way that a working input source ensures the desired behaviour

and the required output voltage.

If an input voltage source fails, any current will flow through its correspondent inductor and therefore

converter, so the output voltage of that converter will be 0 V. Considering System of Equations 3.26, if

a failure is simulated it is possible to induce that the output voltage will be assured only by converter 2,

resulting in an System of Equations as

VO2

R= VI2II2

II2 =1

2

VI2L2

D2T

VO =VI2

1−D

. (4.1)

As the output voltage and load is supposed to remain the same, it will be the average inductor current

of L2 that increases.

In order to test if the system corresponds to the previous described expected behavior, a piecewise

linear voltage source is introduced and placed as VI1 and VI2 , respectively, as shown in Figure 4.1.

65

Figure 4.1: Proposed circuit with piecewise linear voltage sources as input sources for testing purpose.

A piecewise linear voltage source is primarily used to generate a signal as a discrete function of time.

It is parameterized by a dataset of time and value pairs, and its primary application is in providing the

ability to quickly generate an arbitrary signal. However, these sources are also used to generate voltage

or current waveforms by requiring discrete time and value pairs. It has been chosen to use this type of

voltage source in place of a square wave voltage source, for example, due to the possibility of simulating

smother transitions instead of instantaneous ones, since no failure leads to an instantaneous transition

in reality. These transitions were set to take 10 ms each.

This way, using the circuit sizing parameters presented in Table 3.16 and implementing the piecewise

linear voltage generator in each input, with VI1= 2.25 V and VI2= 1.5 V amplitudes and an operating

frequency of 5 Hz, it is possible to analyze the system’s behaviour and test the viability of this solution.

Regarding to the expected behaviour, concerning the output voltage, it is expected to see a DC value

of 3 V with possible voltage breaks at the moment of input sources’ readjustment. This means that, when

both sources are working, the behaviour has to correspond to the one analyzed in Section 3.1.2.A, since

both converters are operating in CCM. It is also expected that the upper converter (converter 1) begins

to behave as the master, once it has a higher average inductor current value, and the lower converter

to behave as slave. When the master fails, its inductor current reaches 0 A after 10 ms, the converter

2 is expected to start behaving as master (since it is the only working converter) and to increase its

average inductor current, in order to maintain the output voltage value. It is in this transition that may

66

occur momentary voltage breakdowns, which are expected to return to the supposed value passed a

few ms.

The simulation results are presented in Figure 4.2.

Figure 4.2: Time diagram of output voltage, VO, and inductors current when a failure occurs in a two converters’system.

Analyzing these results, it is possible to conclude that the previously described behaviour is assured

and the stability of nearly 3 V output voltage is guaranteed.

4.1.1 Load Impedance Variation

It is already known that the pacemaker device has variable impedance values, depending on the

device power consumption which varies with the necessary pacing. When the impedance increases,

pacemaker’s required current decreases and when the impedance decreases to the minimum reported

value (200 Ω for the Vitatron Single Chamber Pacemaker), the required current achieve its higher value.

The proposed solution has to guarantee the delivery , at least, of this maximum required current to

the pacemaker.

Since the 2.8 V Vitatron pacemaker’s lithium-iodine battery delivers a maximum average current of

IMAX =V

Rmin=

2.8V

200Ω= 14 mA, (4.2)

67

the proposed system also has to deliver, at least, this amount when has a 200 Ω impedance as load.

In order to test this variation, a 335 Ω resistor, RT , is placed in parallel with the load and in series

with a switch. This switch will be turned off in the first 50% of the switching period and turned on in the

remaining time fraction of the switching period. When the switch is off, the load impedance value is 500

Ω and when the switch is on, the resistors’ parallel equivalent of the load impedance is about 200 Ω, as

shown in Figure 4.3.

Figure 4.3: Proposed circuit for impedance variation for testing purpose.

With the previously circuit sizing, it is expected that the average current when applied a impedance

of 200 Ω at 3 V to be

IMAX =V

Rmin=

3V

200Ω= 15 mA. (4.3)

The simulation results of this situation is presented in Figure 4.4.

These results correspond to the previously reported expected behaviour, which means that the sys-

tem is able to deliver the maximum amount of current that the pacemaker may require.

Concerning the pacemaker maximum impedance, which happens when it requires the minimum

amount of current, this system has as limiting factor the CCM mode limit condition. This means that the

maximum theoretical impedance value for which the system guarantees its CCM operation is

RMAX =2fsw

Dmin(1−Dmin)2L =

2× 200× 103

0.25(1− 0.25)2× 0.68× 10(−3) = 1.93 kΩ. (4.4)

68

Figure 4.4: Time diagram of output voltage, VO and inductors and output current when decreasing the impedancein a two converters’ system.

However, the maximum load value simulated for the converters to behave in CCM limit is 1.1k Ω, as

shown in Figure 4.5. For this load value, the average output voltage value is decreasing.

Figure 4.5: Time diagram of output voltage, VO and inductors and output current when increasing the impedancein a two converters’ system.

69

4.2 Three Converters Failure

It has been already confirmed that this solution is able to guarantee a specific output voltage when

an input source fails, in a two convert system. However, the more input sources alternatives existence,

the more reliable this system gets.

In order to predict this system configuration expected behaviour, it is needed to take into account the

System of Equations 3.25 and adjust it to a three input sources on. The result is

V 2O

R= VI1II1 + VI2II2 + VI3II3

II1 =1

2

VI1L1

D1D1T +1

2

VI1 − VOL1

(1−D1)(1−D1)T + IOff =1

2

VI1L1

D1T + IOff

II2 =1

2

∆iL2

=1

2

VI2L2

D2T

II3 =1

2

∆iL3

=1

2

VI3L3

D3T

VO =VI1

1−D1

. (4.5)

In this example, the master slave behaviour is also valid, but in this case two slaves exists and the

master converter will also be the one that has higher average inductor current.

If the failure scenario is considered and the master fails, the system’s behaviour will result in the two

converters behaviour, reported in Section 3.1.2.A.

This way, a third input source and its correspondent converter are added to the previously dimen-

sioned circuit. The input voltage source is placed for a third piecewise linear voltage source for testing

purposes, as presented in Figure 4.6.

70

Figure 4.6: Proposed circuit with piecewise linear voltage sources as input sources for testing purpose in a threeconverters example.

This way, using the circuit sizing parameters presented in Table 3.16 and implementing the piecewise

linear voltage generator in the new input, with VI3= 0.75 V amplitude, it is possible to analyze the system’s

behaviour and test the viability of this solution.

Concerning the new input voltage source and the correspondent converter, from previous analyzes

it is known that this converter, if sized to behave in CCM operation, when working simultaneously with

the remaining converters, will maintain its operation mode. This way, this converter has the same sizing

as the remaining ones, but with VI3 = 0.75 V and a duty cycle of 75 %, in order to set the output voltage

as 3 V.

Table 4.1 presents the complete system sizing for a three converters system.

71

Table 4.1: Parameters for a three converters sizing example.

Converter 1 Converter 2 Converter 3

Input Voltage, V I VI1=2.25 V VI2 =1.5 V VI3 = 0.75 vDuty Cycle, D D1=25% D2=50% D3=75%Inductance, L L1=0.68 mH L2=0.68 mH L3=0.68 mH

Schottky Diode Threshold Voltage, VDthVD1Th

=0.144 V VD2Th=0.144 V VD3Th

=0.144 V

Commuting Frequency, fsw 200 kHzCapacitance, C 0.36 µF

Capacitance Resistor, RC 0.3 ΩLoad, RL 500 Ω

Zener Diode Voltage, VDZ4 V

The theoretical values for the inductors currents (and input currents) when the three input sources

are working are the solutions of System of Equations 4.5 with the previous determined values, which is

VO = 3 VII1 = 5.56 mAII2 = 2.76 mAII3 = 1.93 mA

. (4.6)

The correspondent converters set behaviour has been simulated and the results are shown in Figure

4.7.

Figure 4.7: Sized circuit behaviour.

In respect to the expected simulation results in case of any input voltage source failure, concerning

the output voltage, it is expected to see a DC value of 3 V with possible voltage breakdowns at the

72

moment of input sources’ readjustment. This means that, when the three sources are working, the be-

haviour must be similar to the one analyzed in Section 3.1.2.A, since the three converters are operating

in CCM. It is also expected that, when all input sources are working, converter 1 begins to behave as the

master, once it has a higher average inductor current value comparing to the others two, which means

that converters 2 and 3 are expected to behave as slaves. When the master fails, its inductor current

reaches 0 A after 10 ms and the converter that has higher average inductor current (in this case, con-

verter 2) is expected to start behaving as master and converter 3 as slave, maintaining the coherence

with the previous analyzed behaviour. Converter 2 has to increase its average inductor current value, in

order to maintain the output voltage value and converter 3 stays at the CCM limit, concerning its inductor

current. Again, it is at this transitions that may occur momentary voltage breakdowns which returns to

the supposed value passed a few ms.

The input sources have been simulated to fail sequentially: first fails the converter 1 (the supposed

master), with converter 2 as the master and converter 3 as slave, followed by the failure of converter 2

(the master since converter 1 is down) with converter 3 as slave and finally the failure of converters 1

and 2, with converter 3 as the only working input source and behaving as master.

The simulation results are presented in Figure 4.8.

Figure 4.8: Time diagram of output voltage, VO, and inductors current when a failure occurs in a three converters’system.

Analyzing the results, it is possible to conclude that the previously described behaviour is verified

and the stability of nearly 3 V output voltage is guaranteed, for every failure stage. Even when the

lowest converter’s average inductor current is the master, it increases its value in order to guarantee the

required output voltage.

73

5Conclusions

75

Conclusions

Taking advantage of the energy harvesting devices, such as piezoelectric or electrostatic generators,

which gather the best characteristics among other ones, and combining them as multiple input sources

may be a reliable solution of powering IMDs. Although most of these devices are usually low powered,

it is needed to boost the generated voltage, using voltage elevation circuits for this purpose. However,

when using these generators, their continuous working status cannot always be assured.

In this work, a system capable of conditioning the collected voltage from energy harvesters to power

up an IMD is proposed. This system that, being very simple, besides elevating the several sources

voltages, automatically guarantees the existence of a working input power source has been developed

and optimized for a real pacemaker’s example. To achieve that, a detailed study about simultaneous

behaviour of multiple DC-DC converters has been done in order to achieve the best sizing for this

project goal. Ultimately, it has been proved that this system is capable of self-readjusting the input

voltage source into a permanent working one without compromising the circuit performance and fulfilling

the energy requirements of the considered pacemaker.

Future work can be directed to a better simulation of the scavenging sensors or devices, in order to

accomplish a more global and realistic simulation for the prototype. Also, an integrated prototype should

be considered because it is the usual technology used in implants and so a more accurate behaviour

could be verified.

77

Bibliography

[1] A. R. El-Sayed, K. Tai, M. Biglarbegian, and S. Mahmud, “A survey on recent energy harvesting

mechanisms,” Canadian Conference on Electrical and Computer Engineering, vol. 2016-Octob, pp.

0–4, 2016.

[2] M. A. Hannan, S. Mutashar, S. A. Samad, and A. Hussain, “Energy harvesting for the implantable

biomedical devices: issues and challenges,” BioMedical Engineering OnLine, vol. 13, no. 1,

p. 79, 2014. [Online]. Available: http://biomedical-engineering-online.biomedcentral.com/articles/

10.1186/1475-925X-13-79

[3] X. Wei and J. Liu, “Power sources and electrical recharging strategies for implantable medical

devices,” Frontiers of Energy and Power Engineering in China, vol. 2, no. 1, pp. 1–13, 2008.

[4] A. B. Amar, A. B. Kouki, and H. Cao, “Power approaches for implantable medical devices,” Sensors

(Switzerland), vol. 15, no. 11, pp. 28 889–28 914, 2015.

[5] N. Mano, F. Mao, and A. Heller, “Characteristics of a miniature compartment-less glucose-O2 bio-

fuel cell and its operation in a living plant,” Journal of the American Chemical Society, vol. 125,

no. 21, pp. 6588–6594, 2003.

[6] A. Cadei, A. Dionisi, E. Sardini, and M. Serpelloni, “Kinetic and thermal energy harvesters for im-

plantable medical devices and biomedical autonomous sensors,” Measurement Science and Tech-

nology, vol. 25, no. 1, 2014.

[7] D. M. Rowe, “CRC Handbook of Thermoelectrics,” New York, vol. 16, no. 1-4, pp. 1251–1256,

1995. [Online]. Available: http://www.crcnetbase.com/doi/book/10.1201/9781420049718

[8] Y. Yang, X.-J. Wei, and J. Liu, “Suitability of a thermoelectric power generator for implantable

medical electronic devices,” Journal of Physics D: Applied Physics, vol. 40, no. 18, pp. 5790–

5800, 2007. [Online]. Available: http://stacks.iop.org/0022-3727/40/i=18/a=042?key=crossref.

3be06556157bb503f0c2fdf1c77a3403

79

http://biomedical-engineering-online.biomedcentral.com/articles/10.1186/1475-925X-13-79

http://biomedical-engineering-online.biomedcentral.com/articles/10.1186/1475-925X-13-79

http://www.crcnetbase.com/doi/book/10.1201/9781420049718

http://stacks.iop.org/0022-3727/40/i=18/a=042?key=crossref.3be06556157bb503f0c2fdf1c77a3403

http://stacks.iop.org/0022-3727/40/i=18/a=042?key=crossref.3be06556157bb503f0c2fdf1c77a3403

[9] J. A. Nagel, I. Sieber, U. Gengenbach, H. Guth, G. Bretthauer, and R. F. Guthoff, “Investigation

of a thermoelectric power supply for the artificial accommodation system,” 2010 3rd International

Symposium on Applied Sciences in Biomedical and Communication Technologies, ISABEL 2010,

pp. 0–4, 2010.

[10] J. Paulo and P. D. Gaspar, “Review and future trend of energy harvesting methods for portable

medical devices,” WCE 2010 - World Congress on Engineering 2010, vol. 2, pp. 909–914,

2010. [Online]. Available: http://www.scopus.com/inward/record.url?eid=2-s2.0-79959817858&

partnerID=40&md5=d2a5fad5144cd3b3f3805a7aad86cddb

[11] M. T. D. D. P. D. W. M. L. R. L. T. H. Irani, Afraaz; Bianco, “Energy Generating Systems for Implanted

Medical Devices,” 2009.

[12] H. Goto, T. Sugiura, Y. Harada, and T. Kazui, “Feasibility of using the automatic

generating system for quartz watches as a leadless pacemaker power source,” Medical &

biological engineering & computing, vol. 37, no. I, pp. 377–380, 1999. [Online]. Available:

http://www.ncbi.nlm.nih.gov/pubmed/10505390

[13] P. Roberts, G. Stanley, and J. M. Morgan, “Harvesting the Energy of Cardiac Motion to Power a

Pacemaker,” Circulation, vol. 118, no. Suppl 18, pp. S 679 LP – S 680, 10 2008.

[14] V. Luciano, E. Sardini, and M. Serpelloni, “An electromechanical generator implanted in human

total knee prosthesis,” Lecture Notes in Electrical Engineering, vol. 162 LNEE, pp. 25–30, 2014.

[15] V. Luciano, E. Sardini, M. Serpelloni, and G. Baronio, “An energy harvesting converter to power

sensorized total human knee prosthesis,” Measurement Science and Technology, vol. 25, no. 2,

2014.

[16] A. Nasiri, S. A. Zabalawi, and D. C. Jeutter, “A linear permanent magnet generator for powering

implanted electronic devices,” IEEE Transactions on Power Electronics, vol. 26, no. 1, pp. 192–199,

2011.

[17] R. Morais, N. M. Silva, P. M. Santos, C. M. Frias, J. A. Ferreira, A. M. Ramos, J. A. Simoes,

J. M. Baptista, and M. C. Reis, “Double permanent magnet vibration power generator for smart

hip prosthesis,” Sensors and Actuators, A: Physical, vol. 172, no. 1, pp. 259–268, 2011. [Online].

Available: http://dx.doi.org/10.1016/j.sna.2011.04.001

[18] M. Miyazaki, H. Tanaka, G. Ono, T. Nagano, N. Ohkubo, T. Kawahara, and K. Yano, “Electric-Energy

Generation Using Variable-Capacitive Resonator for Power-Free LSI: Efficiency Analysis and Fun-

damental Experiment,” Proceedings of the International Symposium on Low Power Electronics and

Design, pp. 193–198, 2003.

80

http://www.scopus.com/inward/record.url?eid=2-s2.0-79959817858&partnerID=40&md5=d2a5fad5144cd3b3f3805a7aad86cddb

http://www.scopus.com/inward/record.url?eid=2-s2.0-79959817858&partnerID=40&md5=d2a5fad5144cd3b3f3805a7aad86cddb

http://www.ncbi.nlm.nih.gov/pubmed/10505390

http://dx.doi.org/10.1016/j.sna.2011.04.001

[19] P. Miao, a. S. Holmes, E. M. Yeatman, and T. C. Green, “Micro-Machined Variable Capacitors for

Power Generation,” Electrostatics, no. 178, pp. 53–58, 2003.

[20] H. Basaeri, D. B. Christensen, and S. Roundy, “A review of acoustic power trans-

fer for bio-medical implants,” Smart Materials and Structures, vol. 25, no. 12, p.

123001, 2016. [Online]. Available: http://stacks.iop.org/0964-1726/25/i=12/a=123001?key=

crossref.919b23d30da8062ea375b8d734bcadd6

[21] R. Tashiro, N. Kabei, H. Kotera, K. Katayama, Y. Ishizuka, F. Tsuboi, and K. Tsuchiya, “Develop-

ment of an Electrostatic Generator that Harnesses the Motion of a Living Body. Generation Using

Heartbeat.” Transactions of the Japan Society of Mechanical Engineers Series C, vol. 67, no. 659,

pp. 2307–2313, 2001.

[22] R. Tashiro, N. Kabei, K. Katayama, F. Tsuboi, and K. Tsuchiya, “Development of an electrostatic

generator for a cardiac pacemaker that harnesses the ventricular wall motion,” Journal of Artificial

Organs, vol. 5, no. 4, pp. 239–245, 2002.

[23] P. Miao, P. D. Mitcheson, A. S. Holmes, E. M. Yeatman, T. C. Green, and B. H. Stark, “Mems inertial

power generators for biomedical applications,” Microsystem Technologies, vol. 12, no. 10-11, pp.

1079–1083, 2006.

[24] S. P. S. Platt, S. F. S. Farritor, and H. H. H. Haider, “On low-frequency electric power generation

with PZT ceramics,” IEEE/ASME Transactions on Mechatronics, vol. 10, no. 2, pp. 240–252, 2005.

[25] S. R. Platt, S. Farritor, K. Garvin, and H. Haider, “The use of piezoelectric ceramics for electric

power generation within orthopedic implants,” IEEE/ASME Transactions on Mechatronics, vol. 10,

no. 4, pp. 455–461, 2005.

[26] S. Almouahed, M. Gouriou, C. Hamitouche, E. Stindel, and C. Roux, “The use of piezoceramics as

electrical energy harvesters within instrumented knee implant during walking,” IEEE/ASME Trans-

actions on Mechatronics, vol. 16, no. 5, pp. 799–807, 2011.

[27] ——, “Design and evaluation of instrumented smart knee implant,” IEEE Transactions on Biomedi-

cal Engineering, vol. 58, no. 4, pp. 971–982, 2011.

[28] D. K. Cheng, “Field and Wave Electromagnetics,” pp. 27–28, 1986.

[29] B. E. Lewandowski, K. L. Kilgore, and K. J. Gustafson, “Design considerations for an implantable,

muscle powered piezoelectric system for generating electrical power,” Annals of Biomedical Engi-

neering, vol. 35, no. 4, pp. 631–641, 2007.

81

http://stacks.iop.org/0964-1726/25/i=12/a=123001?key=crossref.919b23d30da8062ea375b8d734bcadd6

http://stacks.iop.org/0964-1726/25/i=12/a=123001?key=crossref.919b23d30da8062ea375b8d734bcadd6

[30] L. Zhou, A. C. Abraham, S. Y. Tang, and S. Chakrabartty, “Approaching the limits of piezoelectricity

driven hot-electron injection for self-powered in vivo monitoring of micro-strain variations,” Proceed-

ings - IEEE International Symposium on Circuits and Systems, vol. 2016-July, pp. 1810–1813,

2016.

[31] C. Huang, N. Lajnef, and S. Chakrabartty, “Calibration and characterization of self-powered floating-

gate usage monitor with single electron per second operational limit,” IEEE Transactions on Circuits

and Systems I: Regular Papers, vol. 57, no. 3, pp. 556–568, 2010.

[32] M. Deterre, B. Boutaud, R. Dalmolin, S. Boisseau, J.-J. Chaillout, E. Lefeuvre, and E. Dufour-

Gergam, “Energy harvesting system for cardiac implant applications,” 2011 Symposium on Design,

Test, Integration & Packaging of MEMS/MOEMS (DTIP), no. May, pp. 387–391, 2011.

[33] M. Deterre, E. Lefeuvre, Y. Zhu, M. Woytasik, B. Boutaud, and R. D. Molin, “Micro blood pressure

energy harvester for intracardiac pacemaker,” Journal of Microelectromechanical Systems, vol. 23,

no. 3, pp. 651–660, 2014.

[34] Y. Wang, Y. Yang, and Z. L. Wang, “Triboelectric nanogenerators as flexible power

sources,” npj Flexible Electronics, vol. 1, no. 1, p. 10, 2017. [Online]. Available: http:

//www.nature.com/articles/s41528-017-0007-8

[35] Y. Zhu, B. Yang, J. Liu, X. Wang, L. Wang, X. Chen, and C. Yang, “A flexible and

biocompatible triboelectric nanogenerator with tunable internal resistance for powering wearable

devices,” Scientific Reports, vol. 6, no. December 2015, pp. 1–10, 2016. [Online]. Available:

http://dx.doi.org/10.1038/srep22233

[36] T. C. Hou, Y. Yang, H. Zhang, J. Chen, L. J. Chen, and Z. Lin Wang, “Triboelectric nanogenerator

built inside shoe insole for harvesting walking energy,” Nano Energy, vol. 2, no. 5, pp. 856–862,

2013. [Online]. Available: http://dx.doi.org/10.1016/j.nanoen.2013.03.001

[37] C. Dagdeviren, Z. Li, and Z. L. Wang, “Energy Harvesting from the Animal/Human Body for Self-

Powered Electronics,” Annual Review of Biomedical Engineering, vol. 19, no. 1, pp. 85–108, 2017.

[Online]. Available: http://www.annualreviews.org/doi/10.1146/annurev-bioeng-071516-044517

[38] C.-m. Kyung, Nano Devices and Circuit Techniques for Low-Energy Applications and Energy

Harvesting, 2016. [Online]. Available: http://link.springer.com/10.1007/978-94-017-9990-4

[39] S. Ayazian and A. Hassibi, “Delivering optical power to subcutaneous implanted devices,” Pro-

ceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology

Society, EMBS, pp. 2874–2877, 2011.

82

http://www.nature.com/articles/s41528-017-0007-8

http://www.nature.com/articles/s41528-017-0007-8

http://dx.doi.org/10.1038/srep22233

http://dx.doi.org/10.1016/j.nanoen.2013.03.001

http://www.annualreviews.org/doi/10.1146/annurev-bioeng-071516-044517

http://link.springer.com/10.1007/978-94-017-9990-4

[40] Z. Chen, M. K. Law, P. I. Mak, and R. P. Martins, “A Single-Chip Solar Energy Harvesting IC Us-

ing Integrated Photodiodes for Biomedical Implant Applications,” IEEE Transactions on Biomedical

Circuits and Systems, vol. 11, no. 1, pp. 44–53, 2017.

[41] H. Chen, M. Liu, C. Jia, and Z. Wang, “Power harvesting using PZT ceramics embedded in ortho-

pedic implants,” IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, vol. 56,

no. 9, pp. 2010–2014, 2009.

[42] N. Bourgoine, “Harvest Energy from a Single Photovoltaic Cell,” LT Journal of Analog Innovation,

vol. 21, no. 1, pp. 1–6, 2011. [Online]. Available: www.linear.com

[43] Y. K. Ramadass and A. P. Chandrakasan, “A batteryless thermoelectric energy-harvesting inter-

face circuit with 35mV startup voltage,” Digest of Technical Papers - IEEE International Solid-State

Circuits Conference, vol. 53, pp. 486–487, 2010.

[44] I. Doms, P. Merken, R. Mertens, and C. Van Hoof, “Integrated capacitive power-management cir-

cuit for thermal harvesters with output Power 10 to 1000µW,” Digest of Technical Papers - IEEE

International Solid-State Circuits Conference, vol. 5, pp. 300–302, 2009.

[45] A. Richelli, S. Comensoli, Z. M. Kov, and S. Member, “A DC / DC boosting technique and power

management for ultra low voltage energy harvesting applications,” IEEE Transactions on Industrial

Electronics, vol. 59, no. 6, pp. 2701–2708, 2011.

[46] L. Fontela, “Conversor CC-CC elevador de tensao para aplicacoes de energy harvesting,” 2016.

[47] A. Khaligh, J. Cao, and Y. Lee, “A Multiple-Input DC–DC Converter Topology,” IEEE

Transactions on Power Electronics, vol. 24, no. 3, pp. 862–868, 2009. [Online]. Available:

http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=4760221

[48] Y. M. Chen, Y. C. Liu, and S. H. Lin, “Double-input PWM DC/DC converter for high-/low-voltage

sources,” IEEE Transactions on Industrial Electronics, vol. 53, no. 5, pp. 1538–1545, 2006.

[49] L. Kumar and S. Jain, “A multiple source DC/DC converter topology,” International Journal

of Electrical Power & Energy Systems, vol. 51, pp. 278–291, 2013. [Online]. Available:

http://linkinghub.elsevier.com/retrieve/pii/S014206151300080X

[50] A. Deihimi, M. E. Seyed Mahmoodieh, and R. Iravani, “A new multi-input step-up DC–DC converter

for hybrid energy systems,” Electric Power Systems Research, vol. 149, pp. 111–124, 2017.

[Online]. Available: http://dx.doi.org/10.1016/j.epsr.2017.04.017

[51] M. de Medeiros Silva, Circuitos com transistores bipolares e MOS, ser. Manuais universitarios.

Fundacao Calouste Gulbenkian. Servico de Educacao e Bolsas, 1999. [Online]. Available:

https://books.google.pt/books?id=3jeZAAAACAAJ

83

www.linear.com

http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=4760221

http://linkinghub.elsevier.com/retrieve/pii/S014206151300080X

http://dx.doi.org/10.1016/j.epsr.2017.04.017

https://books.google.pt/books?id=3jeZAAAACAAJ

[52] Vitatron, “E10S pacemaker specifications,” p. 4, 2010. [Online]. Available: http://www.vitatron.com/

downloads/e10-spec.pdf

[53] A. Augello, G. Della Chiara, V. M. Primiani, and F. Moglie, “Immunity tests of implantable cardiac

pacemaker against CW and pulsed ELF fields: Experimental and numerical results,” IEEE Trans-

actions on Electromagnetic Compatibility, vol. 48, no. 3, pp. 502–515, 2006.

[54] V. S. Mallela, V. Ilankumaran, and S. N. Rao, “Trends in cardiac pacemaker batteries,” Indian Pacing

and Electrophysiology Journal, vol. 4, no. 4, pp. 201–212, 2004.

[55] H. H. Klein and W. Knake, “Energy conserving programming of vvi pacemakers: A telemetry sup-

ported, longterm, follow up study,” Clinical Cardiology, vol. 13, no. 6, pp. 409–413, 1990.

[56] A. Haeberlin, A. Zurbuchen, S. Walpen, J. Schaerer, T. Niederhauser, C. Huber, H. Tanner,

H. Servatius, J. Seiler, H. Haeberlin, J. Fuhrer, and R. Vogel, “The first batteryless, solar-powered

cardiac pacemaker,” Heart Rhythm, vol. 12, no. 6, pp. 1317–1323, 6 2015. [Online]. Available:

https://www.sciencedirect.com/science/article/pii/S1547527115002520

[57] STMicroelectronics, “1n5817, 1n5818, 1n5819 Datasheet,” no. July, pp. 1–7, 2011.

[58] V. Electronics, “Si8424DB Datasheet,” pp. 1–8, 1847.

[59] T. Instruments, “LMC555 CMOS Timer Datasheet,” no. May, 2006.

[60] S. A. Sedra and C. K. Smith, “Devices and Basic Circuits,” p. 1450, 2009.

[61] V. Semiconductors, “Vishay Semiconductors Zener Diodes,” pp. 1–4, 2002.

84

http://www.vitatron.com/downloads/e10-spec.pdf

http://www.vitatron.com/downloads/e10-spec.pdf

https://www.sciencedirect.com/science/article/pii/S1547527115002520

AAppendix - Input Voltage Sources

Generators

85

As it has already been mentioned, powering up implantable medical devices by taking advantage

of energy harvesting devices, which convert energy collected from human body activities into electrical

energy, has been increasingly an alternative to fixed density and lifetime batteries. As concluded in

Subsection 2.1.2, among the most promising harvesters for collecting the largest amount of output

voltage, are the electrostatic and the piezoelectric generators.

Analyzing the input voltage sources that have been used in the converters sizing example, it is

possible to conclude that generators capable of deliver voltage amounts of 2.25 V and 1.5 V are needed.

However, it is also possible to use generators that aren’t capable of generating this amount of voltage

combined with the use of charge pump circuits.

Looking at the literature and the harvesters examples studied in Section 2.1.1, one example of a

piezoelectric generator that is suitable for the 1.5 V requirement is the inertial energy scavenger from

heart generated vibrational energy reported by Deterre et al. [32].

This device consists in an inertial energy harvester with viscous damping transduction, which me-

chanical system is presented in Figure A.1, and it is constituted by a proof mass and a frame. When

an external vibration is applied, the movable proof mass makes reciprocating motion and a consequent

displacement, which is related to the mean power of the transducing force.

Figure A.1: Mechanical system of inertial energy harvester with viscous damping transduction [32].

The output power depends on the heartbeats’ acceleration spectrum and the achievable power level

is about 100 µW for a system with dimensions 15 x 7 x 5 mm. If a 25 kΩ load is considered, this device

is able to generate a 1.58 V output voltage, which is suitable for the 1.5 V requirement.

Concerning the 2.25 V input source, a feasible option may be the electrostatic generator that harness

the ventricular wall motion, which as been proposed by Tashiro et al. [22].

87

This harvester is a Variable-Capacitance-Type Electrostatic (VCE) generator that harness ventricular

motion. This generator converts in vivo mechanical into electrical energy based on the following prin-

ciple: The VCES generator has a variable capacitor (VC) and has a capacitance change of between

certain minimum and maximum value. At first, when applying a DC voltage of V0 across the variable

capacitor at the maximum capacitance, the variable capacitor has an electric charge and an electro-

static energy. Next, when the capacitance is decreased to the minimum value by an external force while

preserving the electric charge Q, the voltage increases. This means that the stored electrostatic energy

becomes as large as before. That is, the energy increment is equal to the mechanical work done by the

external force.

Figure A.2 shows a VCE generating system. This system consists of an initial charge supply (ICS), a

variable capacitor (VC), whose capacitance can be changed by an external mechanical force, a capacitor

for energy storage (storage capacitor, C2), and two rectifying diodes (D2, D2). A battery is used only

one time to supply electric charge to the capacitor for the ICS at the very beginning of power generation.

There are two phases in one operating cycle. At first, when the voltage of the VC is low, the ICS

supplies electric charge to the VC in a counterclockwise direction. After that, as the capacitance C1 is

decreased gradually by the external force, voltage V1 increases. Then, the electric charge of the VC

flows into the storage capacitor in a clockwise direction. At the same time, the electric charge returns to

fill up the ICS. The amount of electric charge in the ICS remains constant in one operating cycle. Thus,

electrical energy in the storage capacitor increases, that is, mechanical work done by the external force

is converted to electrical energy. Therefore, electrical generation is performed.

Figure A.2: Variable-capacitance-type electrostatic generating system [22].

A mean generated electric power of 36 µW and a output voltage of approximately 2.4 V was obtained

with this device.

Since the converter’s input voltage source requirement is 2.25 V, the use of this device has proven to

be appropriated.

88

BAppendix - Duty Cycle Generator

89

In order to generate the desired duty cycle values to obtain the required output voltage, it is intended

to implement device capable of generate the dimentioned duty cycle values and at the sized frequency.

This way, it has been chosen to use a LMC555 CMOS Timer [59] in each converter in order to generate

the supposed duty cycle values.

This device is a low power and input voltage CMOS version of the conventional Timer 555 and it is

capable of generating all value of duty cycle between 50% and 100%, being powered by a minimum

supply voltage of 1.5 V. If necessary to have duty cycles with lower values than 50%, an inverter has to

be placed at the device output. Also, if any input source have less voltage than the required, a charge

pump has to be introduced in order to increase that value, at least, until 1.5 V.

As it has already been mentioned, one of the main applications of a 555 Timer is an oscillator that,

when working in astable mode (free running), is capable of generate a continuous stream of rectangular

pulses with a specific frequency, which duty cycle frequency values are set by two external resistors and

one capacitor.

When the circuit is connected as shown in Figure B.1, the device behaves as a variable duty cycle

oscillator. The external capacitor, C, charges through RA+RB and discharges through RB and betweenVS3

and2VS

3.

Figure B.1: LMC555 in Variable Duty Cycle Oscillator Configuration [59].

The charge and discharge times and the frequency are independent of the supply voltage. The

charge time is given by

t1 = 0.693(RA +RB)C, (B.1)

and the discharge time by

t2 = 0.693(RB)C. (B.2)

91

This way, the total period is the sum of the charge and discharge time, which is

T = t1 + t2 = 0.693(RA + 2RB)C. (B.3)

Thus, the frequency of oscillation is

fsw =1

T=

1.44

(RA + 2RB)C. (B.4)

The duty cycle value, as fraction of total period that the output is high, is given by

D =RA +RB

RA + 2RB. (B.5)

For the 50% example, the circuit configuration is presented in Figure B.2.

Figure B.2: LMC555 Configuration for 50% Duty Cycle [59].

In this case, the frequency of oscillation is given by

fsw =1

1.4RCC, (B.6)

and considering the 200 kHz value, the conductor and resistor values areC = 0.0001 µ FRC = 36 kΩ

. (B.7)

This way, knowing that the desired switching frequency is 200 kHz and the value of the duty cycle,

92

conjugating Equations B.4 and B.5, for the 75% value it will be necessary to have, for example,

C = 0.0001 µ FRA = 39 kΩ

RB = 18 kΩ

. (B.8)

Concerning the 25% duty cycle, in order to achieve this value, it is needed to place an inverter in the

output of the 555 timer.

For simulating these duty cycle value generators, a PSIM model of 555 Timer has been used with

the correspondent circuit configuration and the correspondent values of capacitor and resistors for each

desired duty cycle value.

For the 50% duty cycle, it is needed to implement the circuit configuration presented in Figure B.2

and use the already obtained values for RC and C. The circuit configuration and its correspondent

output are presented in Figure B.3.

(a) Circuit configuration for 50% duty cycle. (b) 50% duty cycle output.

Figure B.3: LMC 555 with 50% duty cycle.

For the 25% duty cycle, first it is needed to generate a 75% duty cycle and then invert the out-

put signal. Using the already determined resistor and capacitor value, the circuit configuration and its

correspondent output are presented in Figure B.4.

93



Once that the 75% duty cycle is obtained, it is necessary to invert its output signal. For that, a not

gate is added to the output signal, as shown in Figure B.5, which presents the circuit configuration and

its correspondent output.



94

energy harvesting for implantable medical devices€¦ · energy conditioning for implantable...

Documents