A USER PROGRAMMABLE BATTERY CHARGING SYSTEM

A Thesis

by

JUDY MARIAN AMANOR-BOADU

Submitted to the Office of Graduate Studies of

Texas A&M University
in partial fulfillment of the requirements for the degree of

MASTER OF SCIENCE

Approved by:

Chair of Committee, Edgar Sanchez-Sinencio

Co-Chair of Committee, Sam Palermo
Committee Members Weiping Shi
Hank Walker
Head of Department, Chanan Singh

May 2013

Major Subject: Electrical Engineering

Copyright 2013 Judy Marian Amanor-Boadu

 ABSTRACT

Rechargeable batteries are found in almost every battery powered application. Be it

portable, stationary or motive applications, these batteries go hand in hand with battery

charging systems. With energy harvesting being targeted in this day and age, high

energy density and longer lasting batteries with efficient charging systems are being

developed by companies and original equipment manufacturers. Whatever the

application may be, rechargeable batteries, which deliver power to a load or system,

have to be replenished or recharged once their energy is depleted. Battery charging

systems must perform this replenishment by using very fast and efficient methods to

extend battery life and to increase periods between charges. In this regard, they have to

be versatile, efficient and user programmable to increase their applications in numerous

battery powered systems. This is to reduce the cost of using different battery chargers for

different types of battery powered applications and also to provide the convenience of

rare battery replacement and extend the periods between charges.

This thesis proposes a user programmable charging system that can charge a Lithium ion

battery from three different input sources, i.e. a wall outlet, a universal serial bus (USB)

and an energy harvesting system. The proposed charging system consists of three main

building blocks, i.e. a pulse charger, a step down DC to DC converter and a switching

network system, to extend the number of applications it can be used for. The switching

network system is to allow charging of a battery via an energy harvesting system, while

the step down converter is used to provide an initial supply voltage to kick start the
ii
energy harvesting system. The pulse charger enables the battery to be charged from a

wall outlet or a USB network. It can also be reconfigured to charge a Nickel Metal

Hydride battery. The final design is implemented on an IBM 0.18µm process.

Experimental results verify the concept of the proposed charging system. The pulse

charger is able to be reconfigured as a trickle charger and a constant current charger to

charge a Li-ion battery and a Nickel Metal Hydride battery, respectively. The step down

converter has a maximum efficiency of 90% at an input voltage of 3V and the charging

of the battery via an energy harvesting system is also verified.

iii
 ACKNOWLEDGEMENTS

I would like to thank my committee chair and advisor, Dr. Sanchez-Sinencio for his

support, advice and guidance throughout my research study. I like to acknowledge my

committee members, Dr. Palermo, Dr. Shi and Dr. Walker, for serving on my defense

committee. I also express my gratitude to Dr. Zenon Medina-Cetina for taking time off

his busy schedule to substitute for one of my committee members on the day of my

defense.

I would also like to express thanks to Dake Tuli and Benjamin Sarpong of Texas

Instruments for their support during my years at Texas A & M University. I also extend

my gratitude to MOSIS (Metal Oxide Semiconductor Implementation Service) who

provided a means to fabricate my design through their educational program. Thanks also

goes to my friends, research colleagues and the department faculty and staff for making

my time at Texas A&M University a great experience.

Finally, my gratitude goes to my parents and brother for their encouragement, patience,

love and support.

iv
 NOMENCLATURE

AC Alternating current

DC Direct current

PMOS P-type metal-oxide-semiconductor

NMOS N-type metal-oxide-semiconductor

v
 TABLE OF CONTENTS

Page

ABSTRACT .......................................................................................................................ii

ACKNOWLEDGEMENTS .............................................................................................. iv

NOMENCLATURE ........................................................................................................... v

TABLE OF CONTENTS ..................................................................................................vi

LIST OF FIGURES ........................................................................................................ viii

LIST OF TABLES ......................................................................................................... xiii

1. INTRODUCTION .......................................................................................................... 1

1.1 Thesis organization .................................................................................................. 7

2. ENERGY STORAGE AND ENERGY HARVESTING ............................................... 9

2.1 Energy storage devices ............................................................................................. 9

2.2 Applications of batteries......................................................................................... 10
2.2.1 Portable applications ....................................................................................... 11
2.2.2 Stationary applications .................................................................................... 11
2.2.3 Motive applications ......................................................................................... 12
2.3 The battery.............................................................................................................. 12
2.3.1 Some commonly used battery terms ............................................................... 14
2.3.2 Battery classifications...................................................................................... 17
2.3.3 Major types of battery chemistry..................................................................... 18
2.4 Battery chargers...................................................................................................... 23
2.4.1 Types of battery charging modes .................................................................... 24
2.4.2 Li-ion battery chargers .................................................................................... 28
2.4.3 Li-ion battery charger design considerations .................................................. 34
2.5 Energy harvesting ................................................................................................... 36
2.5.1 Solar energy harvesting ................................................................................... 37
2.5.2 Thermal energy harvesting .............................................................................. 38
2.5.3 Vibration energy harvesting ............................................................................ 39
2.6 The battery and the supercapacitor......................................................................... 40
2.7 Overview of charging system ................................................................................. 42
2.7.1 Description of charging system block diagram ............................................... 43
2.7.2 Description of charging system flow process .................................................. 45
vi
3. PULSE CHARGER...................................................................................................... 47

3.1 Previous works ....................................................................................................... 47

3.2 Conceptual idea of proposed pulse charger............................................................ 56
3.2.1 Flow chart of proposed pulse charger ............................................................. 58
3.2.2 Description of block diagram .......................................................................... 60
3.3 Circuit level implementation .................................................................................. 61
3.3.1 Design of basic building blocks ...................................................................... 62
3.3.2 Design of trickle charging phase ..................................................................... 79
3.3.3 Design of fast charging phase ......................................................................... 84
3.3.4 Protection schemes .......................................................................................... 87
3.4 The pulse charger ................................................................................................... 98
3.4.1 Transformation of pulse charger to trickle charger ....................................... 103
3.4.2 Transformation of pulse charger to constant current charger ........................ 104

4. STEP DOWN DC-DC CONVERTER AND SWITCHING NETWORK

SYSTEM.................................................................................................................... 107

4.1 Step down DC-DC converter ............................................................................... 108

4.1.1 Switched capacitor converter ........................................................................ 110
4.2 Switching network system ................................................................................... 120
4.2.1 Implementation of the switching network system ......................................... 121

5. EXPERIMENTAL RESULTS ................................................................................... 129

5.1 Proposed pulse charger......................................................................................... 130

5.2 Switched capacitor converter ............................................................................... 134
5.3 Switching network system ................................................................................... 137
5.4 Comparison with previous works ......................................................................... 140

6. CONCLUSIONS ........................................................................................................ 143

REFERENCES ............................................................................................................... 144

vii
 LIST OF FIGURES

Page

Figure 1 A typical battery charging system..................................................................... 3

Figure 2 Proposed user programmable battery charging system. ................................... 6

Figure 3 Major applications of batteries. ...................................................................... 10

Figure 4 Chemical and electrical symbols of an electric cell. ....................................... 13

Figure 5 Battery feeding current (I) into a load............................................................. 14

Figure 6 Dependence of cycle life on DOD. ................................................................. 16

Figure 7 Comparison of battery chemistries in terms of energy density and weight. ... 23

Figure 8 Conceptual view detailing the main functions of a battery charger................ 24

Figure 9 Trickle charge profile. ..................................................................................... 25

Figure 10 Constant voltage charge profile. ..................................................................... 26

Figure 11 Constant current charge profile. ...................................................................... 27

Figure 12 Pulse charge profile. ....................................................................................... 28

Figure 13 A typical charge and discharge profile of a Li-ion battery. ............................ 29

Figure 14 Charging profile of a Li-ion battery. ............................................................... 30

Figure 15 Linear and switch mode charger concept. ...................................................... 32

Figure 16 Pulse charger concept. .................................................................................... 34

Figure 17 Conventional capacitor versus supercapacitor. ............................................... 41

Figure 18 Charging system with DC-DC converter and switching network. ................. 43

Figure 19 Flow chart showing the operation of the battery charging system. ................ 46

viii
Figure 20 Proposed block diagram of phase locked battery charger. ............................. 48

Figure 21 Implementation of VFPCS.............................................................................. 50

Figure 22 Block diagram of DVVPCS. ........................................................................... 51

Figure 23 LTC4052. ........................................................................................................ 52

Figure 24 LTC1730. ........................................................................................................ 53

Figure 25 MAX1879. ...................................................................................................... 54

Figure 26 Conceptual block diagram of proposed pulse charger. ................................... 57

Figure 27 Pulse charger flow chart. ................................................................................ 58

Figure 28 Block level diagram of pulse charger. ............................................................ 61

Figure 29 Regenerative comparator showing hysteresis. ................................................ 63

Figure 30 Open loop comparator. ................................................................................... 63

Figure 31 Advantage of hysteresis window. ................................................................... 65

Figure 32 Transistor level diagram of comparator. ......................................................... 66

Figure 33 Plot of various transistor parameters versus inversion level. .......................... 69

Figure 34 Gain and phase plot of amplifier stage in comparator. ................................... 71

Figure 35 Hysteresis of comparator. ............................................................................... 72

Figure 36 Hysteresis window of designed comparator. .................................................. 75

Figure 37 Conceptual idea of DCG. ................................................................................ 77

Figure 38 Positive edge triggered flip flop operation. .................................................... 78

Figure 39 Duty cycle generator. ...................................................................................... 78

Figure 40 Waveforms of duty cycle generator. ............................................................... 79

Figure 41 Trickle charging phase block diagram. ........................................................... 81

ix
Figure 42 Trickle current, ITC, versus resistance, RTC. ................................................ 82

Figure 43 Trickle charging phase operation signals. ....................................................... 83

Figure 44 Fast charging phase block diagram. ................................................................ 86

Figure 45 Fast charging phase operating signals. ........................................................... 87

Figure 46 Safe operating region of circuit and battery. ................................................... 88

Figure 47 Overcurrent protection circuit. ........................................................................ 89

Figure 48 Error amplifier, C2, with boosted output impedance. ..................................... 91

Figure 49 Temperature sense circuit. .............................................................................. 93

Figure 50 Output signal of temperature sense circuit. .................................................... 95

Figure 51 Conceptual idea of overvoltage protection. .................................................... 97

Figure 52 Overvoltage protection and termination. ........................................................ 97

Figure 53 Detailed block diagram of pulse charger. ....................................................... 99

Figure 54 Pulse charger in normal operation connection diagram. ............................... 101

Figure 55 Battery charging using pulse charger. ........................................................... 101

Figure 56 Waveforms showing operation of pulse charger. ......................................... 102

Figure 57 Operation of pulse charger as trickle charger. .............................................. 104

Figure 58 Operation of pulse charger as constant current charger. ............................... 105

Figure 59 Linear regulator (LDO) and switch inductor converter (buck). .................... 109

Figure 60 Switched capacitor converter. ....................................................................... 111

Figure 61 Operation of switched capacitor converter during different clock phases. ... 112

Figure 62 Transistor level implementation of switched capacitor converter. ............... 116

Figure 63 Operation of switched capacitor converter with Vbatt=3V at no load. .......... 117

x
Figure 64 Converter’s output voltage variation with load current. ............................... 118

Figure 65 Converter output variation with varying battery voltage. ............................. 119

Figure 66 Efficiency of switched capacitor converter at 3V input. .............................. 120

Figure 67 Flow chart showing conceptual idea of switching network system.............. 122

Figure 68 Conceptual block diagram of switching network system. ............................ 123

Figure 69 Clocked comparator. ..................................................................................... 123

Figure 70 Operation of clocked comparator. ................................................................ 125

Figure 71 Transistor level implementation of switching network system. ................... 126

Figure 72 Operation of switching network system. ...................................................... 128

Figure 73 Micrograph of user programmable battery charging system IC. .................. 129

Figure 74 Pictorial representation of experimental setup. ............................................. 130

Figure 75 Test bench to verify operation of proposed pulse charger. ........................... 131

Figure 76 Test result verifying operation of proposed pulse charger in normal mode. 132

Figure 77 Trickle mode charging of Li-ion cell using reconfigured pulse charger. ..... 133

Figure 78 Constant current charging using reconfigured pulse charger. ...................... 134

Figure 79 Schematic of switched capacitor test bench. ................................................ 135

Figure 80 Control signal for clock phases and output voltage ripple of converter. ...... 135

Figure 81 Converter's output with a 3V input and increased output capacitance. ........ 136

Figure 82 Output voltage variation versus input voltage variation. .............................. 136

Figure 83 Efficiency of converter. ................................................................................ 137

Figure 84 Schematic of switching network system test bench. ..................................... 138

Figure 85 Operation of switching network for supercapacitor and battery < 3V. ........ 139

xi
Figure 86 Operating signals of switching network system. .......................................... 140

xii
 LIST OF TABLES

Page

Table 1 Different battery chemistries ............................................................................. 19

Table 2 Comparison of types of Li-ion battery charger ................................................. 34

Table 3 Design parameters depending on multi cell or single cell applications ............ 36

Table 4 Energy sources and their estimated harvested power........................................ 40

Table 5 Comparison of previous works ......................................................................... 55

Table 6 Specifications of pulse charger ......................................................................... 61

Table 7 Comparator first stage transistor dimensions .................................................... 70

Table 8 CMOS Schmitt trigger transistor dimensions ................................................... 74

Table 9 Trickle currents and corresponding resistor values ........................................... 82

Table 10 Transistor dimensions of C2 ............................................................................. 92

Table 11 Pulse charger pin descriptions and maximum ratings ..................................... 100

Table 12 Specifications of the EHS ............................................................................... 107

Table 13 Comparison of different step down converters ............................................... 110

Table 14 Clocked comparator transistor dimensions ..................................................... 125

Table 15 Comparison of proposed charger with previous works ................................... 141

xiii
 1. INTRODUCTION

Most systems, from automobiles to smart phones, are powered by energy storage

devices. Energy storage device usage ranges from emergency power to portable devices,

which provide the convenience of movement without the impediment of wires. Energy

storage devices, such as batteries, have proved to be an important factor in the increase

and release of smart and high performance battery powered devices on the market today.

With smart computing and portable electronic devices blooming in the global market,

consumers are pushing for extended battery power usage and longer lasting batteries to

prevent frequent charging and replacement. Companies and original equipment

manufacturers (OEM) are therefore investing in smart battery chargers and battery

management systems (BMS) and are also looking for new and more efficient ways to

make consumers happy in that regard. Battery chargers and BMS continue to be very

important in battery powered systems since once the energy is depleted from the battery,

it has to be replenished for continuous use, and therefore the market for chargers and

BMS will continue to grow as long as there is an increase in demand for rechargeable

batteries. According to a recent report by Global Industry Analysts Inc. (GIA), the global

consumer batteries market is going to reach $55.4 billion by 2017 [1] and this is being

fueled by the rising popularity of portable electronic devices such as smart phones and

tablets.

In this day and age, where the world is moving towards green, renewable and sustainable

energy, energy storage devices are needed in the area of energy harvesting. Renewable
1
energy, which is environmentally friendly, can effectively be stored without any

environmental hazards for a time frame ranging from a few seconds to several months

and reused almost immediately. Renewable energy does not provide a steady stream of

power and hence, energy storage is needed for uninterrupted service. Whether macro

energy harvesting to feed into the grid or micro energy harvesting for portable

applications, the next fifty or more years is going see a lot of investment in renewable

energy, and there is therefore going to be a demand for better harvesting techniques and

energy storage devices. The advancement of smart power management systems will help

support the development of new energy storage technologies. Macro or micro energy

harvesting systems will definitely include a charger to replenish any energy lost from the

storage device.

Whatever the system may be, from energy harvesting to electric vehicles, battery

powered devices go hand in hand with battery chargers and BMS. There is a need to

therefore design efficient, smart and intelligent battery chargers for such systems. These

chargers should incorporate safety mechanisms to ensure safety of the consumer and

also prevent harm to the battery and prolong its life. The state, remaining capacity and

life cycle of a battery depends on the quality of charger used to charge it and therefore,

the selection of the appropriate battery charger for a specific battery is of great

importance when it comes to OEMs. This is to ensure that the life cycle of the battery is

extended to prevent frequent replacement.

2
A typical battery charging system for portable applications is shown in Fig. 1, which

takes power from an input source and replaces the energy depleted from the battery.

Traditionally, batteries are recharged from a wall adapter. With recent improvements in

technology, batteries can be recharged from a Universal Serial Bus (USB) or from an

energy harvesting system.

Input Voltage
Transformerless power supply

L
Rectifier

Rs
Class X2 Cap
N

Wall adapter
AC
rectifier
outlet

regulator

N
transformer Battery
charger +
Battery
USB -
Step up

Energy
harvester Step down

Sources

Figure 1 A typical battery charging system.

From a wall outlet, i.e. AC mains, a device known as a wall adapter is used to provide

the right input voltage for the battery charger. The standard voltage obtained from the

wall outlet is 110Vrms and cannot be directly used to charge a battery. This voltage is

stepped down using a transformer, rectified and regulated to achieve a good input

3
voltage ideal for the charging system. Another method to step down the mains voltage is

by using a reactance, which does not only drop the mains voltage but also limits the

inrush line current. This method is known as transformerless power supply [2]. It uses a

resistor and class X2 capacitor, which could have a voltage rating in the order of

kilovolts [3], to drop the mains voltage to smaller voltage which is then rectified for use

in portable applications. Though the transformerless power supply consumes much less

space and is cheaper than step down transformers, it is not used often because of their

non-isolation properties from the mains, which could lead to potential safety concerns.

The input voltage to the charger can also be provided by connecting it to a USB network.

The USB network consists of a USB socket and a USB cable [4]. The socket consists of

four pins and the cable is constructed using four 20 or 28 American wire gauge (AWG)

wires [5]. The USB socket pins include two inner pins for data transfer and two outer

pins for power supply. These two outer pins provide the charging voltage and current to

the battery charger. Likewise, the USB cable has two wires for power supply and two

wires for data transfer. The construction of the standard USB network provides a clean

and good voltage for a charging system in portable battery powered applications. The

USB outputs a voltage ranging from 4.75V to 5V, allowing a maximum charging current

of 0.5A [6].

Using an energy harvester source requires some type of step up or step down circuit

depending on the available power the harvester is able to produce. The step up or step

down circuit is needed to be able to achieve the right input voltage for the charging

system. Energy harvesters, such as solar cells, can output a high or low voltage
4
depending on the environmental conditions. In cases where the harvester produces a high

voltage, which cannot be applied directly to the charging system, a step down circuit,

such as a buck converter, is used to reduce the voltage to an ideal value. When the

harvester output voltage is low, such as the output power derived from radio frequency

emissions, a step up circuit, such as a boost converter or a charge pump, is needed to up-

convert the voltage to a value which is suitable to charge the battery. In some cases, the

boost or buck converter incorporates the battery charger.

Many battery charging systems use the wall outlet, USB or a combination of the two as

input sources. Energy harvesting systems incorporating chargers make use of only

energy harvesters as their input sources. Many battery chargers do not use a combination

of the three input sources. Also, they do not provide flexibility to users to change the

configuration of the chargers to meet a specific need, i.e. to charge different types of

batteries. In this thesis, the charging system designed seeks to combine all three inputs,

depending on which one is available. It also provides the user with options to change the

configuration of the charger to meet a specific need, i.e. the charger can be reconfigured

from its original design to behave like another type of charger. This can be a pulse

charger, a constant current charger or a trickle charger. This is done to increase the

number of applications which can use this system. The benefits of combining different

input sources is to extend battery power usage and enable charging under different

conditions, such as in applications where a wall outlet is not available, or to make use of

available green energy. It can therefore be used in a multi-purpose environment. In order

to do this, additional blocks are designed to support the battery charger: a step down DC
5
to DC converter to serve as a power supply for an energy harvesting system (EHS), and

a switching network system to enable the charging of the battery from all available input

sources. The conceptual idea for this proposed battery charging system is illustrated in

Fig. 2.

EHS

Charging from EHS

Power supply

AC

Step Down
Pulse Charger Converter
USB
+
Battery
-
Constant current
charger

Trickle charger

Figure 2 Proposed user programmable battery charging system.

When it comes to the load, the charging system designed is used to charge a battery

powering a sensor, preferably a humidity sensor. The system seeks to combine two

energy storage elements, i.e. battery and a supercapacitor, to meet the demands of the

load, i.e. sensor. A supercapacitor is utilized for short term peak power demand and a

battery for constant power demand. The main idea is to increase the life cycle of the

battery and to ensure longer periods between charges.

6
1.1 Thesis organization

The main objective of this thesis is to design a battery charging system capable of

supporting USB, AC-DC and energy harvesting applications. The charging system will

consist of a battery charger, a step down DC-DC converter and a switching network

system to extend the number of input voltage sources that can be used. The main

challenge in designing such systems is ensuring the battery charge method is ideal for

whatever type of battery used, and then ensuring that the charge profile includes

measures to prevent any potential harm to a user and extending the life cycle of the

battery. A charging system incorporating safety measures and good charging profile is

designed in this thesis.

Section 2 first introduces energy storage elements and energy harvesting, delving into

the different types of batteries, charge methods, battery chargers and energy harvesters.

It then provides an overview of the entire charging system to be designed.

Section 3 begins by discussing previous works in literature, and then a detailed design

procedure of the proposed charger. The conceptual idea and a description of the block

diagram are given. Transistor level design of each block is discussed in more detail and

then the section concludes with schematic simulation results to prove the design follows

the ideal charge profile of the battery used and the safety mechanisms employed in the

design work.

Section 4 introduces the two additional blocks, i.e. the step down DC to DC converter

and the switching network system. Different types of step down DC to DC converters
7
are described and the reason behind why a particular topology is chosen over others is

given. The design procedure of the converter is discussed and the simulation results to

prove the designed circuit works at various load conditions is presented. The detailed

process flow of how the switching network work is presented and the transistor level

implementation introduced. The section concludes with simulation results which prove

that the proposed switching network system enables charging of the battery from an

energy harvesting system.

Section 5 takes a look at experimental results showing how the designed system will

work under real life conditions. Comparison is made with previous works reported in

literature.

In section 6, conclusions are drawn from the results obtained and proposed

improvements are discussed.

8
 2. ENERGY STORAGE AND ENERGY HARVESTING

Most systems are powered by energy storage devices and most times these devices need

to be recharged to prevent the inconvenience of constant replacement and additional

cost. Likewise, energy harvesting goes hand in hand with energy storage and that implies

the use of energy storage devices such as batteries and supercapacitors. This section

looks at the different applications that make use of batteries, the battery, different battery

chemistries, battery charging and an introduction to energy harvesting. The section will

conclude with an overview of the entire charging system which includes the charger,

switching network system and step down DC to DC converter designed in this thesis.

2.1 Energy storage devices

Energy storage is basically the storage of energy in a device for later use. The uses of

this stored energy can be for emergency purposes, portable applications or for reliability

reasons. There are many forms of energy storage such as mechanical, thermal, chemical

and electrical [7]-[10]. For this thesis, electrical and electrochemical forms of energy

storage will be examined. Some major examples of energy storage devices of this form

are fuel cells, batteries and capacitors [10]. Fuel cells are generally not for portable

applications [11]. They generate electricity through an oxidation or reduction process

involving hydrogen and oxygen. A battery converts its internal stored chemical energy

into electrical energy while a capacitor stores energy in its electric field [12]. Batteries

and capacitors will be the main focus in this thesis.

9
2.2 Applications of batteries

The usage of energy storage devices, such as batteries, can be found everywhere, from

smart phones to smart grids [10], [13]-[15]. In this regard, battery usage can be

categorized into three major applications [15]-[17]: portable, stationary and motive. Fig.

3 illustrates a summary of the major battery powered applications.

Motive
applications

Batteries

Stationary Portable
applications applications

Figure 3 Major applications of batteries.

10
2.2.1 Portable applications

Portable applications include smart phones, Global Positioning System (GPS),

computers, digital cameras and Personal Digital Assistant (PDA). These applications are

fuelling the increase in the global market demand for consumer batteries. The medical

industry is another area where batteries are used often. This industry is making vast

advances in the area of using efficient and long lasting batteries in wearable medical

devices and portable medical equipment. For portable applications in the military [18],

the ground soldier ensemble is a typical example where the soldier has a wearable

computer, wireless transceivers, a head mounted display and a thermal weapon scope,

which all contain batteries. These enable the soldier to have decreased reaction time and

increased awareness during combat.

2.2.2 Stationary applications

These include standby power plants, which make use of maintenance free batteries, to

power electrical and electronic appliances in case of a power outage. With more and

more people moving towards cloud computing, there is a need for data storage servers to

have uninterruptible power supply to prevent any loss of data. This implies the need for

backup power supplied by batteries in case of a major power outage. The smart grid is

also an emerging area in the use of batteries for stationary applications [9]. Renewable

energy does not constantly provide a steady stream of power, so grid storage batteries

are needed to provide uninterruptible power to consumers. The smart grid maintains and

controls the electrical grid to ensure that it is balanced and reliable when connected to

11
energy harvesters. The smart grid includes energy storage, smart meters and efficient

distributed generation.

2.2.3 Motive applications

The electric vehicle is one major area in this type of application. Car manufacturers are

investing a lot in longer lasting and efficient batteries for electric vehicles as the world

moves towards green energy [19]-[20]. Electric vehicles are not only limited to cars but

include electric motorcycles, boats and airplanes. Marine transport [21]-[22] is also

another major area that makes use of batteries for on board lighting and motor start-up.

The usage of batteries in unmanned underwater vehicles (UUV), unmanned aerial

vehicles (UAV) and unmanned ground vehicles (UGV) [23] enables the military to

undertake dangerous missions without putting military personnel in danger.

2.3 The battery

Having reviewed some applications involving the use of batteries, the battery as an

energy storage device has to be defined. A battery is a collection of two or more electric

cells, which stores energy for later use. An electric cell is a device, which converts a

chemical reaction into electrical energy. A typical electric cell contains a cathode

(negative side), an anode (positive side) and the electrolyte. This is illustrated in Fig. 4.

12
 Load

- Electron
flow
+
- +

Cathode

Anode
- +
+
+
- -

Electrolyte

Electrical symbol
Chemical representation of
an electric cell

Figure 4 Chemical and electrical symbols of an electric cell.

When a load completes a connection between the anode and the cathode of a battery as

shown in Fig. 4, electrons are produced at the anode due to a chemical reaction between

the anode material and the electrolyte. These electrons are then absorbed via a similar

chemical reaction at the cathode. This process of electron flow is known as electricity.

A battery can be simply modeled as a voltage source in series with a resistance as seen in

Fig. 5. The resistance is due to the resistivity of electrodes and electrolyte [24]. The

resistance causes an internal voltage drop when the battery is delivering power to a load,

which causes the current and voltage available to the load to be reduced. Battery internal

resistance can increase due to aging or discharge. The values of the internal resistance

vary from milliohms to a few ohms. A battery with a high internal resistance has a lower

output current capability, while a battery with a lower internal resistance ensures that a
13
high current can be delivered when needed [25]. The relationship between the effective

battery voltage, VB, the ideal battery voltage and the internal resistance, r is shown in

equation 1.

+
Vb

Battery
VB

Load
r

Figure 5 Battery feeding current (I) into a load.

2.3.1 Some commonly used battery terms

 Battery capacity: The available electrical charge stored in a battery for transfer

when connected to a load. Battery capacity is usually rated in Ampere-hour (Ah).

For example, a 1.5V AA Energizer L91 Lithium battery [26] is rated at

3000mAh. This means that the battery can deliver 3000mA current for one hour.

14
 Energy density: This is the energy storage efficiency of a battery, mostly

expressed in Watts-hours per Kilogram (Wh/Kg) [27]. It can also refer to the

ratio of the battery energy to its weight [28].

 Depth of discharge (DOD): This is the amount of energy expended from the

battery expressed as a percentage or fraction of its total capacity. For example, if

the DOD specification of a 1000mAh battery is given as 10%, then only 10% of

the battery capacity can be used by the load, i.e. 100mA can be drawn by the load

in an hour.

 Life cycle: This term explains how long a battery can undergo charging and

discharging cycles before its capacity is reduced appreciably. This depends on

the DOD. A high DOD reduces the life cycle of the battery, while a low DOD

increases the number of life cycles a battery can have. Fig. 6 illustrates this

dependence.

15
 Dependence of battery life vs DOD

5000

Number of cycles
4000

3000

2000

1000

Depth of discharge (DOD) (%)

Figure 6 Dependence of cycle life on DOD.

 Self-discharge: This is the discharge of the battery when it is not connected to

any load. A battery gradually loses its capacity when it is not even connected to a

load over time. This is caused by the electrochemical reactions within the cell,

which is equivalent to applying an external load [29]. The 1.5V AA Energizer

L91 Lithium battery [26] has a shelf life of 15 years which indicates that the

battery will lose its capacity completely over a period of 15years due to self-

discharge.

 Memory effect: This is a term that refers to a condition where a battery

temporarily loses its full capacity after discharging and charging below its full

capacity over time. For instance, if a Nickel based battery is always discharged to

70% of its capacity, when it comes to a time when the battery has to be

discharged below the 70% mark, the battery cannot go below that mark since it

has ‘forgotten’ it has more capacity than that.

16
  C-rate: This term is used to describe the rated current capacity of a battery. For

example, a rating of 1C means that a battery with a capacity of 1.5Ah can

discharge 1.5A to a load in one hour or it can be recharged with 1.5A in one

hour.

 Specific power: This is the ratio of the power output of a battery to its unit

weight. It mostly has to do with how much current the battery can provide. A

battery with a high specific power means it can deliver a large current due to its

low internal resistance.

2.3.2 Battery classifications

Batteries can be classified as either primary or secondary. A primary battery is

collection of electric cells which cannot be recharged once it is depleted of charge. A

secondary battery is also a collection of electric cells that can be charged can be once

depleted. This is because the electrochemical reactions in the secondary battery are

reversible while in the primary battery, they are not.

2.3.2.1 Pros and cons of primary batteries

Primary batteries have the advantages of being inexpensive and lasting longer at smaller

or intermittent current demand since they have high energy density. Primary batteries

also have a low self-discharge rate which means longer storage without extensive

reduction in capacity. For example, the Energizer L91 [26], which is a primary battery,

has a shelf life of 15years which indicates that even after a few years of no usage, the

17
battery still has adequate capacity to provide power when needed. This is due to its low

self-discharge rate.

However, they have the disadvantage of having a high internal resistance meaning it

cannot perform well with high current demand applications. Although the initial cost is

low, frequent replacement makes it expensive since it has to be discarded once the

energy is depleted.

2.3.2.2 Pros and cons of secondary batteries

Secondary batteries carry the advantage of performing well under high current demand

applications and having the convenience of rare replacement.

However, secondary batteries have a higher self-discharge rate compared to primary

batteries meaning they cannot be stored for long periods without periodic recharging.

Though the cost of using secondary batteries is lower in the long term compared to

primary batteries, the initial cost is high, which arises from the cost of the batteries and

the chargers associated with it.

2.3.3 Major types of battery chemistry

There are various battery chemistries when it comes to primary and secondary batteries.

Zinc Carbon and Alkaline are some examples of primary battery chemistries. Nickel

based, Lithium based and Lead acid are some types of secondary battery chemistries.

Secondary or rechargeable batteries will be the priority in this thesis. Table 1

summarizes the different battery chemistries.

18
Table 1 Different battery chemistries

Primary batteries Secondary batteries

Leclanchė Cells e.g. Zinc Carbon Nickel based e.g. Nickel Cadmium
Alkaline cells e.g. Alkaline Manganese Lead acid e.g. Lead Calcium
Dioxide
Lithium based e.g. Lithium Manganese Lithium based e.g. Lithium-ion
dioxide

2.3.3.1 Nickel based batteries

There are two types of nickel based batteries, i.e. nickel metal hydride (NiMH) and

nickel cadmium (NiCd). These types of batteries make use of nickel hydroxide as the

cathode, either a cadmium metal or a hydrogen absorbing electrode as the anode,

depending on whether it is a NiCd or NiMH battery, and an alkaline potassium

hydroxide as the electrolyte.

NiMH batteries are gradually replacing NiCd batteries due to NiCd having lesser

capacity or energy density and, being environmentally unfriendly. They are also more

susceptible to memory effect.

Nickel based batteries have been popular over the years and are found in a number of

devices, such as power tools, toothbrushes, two way radios and recently, electric

vehicles. They have the advantages of operating over a wide temperature range,

performing well under harsh conditions, having rapid and simple charging and the ability

to be deeply discharged to almost empty without being affected. However, they are

19
prone to memory effect and high self-discharge rate and thus Lithium ion batteries are

gradually edging them out in some areas in the global market.

2.3.3.2 Lead acid batteries

Due to its low cost and dependable service under rugged conditions, lead acid batteries

account for about half of the demand of rechargeable batteries [30]. They are usually

used in automobiles and standby power applications. Lead acid batteries are made up of

lead dioxide as the anode, lead as the cathode and sulfuric acid as the electrolyte. There

are different varieties of lead acid batteries. Some are sealed lead acid (SLA), deep

cycle, lead antimony and lead calcium. SLA is designed in such a way to prevent loss of

electrolyte through evaporation and hence, they are maintenance free, while deep cycle

batteries are designed to be completely discharged before recharge. Lead antimony and

lead calcium batteries have electrodes which have been modified with antimony and

calcium respectively, to improve performance and lower cost.

Lead acid batteries have the advantages of being low cost, having no memory effect and

low self-discharge rate and working well under rugged conditions. The drawbacks of

lead acid batteries are their inability to be charged quickly, its heaviness and bulkiness,

and the necessity of having proper storage in order to prevent sulfation (the formation of

crystalline sulfur due to loss of moisture from the electrolyte), which increases its

internal resistance and hence lowers performance.

20
2.3.3.3 Lithium based batteries

The number of battery powered applications using lithium based batteries has been

growing over the years. Mobile computing, plug in hybrid electric vehicles and military

applications are some major market shares in the lithium based batteries market. This is

because these batteries have a high energy density, are light weighted and have a low

self-discharge rate. Though lithium is the lightest of all metals, it is very unstable and

can be potentially dangerous. It has to be combined with other metals to improve

stability and reduce the dangers during usage and transportation.

In lithium based batteries, the anode is usually made of carbon, the cathode is a metal

oxide and the electrolyte is lithium salt. The anode and cathode in lithium based batteries

normally reverse roles depending on whether it is charging or discharging. The popular

lithium based batteries are lithium-ion (Li-ion) and lithium ion polymer which has a

solid electrolyte. These two types may have different types of metal oxide cathodes.

Lithium cobalt oxide, lithium titanate, lithium manganese oxide, lithium nickel oxide

and lithium iron phosphate are some types of cathodes used in Li-ion batteries. The

energy density and nominal voltage vary with these different chemistries. Li-ion

batteries come primarily in four basic formats, i.e. cylindrical (small and large), pouch

and prismatic [31]. Small cylindrical cells are enclosed in a metal cylinder while large

cylindrical cells are enclosed in a hard plastic or metal case. Prismatic cells are enclosed

in a semi hard plastic case while pouch cells are enclosed in a soft foil bag.

21
Li-ion batteries have the advantages of having a high energy and power density, no

memory effect, a low self-discharge rate and low weight. Even though Li-ion batteries

are largely used in many battery powered applications, they have some shortcomings

such as stringent charging requirements and degradation when deeply discharged or

overcharged. These shortcomings are far outweighed by the advantages and hence its

usage in a vast majority of battery powered applications. A study by TechNavio analysts

forecasts a compound annual growth rate of 13.64% for the global lithium battery

market in the next three years [32]. Global Information Inc. also predicts the revenue for

Li-ion batteries in the transportation industry is set to grow by 700% by 2017, from $2

billion in 2011 to $14.6 billion by 2017. This is primarily driven by plug-in hybrid

electric vehicles and battery electric vehicles [33]. The Li-ion battery is becoming an

optimum choice for manufacturers of battery powered products; hence this thesis will

focus on charging a standard Li-ion battery. Fig. 7 shows the energy densities versus

battery weight of different battery chemistries. The figure shows that the Li-ion battery

has high energy density at a lighter weight compared to other chemistries and hence, its

gradual domination in the market shares of the battery market.

22
 Diminishing size

1000

Energy density (Wh/L)

800

Lithium based
600

400
NiMH

200 NiCd
Lead
acid

Light weight

200

400

600

800

1000
Energy density(Wh/Kg)

Figure 7 Comparison of battery chemistries in terms of energy density and weight.

2.4 Battery chargers

Rechargeable batteries go hand in hand with good quality chargers and with the global

market for secondary batteries surging, the research and development of better quality

and efficient chargers is not going to end anytime soon. Once charge is depleted from a

secondary battery, the charge has to be replaced through a series of electrochemical

reactions, a process called recharging. The cycle life and performance of a battery

depends on how efficient and quality its charger is. Bad charging techniques destroy

batteries more than any other factor.

A battery charger has three key functions, i.e. charging, optimizing the charge rate and

termination, or knowing when to stop charging [34]. The charging process detects the

voltage across the battery, then starts charging and keeps charging till a specified value

is reached after which charging is terminated [35]. Different battery types have different
23
charging profiles which gives rise to the different types of charging modes. Fig. 8

provides a conceptual view of the main functions of the battery charger.

Battery charger
Input
Charger circuits
+

Battery
Optimization circuits

r
Termination circuits
-

Figure 8 Conceptual view detailing the main functions of a battery charger.

2.4.1 Types of battery charging modes

Depending on the size, type and capacity of a battery, there are different modes of

charging. No single mode is ideal for all types of battery. Also, some battery chargers

may combine two or three different modes to suit the charging profile of the battery.

Some types of charging modes are:

 Trickle charge mode: This is when a small and constant current is used to

charge the battery. This form of charging is sometimes used when the battery is

deeply discharged and needs some form of preconditioning before the main

24
 charge cycle. A Li-ion battery for example, uses trickle charge for

preconditioning when it is deeply discharged. In other instances, when a battery

is fully charged, trickle charge may be used to sustain its full capacity. Fig. 9

shows the trickle charge profile.

Charge current (A)

Battery voltage (V)

Itrickle

Time (sec)

Figure 9 Trickle charge profile.

 Constant voltage charge mode: This is when the battery is charged at a constant

voltage. Current is sourced into the battery during this mode of charging and it

varies to ensure that the battery voltage remains constant. This form of charging

is used for lead acid batteries. Fig. 10 shows the constant voltage charge profile.

25
 Input voltage

Battery voltage (V)

Charge current (A)

Vfull
Icharge

Time (sec)

Figure 10 Constant voltage charge profile.

 Constant current charge mode: In this mode, the current used to charge the

battery is kept constant. When the battery reaches full capacity, the current does

not decrease but remains constant and if charging is not terminated, it can lead to

overheating which can be potentially dangerous. This form of charging is

normally used for NiCd and NiMH batteries. Fig. 11 shows the constant current

charge profile.

26
 Icharge

Vfull

Battery voltage (V)

Charge current (A)

Time (sec)

Figure 11 Constant current charge profile.

 Pulse charge mode: In this mode, a pulse current or voltage is applied to the

battery. This pulse current or voltage has relaxation periods during which the

battery has low impedance. This low impedance allows the next charge pulse to

go into the battery more efficiently, preventing overheating and overvoltage

gassing, thereby increasing the power transfer [36]. Pulse charge can be used for

lead acid or Li-ion battery. When a discharge pulse is applied during the

relaxation period of a pulse charge, it is known as burp charging. Fig. 12 shows

the pulse charge profile.

27
 Ipulse
Vfull

Battery voltage (V)

Charge current (A)

Time (sec)

Figure 12 Pulse charge profile.

2.4.2 Li-ion battery chargers

The Li-ion battery is the battery used in this thesis considering the fact that it is the most

popular battery of choice in today’s portable battery powered applications. As mentioned

earlier, its light weight, high energy density and low self-discharge rate makes it the

perfect battery for mobile and computing applications. As the charge is depleted from

the Li-ion battery, it has to be recharged so as to regain its full capacity. This is done

using a Li-ion battery charger. The Li-ion battery has stringent charging requirements in

order to prevent over-current and over-voltage. Over-voltage can damage the battery or

reduce its cycle life, while over-current can lead to overheating, which can potentially

lead to safety issues.

There are different types of chargers specifically designed for Li-ion batteries which

have complex charging algorithms to meet the charging profile of the battery. A typical

charge/discharge profile of a Li-ion battery is shown in Fig. 13.

28
 Charge/discharge profile of a typical Li-ion
battery
4.2

Charging

Battery voltage (V)

Discharging

2.3

0 100
State of charge (%)

Figure 13 A typical charge and discharge profile of a Li-ion battery.

Most Li-ion battery chargers combine three charging modes, i.e. trickle charge mode,

constant current mode and constant voltage mode, to ensure that the Li-ion battery is not

susceptible to overvoltage which could damage it. Using only one mode of charging for

Li-ion batteries could lead to potential safety issues due to its stringent charging

requirements. A typical charging characteristic is shown in Fig.14.

29
 Trickle
phase Constant current phase Constant voltage phase

Vfull

ICHARGE

Cell voltage
Charging current

ISTART
IEOC

Charge time

Figure 14 Charging profile of a Li-ion battery.

Trickle charge phase is used when the battery is deeply discharged and needs

conditioning. A small current usually ranging from 0.025 to 0.1 of the full charge current

(Istart) is used. If the battery is in a deeply discharged state and a high current is used for

charging, it can damage the battery. This is because in this state, the rate at which

electrochemical reactions take place is slower than the rate at which current is fed into

the battery, hence the need for preconditioning. The preconditioning period takes a

longer time than any of the other modes. Most 4.1V/4.2V rated Li-ion battery powered

systems are designed such that the battery voltage does not fall below 2.3V as seen in

Fig. 13. The battery management system of the charger ensures that the charger operates

in such a way to prevent the battery from going into trickle charge mode, thereby

ensuring the battery voltage does not fall below 2.3V. When the voltage of the battery is

30
above a predefined value, a constant current (from 50% to 100% of the full battery

charge current, which is indicated on the battery datasheet), Icharge, is used to charge the

battery. The constant current phase is much shorter compared to the trickle phase.

However, charging with higher currents does not reduce the charging time, but instead

causes the battery voltage to increase rapidly leading to over-voltage problems [37]. As

the battery approaches full voltage (Vfull), the constant current phase ends and the

constant voltage phase begins. During this phase, the charging current reduces as the

battery approaches Vfull and charging is terminated when a minimum current is reached

(IEOC). This type of charging profile is known as the constant current-constant voltage

(CC-CV) charging method.

Li-ion battery chargers can be distinguished into three main categories, i.e. linear

chargers, switch mode chargers and pulse chargers. Linear and switch mode chargers

uses the CC-CV method of charging while pulse charging involves feeding current

pulses into the battery. The different categories of chargers are used in specific

applications depending on cost, simplicity and efficiency. For example, for low power

applications such as energy harvesting, switch mode chargers are used because of their

low power consumption.

2.4.2.1 Linear chargers

A linear charger is similar to a linear regulator. A linear regulator usually includes a

linear regulating element, which can be a resistor or a transistor. This regulating element

drops the regulator input voltage to a precise output voltage. The difference between the

31
linear regulator and the linear charger is that the charger has added circuitry for battery

charge control and protection.

Linear chargers are simple and less expensive compared to the other chargers. However,

there is constant current continuously flowing through the regulating element, Tp, which

causes a lot of heat dissipation, resulting in the linear charger being inefficient. Fig. 15

shows the concept of the linear charger.

BMS BMS- battery management system

VIN L

_ +
+ Tp DC
C -

R1 +

R2
- BMS
Linear charger Switch mode charger

Figure 15 Linear and switch mode charger concept.

2.4.2.2 Switch mode chargers

Switch mode charges are similar to the switch mode power supplies (SMPS). SMPS

usually have their output voltages regulated through a feedback system and controlling a

series of switches at some frequency depending on the system operation. These power

32
supplies make use of DC-DC converters such as buck, boost or buck-boost depending on

the input voltage. The difference between the switch mode chargers and the SMPS is

that the charger incorporates a series of complex circuitry to regulate charging and

provide protection for the battery.

Switch mode chargers are more efficient than linear chargers since the switches are not

always on, hence dissipating less heat and consuming less power. However, they are

very complex and costly compared to the linear charger. Fig. 15 shows the concept of

the switch mode charger.

2.4.2.3 Pulse chargers

Pulse chargers use the pulse charge mode for charging. This involves a pass transistor

between the input and output which switches on and off, pulsing current into the battery

to charge it. Pulse chargers may contain added circuitry to control the pulse width and

period to make charging more efficient and faster.

Pulse chargers are less complex compared to switch mode chargers and more efficient

compared to linear chargers. Pulse chargers require a carefully controlled current limited

input voltage. Since this requirement is stringent, there is an increase in cost associated

with it. Fig. 16 shows the concept of a pulse charger.

33
 VIN
+
-

BMS
Figure 16 Pulse charger concept.

Table 2 summarizes the pros and cons of each Li-ion battery charger.

Table 2 Comparison of types of Li-ion battery charger

Charger type Pros Cons

Switch mode • low power consumption • Complex

• High efficiency • Large board area
Linear • Simple • High power
consumption
Pulse • Simple • Expensive
• Efficient

2.4.3 Li-ion battery charger design considerations

Before designing a Li-ion battery charger, there are some system parameters that need to

be considered. Some parameters such as input voltage, output voltage regulation, charge

termination method [37], cell protection, cell balancing and charge control are very

important considerations.

34
Input voltage is an important criterion when it comes to battery charger design. It

enables the designer to select a topology suitable for charging. If the input voltage is

higher than the battery voltage, it is important to select a charger topology that will step

down and regulate the output voltage to a precise value. In such instances, a pulse/linear

charger or switch mode using a buck converter can be used. When the input voltage is

lower than the battery voltage, a step up converter will be needed to bring up the voltage

to meet specification.

Output voltage regulation is another parameter to consider. If the output voltage is not

set right, the battery can either be in under-voltage mode, reducing battery capacity or

can be in over-voltage mode, raising safety concerns. The output voltage has to be set

right to achieve the right battery voltage at full charge.

Charge termination methods differ according to what the designer feels best for the

battery. Charge termination is important because if charging is not properly terminated,

it can lead to potential safety concerns and reduced life cycle of the battery. Charge

termination can be done by using a timer or by terminating when charging current

reaches a certain minimum threshold.

Charge control is very essential in preventing over charging. An appropriate mode of

charging should be used according to the state of the battery. When the battery is deeply

discharged, an appropriate method of charging is by using a trickle charge. It is

important to be monitoring the state of the battery in order to use the appropriate method

for charging.
35
Cell protection is a very important parameter when it comes to designing battery

chargers. This is to prevent potential danger to the consumer or the device. Additional

circuitry is needed to monitor charger parameters such as temperature, current and

voltage in order to ensure a safe charging environment.

Cell balancing is done if the charger is going to be used to charge multiple cells. This is

to ensure that charge is equalized over all the cells and no cell is being undercharged or

overcharged. It further increases the life cycle of the batteries and ensures efficiency

when delivering power.

Table 3 summarizes the design considerations based on whether the charger is charging

a single cell battery or multi cell battery.

Table 3 Design parameters depending on multi cell or single cell applications

Design considerations
Number Input Output Charge Charge Cell Cell
of cells voltage voltage termination control protection balancing
Single yes yes yes yes yes no
cell
Multi yes yes yes yes yes yes
cell

2.5 Energy harvesting

Energy harvesting [38] is simply taking energy from the environment and converting it

into electrical energy for daily use. Environmental harvesting [39] is said to be green

36
since it does not contribute to environmental pollution when being harnessed.

Environmental harvesting systems such as hydro power plants, wind mills and solar

farms have been around for a long time. The output power of these large scale energy

harvesters [40] are huge and are fed directly into the electric grid system to be

distributed to consumers. Small scale energy harvesters [39]-[41] which are still

emerging in today’s market produce very little output power and are usually used for

low powered portable applications such as sensors. Micro energy harvesting [40] seeks

to reduce maintenance cost in the frequent replacement of batteries, reduce pollution and

enable the long term powering of devices in places that are not easily accessible. The

most popular sources for micro energy harvesters include solar, thermal, vibration and

radio frequency emissions [39]. The energy available from harvesting radio frequency

emissions is an order of magnitude less than the other three energy harvesters [40]. All

these energy harvesters need some form of power conditioning for it to be able to

provide the order of power needed to power portable electronics.

2.5.1 Solar energy harvesting

This is the most popular type [42]-[44] of energy harvesting and can either be obtained

from indoor lighting or the sun [45]. Solar energy harvesting is done using solar cells or

photovoltaic cells which convert light energy into electrical energy. Depending on

factors such as the intensity of the light, size and number of the solar cells, the output

voltage and current levels may vary.

37
The most popular solar powered device is the solar calculator. There are emerging

markets for solar powered devices such as battery chargers. An example is the Voltaic

amp solar charger, which is designed to charge smart phones, tables, digital cameras and

GPS [46].

2.5.2 Thermal energy harvesting

This involves the use of thermoelectric generators to produce electrical energy when

there is a temperature difference across two different metals through a phenomenon

called the Seebeck effect [38]-[39], [47]-[48]. The Seebeck effect is observed when a

thermocouple made up of two dissimilar conductors produce electricity when the

conductors are maintained at different temperatures [48]. In instances where light is not

present and there is a substantial amount of heat such as in automobiles, thermoelectric

generators can produce some adequate amount of power for low powered applications.

In order to obtain a considerable amount of voltage, there has to be a large temperature

difference. This problem can be solved by increasing the number of generators but the

drawback is that it becomes inefficient because more power is lost across the increased

resistance. Recently, integrated circuits have been designed to boost the output power

from these thermoelectric generators, which could open the market for wearable devices

that make use of body heat. A good example is Seiko Thermic, a wristwatch powered by

body heat using thermoelectric generators [49].

38
2.5.3 Vibration energy harvesting

This is the conversion of kinetic energy of vibrating surfaces into electrical energy. Not

all vibrating surfaces can be used to produce electrical energy. To be able to convert

vibrations into electrical energy, the vibrations should have a resonant frequency.

Electricity produced from vibration energy is alternating and must be rectified in order to

be used for portable applications. Examples of converters which convert vibrations into

electrical energy are piezoelectric, electromagnetic and electrostatic transducers [48].

Piezoelectric transducers use piezoelectric materials, which have the property of

generating electricity under strain or stress. Electromagnetic transducers produce

electricity based on the movement of a magnet coil, while electrostatic transducers are

based on the movement of the plates of a capacitor, which varies capacitance producing

varying voltage.

There are commercial products, which make use of vibration energy today even though

there are still challenges being faced in this industry. The Nike smart shoe is a typical

example, where the speed and number of steps a person takes are wirelessly linked to a

phone or some kind of application. The vibrations from movement are converted into

electrical energy to power the sensors in the shoes. Also, the nPower PEG (personal

energy generator), which can be placed in anything that moves, can harvest enough

energy from the kinetic motions to recharge portable electronics [50].

Table 4 summarizes different energy sources and their estimated harvested power.

39
Table 4 Energy sources and their estimated harvested power

Energy source Harvester Estimated harvested

power [40],[48]
Human motion Piezoelectric transducers 4µW/cm2
Industrial machine Electromagnetic/electrostatic 100-800 µW/cm2
vibrations transducers
Thermal Thermoelectric generators 25-60 µW/cm2
Direct sunlight Solar panels 10-100 mW/cm2
Indoor light Solar panels 10-100 µW/cm2
Radio frequency Antennas < 1 µW/cm2

2.6 The battery and the supercapacitor

The battery and the supercapacitor, also known as electrochemical or ultracapacitor,

complement each other in systems where both are present. The supercapacitor differs

from conventional capacitors by not having an intervening dielectric material between

electrodes [51]. A supercapacitor has its two electrodes immersed in an electrolyte with

a separator between the electrodes [52]. The electrodes have much higher surface area

and decreased distance between them, hence having much higher capacitance and energy

than the conventional capacitors [53]. Fig. 17 shows the conventional capacitor versus

the supercapacitor in terms of operation and equivalent circuit. A supercapacitor can be

modeled to the first order as a conventional capacitor but due the porous material used to

make its electrodes, it exhibits a behavior closely related to transmission lines [54].

40
 anode cathode anode cathode
+ - + - + -- + -
+ + -
+ - + - + - + -
+ - + - + -- + -
+ + -
+ - + - + - + -
+ - + -
load dielectric load
separator

DC DC

Conventional capacitor schematic Supercapacitor schematic

Rp L R1 R2 Rn
+
L
Rp C1 C2 Cn
Rs

Equivalent circuit of a conventional capacitor Equivalent circuit of a supercapacitor

Figure 17 Conventional capacitor versus supercapacitor.

A battery can store more energy than a capacitor, i.e. high energy density, but it cannot

deliver it quickly, i.e. low power density, hence supercapacitors are ideal for applications

such as sensors where short peak current pulses are needed. Supercapacitors have the

advantage of having the ability to be charged rapidly and hence they are ideal for energy

harvesting applications. They also have a longer shelf life and can perform well under

rugged conditions, such as relatively higher and lower temperatures, compared to

batteries. They can undergo many charge and discharge cycles without affecting their

capacity because energy is stored in their electric field compared to a battery which

stores energy based on a chemical reaction. Despite these wonderful characteristics, they

41
cannot replace batteries since they have high self-discharge rates and are expensive. This

can be explained by comparing the equivalent circuit of the battery, shown in Fig. 5, to

that of the supercapacitor (Fig. 17). The resistance, Rp, is negligible in the battery but

present in the supercapacitor, and is responsible for high leakage current between the

electrodes.

A battery cannot deliver short term peak currents to a load, while a supercapacitor

cannot meet the demands of long term peak current. This is because a supercapacitor

stores energy in its electric field and not due to chemical reactions and hence can deliver

power quickly from its electric field. Therefore, a battery-supercapacitor combination

allows peak current pulses to be delivered to a load whenever needed and at the same

time sustaining longer term load power demand. This ensures the battery runs longer,

safer and at reduced output impedance, which prevents unwanted transients [53]. These

two energy storage elements will therefore complement each other in the charging

system designed in this thesis.

2.7 Overview of charging system

The main aim of this thesis is to design a battery charging system, which can be used

with an energy harvesting systems, USB applications and wall adapters and also

reconfigurable to be used as a constant current or trickle charger. This charging system

should be able to support sensors in remote locations, i.e. ensuring that battery life lasts

longer. Typical voltages for sensors to be used as load for this charging system range

from 2.7V to 5.5V, hence a Li-ion coin or prismatic cell with voltages of 3V or 4.2V

42
respectively will be the target cell to be charged. The charging system seeks to add

additional blocks to improve functionality in order to support an energy harvesting

system. These additional blocks are a step down DC to DC converter and a switching

network system to enable the charging of the battery from an energy harvesting system.

2.7.1 Description of charging system block diagram

A block diagram of the charging system is shown in Fig. 18. The energy harvesting

system (EHS) is treated as a black box providing a certain output current and voltage in

this project. It is treated as another input source.

Input voltage Proposed battery charging system

Vsupply
Current limited
wall adapter Pulse charger Step down
in out
AC
`

+
USB _ battery

EHS charge battery

Vsupply

Energy Switching
harvesting EHS output network
system

EHS charge Csuper

Load
Csuper

Figure 18 Charging system with DC-DC converter and switching network.

43
A pulse charger is chosen due to its simplicity, faster charging and less power

consumption. The pulse charger will take its input either from a current limited wall

adapter or from a USB. It will then charge the battery to its full voltage and monitor

system parameters such as input and battery voltage, charge current and temperature to

ensure safe charging and battery protection.

The step down DC-DC converter is needed to down convert the battery voltage to an

ideal voltage suitable to be used as supply for the energy harvesting system. Energy

harvesters provide output voltages that are very low, i.e. microvolts or millivolts. These

small voltages are then boosted by charge pump circuits or boost converters to a higher

voltage, which can then be used to power an electronic circuit. These booster circuits

make use of complementary metal-oxide semiconductor (CMOS) transistors or bipolar

junction transistors (BJT) which need a relatively higher voltage for operation. The low

voltages produced by the harvesters are not enough to power up the CMOS transistors or

BJTs, therefore an initial battery voltage is needed to kick start the system. The step

down DC-DC converter is necessary to achieve this initial voltage for startup. The

converter can also produce a voltage which might be needed somewhere in a larger

system.

To prevent reduced efficiency, the switching network system is designed to bypass the

pulse charger and enable battery charging directly from an EHS. It also enables the

supercapacitor being charged by the energy harvesting system to deliver short term peak

current to a load when needed without affecting the battery, thereby enabling both the

44
battery and supercapacitor to complement each other. This block ensures that the EHS

and the pulse charger operate independently from each other and each can be used as a

standalone device.

2.7.2 Description of charging system flow process

When the charging system recognizes that an input voltage, i.e. USB or wall adapter is

present, it prevents the battery from being charged from the energy harvesting system.

The battery is then charged to a certain voltage which can support the EHS. This is

achieved by stepping the battery voltage down. This voltage serves as power supply to

power up the EHS to charge the supercapacitor. When the supercapacitor voltage is high

enough to support the load and the wall adapter and USB inputs to the pulse charger are

not available, the switching network system turns on to charge the battery from the

energy harvesting system. When both battery and supercapacitor voltages are good, the

switching network system has the option to operate in such a way as to use the

supercapacitor for peak load currents and the battery for constant load demand. The

above description is shown in the flow chart in Fig. 19.

45
 Input voltage Energy harvesting
system connected

Yes Yes
No

USB/AC Pulse charger

No Good supply
available off

Yes Yes

Yes Charge
Start charging
supercapacitor

Yes Yes
No No

Battery Supercapacitor
Yes voltage good
voltage good

Yes Yes
Start switching
Supply load
Stop charging mechanism

Yes

Start step down

DC-DC converter

Figure 19 Flow chart showing the operation of the battery charging system.

46
 3. PULSE CHARGER

The type of battery charger designed in this thesis is the pulse charger. It is chosen over

the switch mode and linear chargers because of its lower power consumption and

simplicity. Pulse charging can be applied to both nickel based batteries and Li-ion

batteries. This thesis looks at both NiMH and Li-ion batteries. Tests performed by [55]

indicate that pulse charging eliminates concentration polarization, increases power

transfer rate and lowers charge time by getting rid of the constant voltage charging

phase. Concentration polarization is when ion mobility is hindered in the electrolyte of

the battery, increasing internal resistance and voltage drop, thereby preventing the full

voltage of the battery from being reached. In pulse charging, the period between pulses

allows the chemical reactions in the battery to stabilize before charging begins, enabling

the electrochemical reactions to keep pace with the rate of electrical energy input [34].

Pulse charging seems to be an effective method for increasing the life cycle of a battery

and achieving a higher discharge capacity [55]. This makes it an ideal choice for the

application in this thesis. This section introduces some previous works and commercially

available pulse chargers, the conceptual idea of the proposed pulse charger, a description

of the block level diagram, a detailed transistor level design and finally presents

simulation results to prove the proposed charger design works.

3.1 Previous works

Published works making use of the pulse charging strategy to charge Li-ion batteries

include the phase locked battery charger proposed in [56] which uses a phase-locked
47
loop (PLL) to charge a 4.2V, 600mAh Li-ion battery. The phase locked battery charger

makes use of the inherent properties of a PLL, such as auto locking, auto tracking and

high accuracy, to charge the battery. A PLL consists of a phase comparator, a low pass

filter and a voltage controlled oscillator. The battery is incorporated in the low pass filter

in order to construct a fast, safe and cheap battery charge system. In this published work,

the charging process consists of three phases: a constant current charge phase where an

upper bound current is used to charge the battery, a variable current charge phase where

a decreasing current is used for charging and a float charge phase where a small current

is used to maintain the battery at a preset voltage. In this work, the 600mAh Li-ion

battery is able to charge to 4.2V in 150 minutes proving that though the designed circuit

is simple, it can provide a suitable charge current, making the PLL a perfect battery

charge circuit topology [56]. Fig. 20 shows the block diagram of the proposed phase

locked battery charger.

Figure 20 Proposed block diagram of phase locked battery charger.

48
A variable frequency pulse charge system (VFPCS) is presented in [57] that can

dynamically track the optimal charge frequency to improve battery-charge response.

This work seems to overcome the trial and error method used in commercial battery

pulse charge systems to determine the optimal frequency of the pulse charges. The

VFPCS can determine what optimal pulse charge frequency is needed during charging

and supply that optimal charge pulse to increase charge speed. The VFPCS is

implemented using a microprocessor to detect and control the frequency optimally. A

voltage regulator, transistors and a sensing resistor serve as a variable frequency pulse

generator. Experimental results verified the performance of the VFPCS by charging a

600mAh Li-ion battery. It showed an improvement of 24% in charge time compared to

the standard constant voltage-constant current (CC-CV) charging, proving that the

VFPCS can provide pulses with an optimal frequency to charge a battery at a reduced

charge time [57]. Fig. 21 shows the implementation of the proposed VFPCS.

49
Figure 21 Implementation of VFPCS.

A duty varied voltage pulse-charge strategy (DVVPCS) is proposed in [58] to detect and

dynamically track the suitable duty of the charge pulse to improve battery charge

performance. Commercial battery pulse charge systems use a trial and error method

which might not be suitable. The DVVPS therefore detects the suitable charge pulse-

charge duty and supplies that suitable charge pulse to the battery to decrease charge time

and improve efficiency. The DVVPCS consists of a controller, a duty varied pulse

generator (DVPG), an average current sensor and two analog-to-digital (ADC)

converters. The controller and ADC are implemented using a microprocessor to detect

the charging modes of the battery. The average current sensor is implemented using an

50
operational amplifier, a current sensing resistor and a capacitor to detect the average

charging current. When the charge current is determined, a corresponding duty,

generated by the DVPG, is obtained to increase battery charging speed and efficiency.

Experimental results proved that the charge speed of the DVVPCS is improved by 14%

compared to the standard CC-CV method. The results indicate that the DVVPCS can

actually provide charging pulses with suitable pulse-charge duty for Li-ion battery to

obtain fast-charge process. Fig. 22 shows the block diagram of the proposed DVVPCS.

Figure 22 Block diagram of DVVPCS.

Commercial pulse chargers for portable applications on the market today include the

LTC4052, LTC1730 and MAX1879. LTC4052 integrated circuit (IC) is a complete

standalone pulse charger with an integrated metal oxide semiconductor field effect
51
transistor (MOSFET) for single cell Li-ion batteries. It is ideal for cell phones and

handheld computers. This IC has over-current protection, a thermal shutdown, defective

battery detection and a programmable charge termination. The charge current is set by

the wall adapter. It has a trickle charge mode which activates when the battery is below

2.45V. The charger goes into a full charge mode when the battery goes above this

voltage. When the battery approaches the programmed float voltage, the MOSFET

begins switching with a gradually decreasing duty cycle and termination is achieved by a

programmable timer [59]. Fig. 23 shows the block diagram of the LTC4052.

Figure 23 LTC4052.
52
LTC1730 is a complete pulse charger for single cell Li-ion batteries. It can charge a

completely discharged cell by allowing the current limited input power source to provide

charge current to the battery through the internal MOSFET. It has the ability to either

charge a 4.1V or 4.2V battery by setting the SEL pin. Its trickle charge mode will only

activate when the battery voltage is less than 2.45V, after which full charge mode

operation starts. This is when the internal MOSFET is turned on. It also has over-current

protection, thermal shutdown and a programmable charge termination timer [60]. Fig. 24

shows the block diagram of the LTC1730.

Figure 24 LTC1730.

53
MAX1879 is a single cell Li-ion battery charger which uses pulse charging strategy to

minimize power consumption. It combines the efficiency of switch mode chargers and

low power consumption of linear chargers. The IC monitors voltage and temperature

continuously and has a preset charger time out function. It can automatically shut down

when it detects that input power has been removed to minimize the current drain of the

battery. It also has the ability to charge deeply discharged cells and restart charging at

4V. It can be used in many handheld devices, such as cell phones, digital cameras and

pagers [61]. Fig. 25 shows the functional diagram of the MAX1879.

Figure 25 MAX1879.

Table 5 summarizes a comparison of the previous works and commercial pulse chargers

discussed above.
54
Table 5 Comparison of previous works

VFPCS DVVPCS PLL MAX1879 LTC4052

[57] [58] based [61] /LTC1730
[56] [59],[60]

Control loop PFM PWM - PWM PWM

Cell Voltage 4.2V 4.2V 4.2V 4.2V 4.2V/4.1V

User no no no Partially partially

programmable

Precondition Not Not indicated Not present present

phase indicated indicated

Over Absent Absent absent present present

temperature
protection

Design Software/ Software/ discrete analog analog

implementation discrete discrete

Power - - - minimal minimal

consumption

Charge current Average Average Average User User

current current current defined defined

55
3.2 Conceptual idea of proposed pulse charger

The main idea behind this work is to make the pulse charger completely versatile. [56]-

[58] propose pulse chargers which reduce charging time. It does not address the issue of

a deeply discharged Li-ion battery and also it is discretely implemented. The proposed

pulse charger designed in this thesis not only addresses fast charging but also provides a

pre-conditioning phase for deeply discharged cells. [59]- [61] are commercial pulse

chargers with provision for deeply discharged cells and have important safety

mechanisms in place to protect both the user and the battery. However, these chargers

can only be used as pulse chargers and does not have the versatility of the pulse charger

designed in this thesis. The proposed pulse charger can be used as a standalone pulse

charger, a constant current charger or a trickle mode charger. It also seems to provide the

user with full programmability options. The trickle current, for example, in [56] - [58] is

preset during design and cannot be changed by the user. The charger designed in this

thesis provides an option to the user to set the value of the trickle current. The full

charge current is set by the current limited wall adapter, similar to that in [59] - [61]. Fig.

26 shows the conceptual idea of the proposed pulse charger.

56
 M1
VIN

Decision
User set
block Trickle mode
User set
Overcurrent
protection

Overvoltage Over Fast charge/

protection temperature Constant current
User set User set
protection mode

Figure 26 Conceptual block diagram of proposed pulse charger.

Most pulse chargers have a control loop that controls the current pulses and that loop is

not accessible to users. They use a form of empiricism to determine the duty cycle. The

proposed pulse charger seeks to give control of the period and frequency of the current

pulses to the user of the device. The charge current, overvoltage set point, over

temperature set point, pre-charge current, power good set point and current limiting can

all be user set. This is to enable a wider range of Li-ion coin, prismatic and cylindrical

cells with different nominal voltages to be charged and possibly nickel based batteries.

The proposed pulse charger is set to charge a maximum of 3.3V rated battery. This is

because the maximum voltage rating of transistors used in the design is limited to 3.6V.

To prevent permanent damage or device failure, the proposed pulse charger must limit

57
its input voltage to 3.6V. Specifications of the 3.3V/2.3Ah/7.6Wh OlevinPower Li-ion

single cell battery (A123) [62] is used to set design parameters in this thesis.

3.2.1 Flow chart of proposed pulse charger

A flow chart of the proposed pulse charger is shown in Fig. 27.

Input

Yes

Battery
connected,current, Disconnect charging
No
temperature & input cycle
voltage ok

Yes

Precharge
Vbatt< Fast charge
setpoint<Vbatt<
Precharge No No setpoint<Vbatt<full No Vbatt>full charge Yes Stop charging cycle
Fast charge
setpoint charge
setpoint

Yes Yes Yes Yes

Start trickle charge Start fast charge Start slow charge M1 off
phase phase phase

Disable trickle
Enable trickle current source, M1 on, current All unneeded
current source, M1 on, current pulses at 25% comparators
M1 off pulses at 75% duty cycle off
duty cycle

Figure 27 Pulse charger flow chart.

When the input is connected, i.e. a current limited wall adapter, the input current, voltage

and current temperature of the device is checked to ensure that they are all within the

safe operating range (SOR). If these parameters are within the SOR and a battery is
58
detected, the charger is given the green light to proceed with charging. These parameters

are continuously checked throughout the entire charging process and when any of the

parameters fall outside the SOR, the charger is instantly disconnected to prevent any

damage.

When charging commences, the battery voltage is determined to see which charging

phase is appropriate. If the battery voltage is less than the user set pre-charge voltage, the

trickle charge phase is initiated. During this phase, the main power switch, M1 is off.

This is to allow an internal current source to pre-charge the battery to a voltage set by the

user. The current for pre-charging is also user set and this portion of the device can be

used as a trickle charger by disconnecting all other phases.

Once the battery voltage is above the set pre-charge point, the current source for pre-

charging is turned off and M1 becomes operational. If the battery voltage is above the

set pre-charge voltage, the charger determines whether the battery voltage is above the

fast charge voltage set point or below. If it is below, the charger proceeds with charging

with a higher duty cycle. During this phase, the input current is pulsed at 75% duty cycle

into the battery. This ensures a faster charge time and at the same time enables

electrochemical reactions to keep pace with the input of electrical energy. When the

battery voltage approaches full charge, i.e. above the fast charge set point, the input

current is pulsed at 25% duty cycle to ensure that maximum capacity is achieved and

battery charge efficiency is increased. The user can decide on the frequency of the pulses

59
by providing a clock with a desired frequency. This clock pin can be manipulated to

enable the pulse charger serve as a constant current charger.

Termination occurs when the battery voltage crosses the full charge set point, which is

also user set. When termination occurs, M1 and all unneeded comparators are turned off

to reduce power consumption if the input is still connected.

3.2.2 Description of block diagram

The block diagram of the proposed pulse charger based on the flow chart is shown in

Fig. 28. In Fig.28, the pulse charger is in normal mode and has three charging phases,

i.e. pre-charge, fast charge and slow charge. The pre-charge phase includes the current

source, IT, and the comparator, TC. The fast and slow charge phase is controlled by the

FC comparator and a series of logic circuits in the decision block of the charger.

Termination is achieved by the OV comparator in connection with logic circuits in the

decision block to ensure the battery is not overcharged. M1 is the main power transistor

which switches on and off to pulse the input charge current into the battery. M2 is a

fractional size of M1 and in combination with M3, comparator C1 and error amplifier C2

form the current limiting branch. The current limit is set by connecting a resistor to the

RLIM pin. The PG comparator ensures that the input voltage is always good enough for

charging while the temperature sense block ensures that the circuit does not operate

under high temperature conditions which can cause device failure. Table 6 shows the

specifications of the pulse charger.

60
 M1
VIN BATT

M3
RLIM TC
M2 IT PRECHARGE
VIN
PHASE
C1

C2

DECISION BLOCK
FC
VREF FAST CHARGE
PHASE

OV
POWER GOOD OVERVOLTAGE
SET VREF PG SET

CLOCK
TEMPERATURE
SENSE

Figure 28 Block level diagram of pulse charger.

Table 6 Specifications of pulse charger

SPECIFICATION Min Typ Max UNIT

Input voltage, Vin 3.3 3.3 3.6 V
Input current, Iin 200 500 mA
o
High temperature cutoff, Toff 85 C
Pre-charge current, IT 5 10 20 mA
Battery voltage, Vbatt 3 3 3.3 V
Reference voltage, Vref 1.2 V

3.3 Circuit level implementation

This section seeks to describe the building blocks of the pulse charger in more detail.

The basic building blocks, including the comparator, which is part of every decision
61
making block and the duty cycle generator, which forms the core of the fast charging

phase of the charging process, are discussed. Secondary building blocks, such as the

trickle and fast charging blocks, are also discussed in more detail. Circuits such as the

current limiting and over temperature blocks which ensure protection and safety during

operation of the pulse charger are also presented. Simulation results showing the

operation of each building block are also provided.

3.3.1 Design of basic building blocks

The main basic circuit, which must have good properties due to its frequent use in the

charger, is the comparator. The comparator is used at every stage of the charging process

and also in the protection schemes. It is necessary to ensure that the designed comparator

achieves low power consumption. The duty cycle generator is another basic block for the

charger. It is used in the fast charging phase to ensure a smooth transition from one

phase to another.

3.3.1.1 Design of comparator

Comparators are circuits that output either a low or high signal depending on the two

input signals it receives. Comparators can be classified as open loop comparator,

regenerative or a combination of both. An open loop comparator employs an operational

amplifier with no compensation while regenerative comparators employ positive

feedback. Fig. 29 shows a regenerative comparator using resistors in positive feedback.

62
 Vo

Vout VOH
_
Vin
+

R1 VTL VTH Vin

R2

VOL

𝑅2
ℎ𝑦𝑠𝑡𝑒 𝑒𝑠𝑖𝑠 𝑤𝑖𝑛𝑑𝑜𝑤 = ( 𝑂𝐻 𝑂𝐿 )
𝑅2 + 𝑅1

Figure 29 Regenerative comparator showing hysteresis.

Fig. 30 shows an open loop comparator with two analog input signals, V+ and V-, and a

binary output signal, Vout .

VDD
V+ < V- ; Vout = VOL
V+ +
Vout VOH V+ > V- ; Vout = VOH
VOL
V- -
VSS

Figure 30 Open loop comparator.

If V+ is greater than V-, the comparator outputs a high voltage, VOH, and when V- is

greater than V+, the comparator outputs a low voltage, VOL. In this project, the

63
comparators are used in decision circuits to ensure smooth transition from one charging

phase to another. The comparator can therefore be seen as a single-bit analog to digital

converter.

One important feature needed in the comparator is hysteresis. Hysteresis is the difference

between input signals at which the comparator outputs stays at either a high or low

voltage depending on its value before it entered the hysteresis window. The window of

hysteresis is shown in Fig. 31. The figure also shows an analog input signal

superimposed with noise and as it crosses the reference voltage, multiple transitions are

triggered at the output of the comparator. If a comparator is designed with hysteresis, the

output signal retains its value in the hysteresis window, VTH-VTL, even if the signal

crosses the reference voltage many times. The output of the comparator will change once

the input signal crosses one of the thresholds, i.e. VTH or VTL. Hysteresis ensures that a

noisy signal does not cause these undesirable multiple transitions between states when

the input signal is at the switching threshold, hence providing some form of filtering.

64
 VTH- Upper threshold voltage
VTL- Lower threshold voltage
VTH-VTL = hysteresis window
Input signal

VTH
Vref Vref
VTL

Output signal

Without hysteresis With hysteresis

Figure 31 Advantage of hysteresis window.

In this thesis, the comparator designed is shown in Fig. 32. It is composed of a single

stage operational amplifier with low offset and enough gain, a CMOS Schmitt trigger to

provide hysteresis and a latch to provide stability and a sample and hold function [63].

65
 EN MP3
MP6
MP5
B
MP1 MP2

MP4
Vis Vo
OUT

IN+ IN- MN6

MN1 MN2
A Latch
IN MN8
MN7

MN4 MN3
EN MN5

Single stage amplifier CMOS Schmit trigger

Figure 32 Transistor level diagram of comparator.

The single stage amplifier is designed to incorporate an enable/disable function, have

low power consumption, a gain greater than 40dB and a low offset voltage. To improve

matching in the current mirrors (MP1, MP2, MN3 and MN4), long length devices are

used. MP6 and MN5 implement the enable/disable function. If the input signal to MP6 is

low, the transistor turns on and shuts off the current mirror load (MP1 and MP2).

Likewise, if the input signal to MN5 is high, it turns on and shuts off the current mirror

(MN3 and MN4). This function ensures the comparator is turned off at a time it is not

needed in order to reduce power consumption in the entire pulse charger. Another

technique to ensure low power operation is to design in the subthreshold or weak

inversion region.

66
In the subthreshold region, the MOSFET gate is held below the threshold voltage and

diffusion current dominates as compared to drift current in saturation region [64].

Diffusion current is the flow of current from the source to the drain of the transistor

while drift current is the flow of current controlled by a voltage applied to the gate of the

transistor. Equation 2 shows that the drain current, Ids, in the subthreshold region is

independent of the drain to source voltage, Vds if Vds is greater than the thermal voltage,

VT [65]. The transconductance equation, shown in equation 3, shows a linear relationship

between the transconductance, gm and drain current. The transconductance to drain

current ratio for a MOSFET is maximized in subthreshold operation [66].Hence, a

moderate gain, low noise and low power consumption can be achieve in this area of

operation. Gain, Av is increased and noise, vn2 is decreased as gm is increased as seen in

equations 4 and 5 respectively.

(𝑒 )( 𝑒 )
𝐿

where m is the body effect coefficient and Kp,n is a process parameter.

where 𝑅 .

67
where Kf is a fitting parameter, f is frequency, T is temperature, Cox is gate oxide

capacitance, Ro is output impedance, W and L are width and length of transistor

respectively.

MN1 and MN2 are designed in sub threshold with an inversion level of one. This is

because higher inversion level means smaller area and higher speed, but higher power

consumption. At smaller inversion levels, power consumption is low but circuit

operation is slow, hence an inversion level optimizing area, speed and power conversion

is appropriate. This can be inferred from Fig. 33 which shows that at the higher inversion

region, transistor sizes are small, i.e. small area, but there is a large consumption of

power and higher speed. At smaller inversion region, power consumption is low while

area is large and speed is smaller due to a smaller bandwidth of operation. Between the

inversion levels of 0.1 and 10 where there is just a slight change in power consumption;

a smaller area with better bandwidth operation can be picked. Fig. 33 shows that a good

inversion level can be chosen at the intersection of the normalized power consumption,

area and speed curves.

68
 3
plot of various transistor parameters vrs inversion level
10
normalized ft
normalized silicon area
normalized power consumption
2
10

1
10
normalized ft, power and area

0
10

-1
10

-2
10

-3
10
-3 -2 -1 0 1 2 3
10 10 10 10 10 10 10
inversion level,(iF)

Figure 33 Plot of various transistor parameters versus inversion level.

With a gain of 40dB, Ids of 1uA, the gm of the transistors can be calculated by following

the procedure below.

For an inversion level of one, i.e. if=1,

𝑛 √ 𝑖

The dimensions of the transistors can be calculated as follows,

𝐿 √ 𝑖
69
𝐿

After obtaining the initial dimensions of the transistors, the circuit is adjusted to achieve

a good gain at low power consumption. The final transistor dimensions are shown in

table 7. The gain and phase margin plot is shown in Fig. 34.

Table 7 Comparator first stage transistor dimensions

Transistor W/L (um)

MN1 4/1.5 * 1
MN2 4/1.5 * 1
MN3 4/1.5 * 3
MN4 4/1.5 * 3
MN5 4/1.5 * 3
MP1 4/1.5 * 4
MP2 4/1.5 * 4
MP6 4/1.5 * 2

70
Figure 34 Gain and phase plot of amplifier stage in comparator.

The CMOS Schmitt trigger, shown in Fig. 32, is added to introduce hysteresis into the

comparator. Transistors MP5 and MN8 introduce hysteresis by operating as source

followers providing some form of positive feedback where they feedback the output

voltage, Vo to node B and A respectively [63]. When the input, Vis, to the CMOS

Schmitt trigger is low, transistors MP3, MP4 and MN8 are on operating in the triode

region, while MN6, MN7 and MP5 are off and Vo is equal to the supply voltage. As Vis

increases to a value above the threshold voltage, MN7 begins to turn on. At this point,

both MN7 and MN8 are on and form a biasing network to bias MN6. The condition of

the P transistors does not change. The voltage at node A, VA begins to decrease. When

Vis reaches the upper threshold voltage, VTH, MN6 turns on fully and transition starts

with Vo starting to go low. MP4 and MP3 start to turn off when MP5 turns on, further

reducing Vo due to the source follower operation. The cycle then repeats itself when the
71
input goes low again [63], [67]. The hysteresis at the output of the latch is shown in Fig.

35.

Vo

VOH

VHW

VOL
VLI VHI Vin

Figure 35 Hysteresis of comparator.

VOH and VOL is the maximum and minimum output voltage respectively. VOH and VOL

are dependent on saturation levels of the amplifier. VHI is the input voltage at which

output switches from VOH to VOL and VLI is the input voltage at which the output

switches from VOL to VOH. VHW is the hysteresis window. To determine the VOH, VOL,

VLI and VHI, the following equations are used [68].

At the time when the output is switching from high to low, MN7 and MN8 are in

saturation and MN6 is off,

72
where β is the transconductance parameter of the transistors if .

The current through MN7 and MN8 is the same and both are in saturation.

From equation 11, VHI from Fig. 35 can be determined if VHI=Vin.

[ (√ )]

To determine VLI, just before the output switches again, MP3 and MP5 are in saturation.

Similarly,

If ,

Similarly, the current through MP3 and MP5 is the same and both are in saturation.

73
 √

( √ )

For a VLI of 1.45V, VHI of 1.55V and a VHW of 100mV, using equations 12 and 17, the

transistor dimensions can be determined.

𝐿
𝐿

𝐿
𝐿

The transistor dimensions of the Schmitt trigger are shown in table 8. Fig. 36 shows that

the comparator has a hysteresis window of 100mV.

Table 8 CMOS Schmitt trigger transistor dimensions

Transistor W/L (µm)

MP3 4/0.5 *4
MP4 4/0.5 *4
MP5 4/0.5 *5
MN6 4/0.5 *2
MN7 4/0.5 *2
MN8 4/0.5 *2

74
Figure 36 Hysteresis window of designed comparator.

3.3.1.2 Design of duty cycle generator

In many pulse chargers, users do not have access to the switching frequency of the main

power switch. The duty cycle and frequency of the charge pulses are normally obtained

through a trial and error method [57]. This can cause the main switch to be on for longer

periods than needed and might contribute to extra power consumption or even

overcharging. There is no clear literature to specify the optimum duty cycle for pulse

chargers. Most internal regulation loops are similar to a pulse width modulation (PWM)

or pulse frequency modulation (PFM) system, which varies the pulse width or

frequency, respectively, with respect to the battery voltage to achieve a fully charged

battery.

75
A study by [69] indicated that better charging response is obtained when lower

frequencies are used for pulse charging. In [70], it was suggested that charging at low

frequencies is faster than charging at high frequencies. It also suggests that higher duty

cycle pulses are more efficient in charging batteries than lower duty cycle. A higher duty

cycle will result in faster charging but will cause the battery to be prone to overcharging

and overvoltage conditions, while a lower duty cycle will result in longer charge time

but good battery charge efficiency. It is therefore normal to conclude that the best duty

cycle is 50%. This conclusion is unacceptable since the electrolyte concentration differs

at different times during charging [58]. Thus, it is necessary to provide a method where

faster charging can be achieved and at the same time achieve a good battery charging

efficiency without entering overvoltage or overcharge condition. In order to do this, the

charging phase is divided into two parts: one part to provide a fast charging time using a

higher duty cycle and the other part using a lower duty cycle to enable better battery

charging efficiency. Due to the stringent charging requirements of a Li-ion battery, a

duty cycle of 0.7 or higher is appropriate for the initial stage of charging which then

reduces as full charge is approached. The duty cycle for faster charge is set at 0.75 while

that for better charge efficiency is set at 0.25. This can be generated using a duty cycle

generator. A conceptual idea of the duty cycle generator (DCG) is presented in Fig. 37.

76
 Voltage detector

Deeply discharge<Vbatt<full charge Vbattà full charge +

battery
Fast charge Better charge efficiency -
75% duty cycle 25% duty cycle

Figure 37 Conceptual idea of DCG.

To provide versatility, giving the user access to the frequency at which the switch is

pulsed and also providing options to use the pulse charger as other type of chargers, a

duty cycle generator is designed and added to the proposed pulse charger. The duty cycle

generator uses flip flops. A flip flop is a memory device that is controlled by a clock

signal. It stores only one bit of data. A positive edge triggered D flip flop operates by

changing state at the positive or rising edge of the clock input. If the data bit, D, is high

when a clock pulse is applied, the output, Q, is also high, i.e. the flip flop sets. Likewise,

if D is low when the clock pulse is applied, Q goes low and the flip flop resets. The set

and reset, also known as clear, inputs take precedence over the clock input, i.e. the flip

flop will only operate if both set and clear are high. Fig. 38 shows the timing diagram

and truth table of the positive edge triggered D flip flop.

77
 Data

Data D
SET
Q out Clock Data Clock out out
0 0 1
Clock out 1 1 0
CLR Q out
out
Truth table
D flip flop

Timing diagram of positive edge

triggered D flip flop

Figure 38 Positive edge triggered flip flop operation.

The duty cycle generator circuit, shown in Fig. 39, consists of two positive edge

triggered D flip flops. The first flip flop divides the frequency of the clock by two at

node A, while the second flip flop serves as a modulo two divider using the divided

clock from the first flip flop at its clock input [71]. The NAND gates will then generate

the waveforms with 25% and 75% duty cycle. Fig. 40 shows the waveforms at node A,

B, C and D. The proposed pulse charger clock input frequency should be less than 1

KHz since charging at lower frequencies shows some superiority to higher frequencies

[70].

VDD
C
A 25% duty cycle

SET SET
D Q D Q D
B 75% duty cycle
CLOCK

CLR Q CLR Q

Figure 39 Duty cycle generator.

78
Figure 40 Waveforms of duty cycle generator.

3.3.2 Design of trickle charging phase

Deeply discharged cells have reduced capacity and hence many battery management

systems have protection circuits in place to prevent cells from attaining a deep discharge

state. However, there are times when a cell might be deeply discharged and will have to

be charged to its full capacity. If the Li-ion battery is in a deeply discharged state and a

high current is initially used for charging, the battery could become damaged so

typically, a small current, about ten percent of full charge current, is used to charge the

battery until a moderate voltage is reached after which full charge current, 1C, can be

used.

The trickle charging phase is included in the design primarily for the preconditioning of

deeply discharged Li-ion cells. From table 6, a maximum of 20mA can be supported

79
during this phase. The trickle charge phase is shown in Fig. 41 and is only enabled

during normal mode if the following conditions exist: die temperature is below 85oC

(Tdie), input voltage is present (PG) and at a good level (the input voltage ranges from

3.3V to 3.6V), the battery is not in overvoltage state (OV) and the battery voltage is less

than a predefined threshold, VTS, which can be set by the user. VTS is also the minimum

voltage below which the battery becomes deeply discharged. Recommended V TS for the

OlevinPower 3.3V Li-ion battery [62] is 2V.

Since the trickle charge phase had to be enabled by the factors mentioned above, the

NOR gate can be used for implementation. Equation 18 shows that the trickle charge

phase will only be enabled (PMOS enable) if all inputs are low and will immediately be

disabled when any of the inputs go high.

𝑖 𝑒 𝑂

80
 EN

Vref
TC

PG
Trickle current enable
OV

ITC

Li-ion battery +
R1
RTC -
R2

Figure 41 Trickle charging phase block diagram.

The current source, ITC, is designed to have large output impedance and sized to handle

at most 25mA. Long channel length devices are used to reduce the effects of mismatch.

Resistors R1 and R2 set the VTS by using equation 19.

𝑅
( )
𝑅

If VTS is 2V,

𝑅 𝑅

Using commercial resistor values (5% tolerance),

The current into the Li-ion battery can be varied from 5mA to 20mA by changing the

resistance of the potentiometer, RTC. The trickle currents and its corresponding

81
resistance can be obtained from table 9. The error in current using commercial resistor

values is less than 5%. Fig. 42 shows the characterization of trickle current source IT

with respect to varying resistance values.

Table 9 Trickle currents and corresponding resistor values

Commercial Resistor value (5%) ITC

1.5KΩ 20mA

2.2KΩ 15mA

3.9KΩ 10mA

10KΩ 5mA

Figure 42 Trickle current, ITC, versus resistance, RTC.

82
Fig. 43 shows the signals OV, PG, trickle current enable, the output of the comparator,

TC_out, the battery voltage, Vbatt and then the trickle current. Modeling the battery as a

10mF capacitor in series with a 100mΩ resistor, the trickle current is set to 10mA and

VTS is 2V. From the figure, when Vbatt is less than 2V, the input power is good, PG, and

the battery is not in overvoltage, OV, the trickle charging phase is active and the battery

is charged with a 10mA current. Immediately, the battery voltage exceeds VTS, at point

A, the trickle phase is disabled.

Figure 43 Trickle charging phase operation signals.

83
3.3.3 Design of fast charging phase

Pulse charging decreases battery charge time at the same time increasing the life cycle of

batteries. It provides rest periods during which the battery electrolyte’s ions has the

opportunity to be distributed evenly in order to obtain better battery charge efficiency

[57]. This phase seeks to combine fast charging with increased battery charging

efficiency by using a large duty cycle, 0.75, initially to decrease charge time and a small

duty cycle, 0.25, to achieve better battery charging efficiency.

The process begins when the Li-ion battery reaches VTS. Charging at large current is

done during this phase. The fast charging phase is divided into two stages, i.e. charging

at 0.75D and charging at 0.25D. These two different duty cycles are generated using the

duty cycle generator (DCG) discussed. For this phase to be triggered, the battery should

not be in overvoltage state. Also, the die temperature, the temperature of the Li-ion

battery, the input voltage (PG) and input current (Iin) must all be within specifications. If

any of these parameters fall outside the specified range, the fast charging phase will not

be enabled. This is to protect the IC and the battery from damage. The designed pulse

charger has a maximum allowable charge current of 500mA during the fast charging

phase. This current is set by the current limited input adapter.

Once the fast charging phase is enabled, the battery voltage is determined. If the voltage

lies between VTS and VN, where VN is nominal voltage and can also be user set, the duty

cycle generator provides pulses at 0.75D to drive the PMOS switch, M1, which then

pulses the charge current from the input adapter into the battery at 0.75D. The output of

84
the SW comparator serves as the select signal for the multiplexer. Depending on whether

the battery voltage is below VN or not, 0.75D or 0.25D is selected to drive M1. Chemical

reactions in the battery are stabilized during rest periods allowing electrochemical

reactions to keep pace with the rate of input of electrical energy input [34]. This

increases power transfer rate and lowers charge time [55]. When the battery is

approaching full charge, i.e. above VN, M1 is driven at 0.25D, hence pulsing charge

current at 0.25D into the battery. The lower duty cycle is to ensure full charge is reached

without causing overvoltage concerns and also achieve better battery charging

efficiency. As long as the battery voltage is between VN and OV, the battery is charged

at 0.25D. In the case of the OlevinPower 3.3V Li-ion single cell battery, VN is 2.8V and

OV is 3.3V. VN is set by R3 and R4 by using equation 20. VTS and VN can be changed to

reflect what the user wants.

𝑅
( )
𝑅

If VN is 2.8V,

𝑅 𝑅

Using commercial resistor values (1% tolerance),

The operation of the fast charge phase can be summarized using equations 21-23. The

output of the DCG will pass to drive M1 as long as the mux_enable signal is high,

pulsing at 0.75D or 0.25D depending on the voltage of the battery, determined by the

output of comparator SW (SWout). The fast charging phase block diagram is shown in

Fig. 44.

85
( ) 𝑂

OV PG Tdie TS Current good

VIN

25% EN

Duty
Clock MUX M1
cycle
generator
75%
+
Battery
-
OV- Overvoltage Vref R3
PG- power good SW
Tdie-Die temperature
TS-external temperature sense
R4

Figure 44 Fast charging phase block diagram.

Fig. 45 shows the operation of the fast charging phase. Before, Vbatt reaches VTS, the

trickle charging phase is in operation. Once, Vbatt reaches VTS, point A in the figure, the

pulse charger switches to the fast charging phase, where a charge current of 125mA is

pulsed into the battery at 0.75D. The signal to the gate of the PMOS switch, M1, is

out_mux. When Vbatt reaches VN, point B, the charge current is pulsed at 0.25D and the

battery gradually approaches full charge. The OV signal goes high, point C, when the

86
battery attains full charge. This turns M1 and all unneeded comparators off. Input power,

PG, and die temperature, Temp, must be within specifications during this charging

phase.

Figure 45 Fast charging phase operating signals.

3.3.4 Protection schemes

The designed pulse charger includes protection schemes to ensure safe operation within

specified limits. It is very important to charge the battery within the specified limits and

at the same time ensuring the battery temperature is also within a safe operating region.

These protection schemes are external and internal temperature sense circuitry and

87
overvoltage and overcurrent protection. These circuits ensure that both the Li-ion battery

and IC are protected from abnormal conditions that can cause battery and circuit failure.

Fig. 46 shows the safe mode operating region of the battery and the circuit.

Vmax I_chargemax
Voltage (V)

Current(A)
0 45 85
Temperature (oC)

Battery safe zone Danger zone IC safety zone

Figure 46 Safe operating region of circuit and battery.

Temperature wise, the Li-ion battery must be charged within the safe operating

temperatures of 0oC to 45oC [72]. Also, the internal operating temperature of the IC, i.e.

die temperature, must not exceed the specification of 85oC. Overvoltage and overcurrent

protection ensures that the battery voltage and charging current respectively does not

exceed their specified limits. Fig. 22 also shows the danger region where the battery and

IC are most likely to fail when they operate outside the designed specifications.

88
3.3.4.1 Overcurrent protection

To prevent damage in case of the charge current exceeding a set threshold, a current

limiting block is added to the design. This block serves as an overcurrent protection

circuit. The overcurrent threshold is set by the R_LIM resistor which tries to limit the

charge current through M1 in case the charge current tries to exceed the preset threshold.

A M1
VIN BATT

X M3 B
M2
R_LIM

C1
DECISION
CIRCUIT
C2

VREF

Figure 47 Overcurrent protection circuit.

The current limiting circuit is shown in Fig. 47. The circuit consists of a current sensing

FET, M2, which is connected to M1 to mirror the charging current. M1 is 1500 times

M2. R_LIM is used to set the limiting current through M3 and M2, thereby setting the

overcurrent threshold. When the current through M1 is greater than 1500 times the

current through M2, the comparator, C1, comparing the nodes A and B outputs a signal

to the decision circuit to turn M1 off. This always ensures the charge current through M1
89
is below the preset threshold. The amplifier, C2, is used to force the voltage at node X to

be equal to VREF, hence the limiting current, I_LIM , is set by using equation 24. This

current can be user set and must not exceed 300µA.

Since M1 must handle a maximum current of 500mA. It has to be sized as such with

minimum on-resistance (1/gds) to reduce power consumption during operation.

| |
𝐿

If Vin=3.3V and gds<10Ω-1, then from equation 25,

( )
𝐿

Also,

( ) ( )
𝐿 𝐿

C2 is shown in Fig. 48. The one stage amplifier has a better gain than the regular single

stage amplifier comprising of a differential pair and current mirror load. This is because

the current mirror used as load for the differential pair has its output impedance boosted

by the current amplifier which consists of MP5, MP6, MN2 and MN3. The amplifier has

two feedback loops, i.e. a negative loop that decreases the impedance at node x and a

positive loop that increases the impedance, Z at node y. By including the current

amplifier with a gain of Ac, the overall transconductance, gm of C2 is increased and

90
hence the one stage amplifier can provide a relatively large gain of AcgmZ. This

amplifier is presented in [73].

MP1 MP2

IP

IN+ MP3 IN-

MP4
x y
MP5 MP6

OUT

IN
MN4 MN5

MN1 MN3
MN2

Figure 48 Error amplifier, C2, with boosted output impedance.

The amplifier gain is expressed as,

𝑜𝑡 𝑒 𝑖𝑛

where go is the output conductance of the transistor.

For if=10 and Ids=1µA, using equation 6,

The dimensions of the transistors can be calculated using equation 7,

( )
𝐿

91
The transistor dimensions are shown in table 10.

Table 10 Transistor dimensions of C2

Transistor (W/L)µm
MP1 (4/2)*4
MP2 (4/2)*4
MP3 (4/2)*4
MP4 (4/2)*4
MP5 (4/2)*1
MP6 (4/2)*1
MN1 (2/1.5)*2
MN2 (2/1.5)*2
MN3 (2/1.5)*2
MN4 (4/1.5)*2
MN5 (4/1.5)*2

3.3.4.2 Thermal protection

Thermal protection includes internal and external temperature sense circuitry. The

internal temperature sense circuitry monitors the die temperature of the IC to prevent it

from exceeding the predetermined limit of 85oC. This is to prevent the die temperature

from rising to a point where the IC can become damaged. When die temperature

increases, leakage currents increase which can cause the circuit to fail. Extreme

temperatures accelerate circuit aging and reduce circuit performance.

92
 VDD

Vref

Rp

RT1 O/P

Qn

RT2 Vbe

Figure 49 Temperature sense circuit.

Fig. 49 shows the temperature sense circuitry. The circuit consists of a bipolar NPN

transistor, Qn, whose base-to-emitter voltage, Vbe, has a negative temperature coefficient.

The resistors, RT1 and RT2, set the voltage at the base of Qn such that at a particular

temperature, Qn turns on and pulls the output, O/P, to ground. In normal operation, Qn is

off and O/P is pulled high through Rp. O/P can be used as a logic signal to either turn the

IC off or on. Using equations 27 and 28, RT1 and RT2 are preset to set the base voltage of

Qn at 0.484V which corresponds to a cut off temperature of 85oC.

where Ic is the collector current, Is is the reverse saturation current and is the

thermal voltage. For a constant Ic, the temperature dependence of Vbe is

93
To set the temperature at which the IC shuts off, i.e. 85oC, the Vbe of Qn at that

temperature must be determined.

| | ( ) |

where | is the base-emitter voltage at which the temperature sense circuitry

shuts off the IC, | is base emitter voltage at room temperature which is 0.6V and

and are cut off and room temperatures respectively.

At room temperature, Vbe is 0.6V, which implies at 85oC, Vbe will be equal to 0.484V.

Based on this voltage, RT1 and RT2 can be determined by using equation 29 when Vref is

1.2V.

𝑅
𝑅

Fig. 50 shows the output signal of the circuit across temperature. The signal goes low as

the temperature approaches 85oC. This signal is fed into the decision block of the pulse

charger to turn it off.

94
Figure 50 Output signal of temperature sense circuit.

External temperature sense ensures that the battery temperature is within safe operating

limits during charging. A pin is provided so that the user can connect a battery

temperature sense circuitry if needed. This external temperature sense circuitry senses

the ambient temperature, ensuring that any factor that might cause rise in battery

temperature such as internal battery fault or ambient temperature does not cause harm to

the battery. In cases where the battery is exposed to high temperatures, this serves as a

protection scheme to prevent the battery from charging when its electrolyte is too cold to

charge without damage or too hot to safely charge [74]. Thermistors can be used to serve

as external temperature sense circuits. They are temperature sensitive resistors which can

either have a positive temperature coefficient (PTC) or a negative temperature

coefficient (NTC). They convert sensed temperatures to voltage signals which can be

used to control the charger, i.e. to turn it off or on depending on battery condition.

95
3.3.4.3 Overvoltage protection

Overvoltage leads to overcharging which can lead to reduced life cycle of batteries.

Overcharging can also lead to thermal runaway. Battery thermal runaway is the rapid

repetitive increase in temperature of the battery due to overheating or overcharging

which can lead to fire or explosion. Heat generation during overcharge is almost

proportional to charging current [75]. This occurs due to increase in gases produced

during overcharge reactions in the battery [75] – [76]. Therefore, there must be measures

in place to prevent this from happening. In this design, the overvoltage protection circuit

serves two purposes, i.e. battery charge termination and overvoltage protection. Most

pulse charger designs use timers to terminate charge [59] – [61]. One problem with this

is that if the timers are not correctly programmed, overcharging can occur which can

accelerate the failure rate of the battery. To prevent this, battery charge termination and

overcharge are incorporated in the overvoltage protection block. Fig. 51 presents the

conceptual idea.

96
 Vbatt sense

Overvoltage
block Charge termination

disable
VIN charge
+
battery
-

Figure 51 Conceptual idea of overvoltage protection.

The comparator, OV, shown in Fig. 52 will monitor the output voltage of the battery and

once the threshold is reached, sends a signal to M1 to shut it off. Signals are also sent to

unneeded comparators and circuits to shut off in order to reduce power consumption.

ROV1
Li-ion battery

DECISION
Vref OV CIRCUIT M1
ROV2 Vref

disable
Vin

Figure 52 Overvoltage protection and termination.

97
The overvoltage threshold is set by resistors ROV1 and ROV2 by using equation 30. Based

on the specifications of OlevinPower 3.3V Li-ion single cell battery, the overvoltage

threshold is set to 3.3V.

𝑅
( )
𝑅

If VOV=3.3V, using commercial resistors (1% tolerance), ROV1=1.74MΩ and

ROV2=1MΩ.

3.4 The pulse charger

A detailed block diagram of the designed pulse charger with the relevant pins is shown

in Fig. 53. The use of each pin is shown in table 11. All the charging phases are

combined and simulation results are presented. The pulse charger can be operated as a

trickle or constant current charger to serve as an overnight or NiMH battery charger.

This brings versatility to the design.

98
 M1
VIN BATT

M3
R_LIM
M2 OV_set
DRIVER

C1
C2

VREF

EN
FC_set FC FC_out
VREF
VDD

D
SET
Q D
SET
Q
FC_EN
CLK
CLR Q CLR Q
DISABLE
Trickle_EN

PG_set TC_out
PG
VREF
EN
Trickle_set
INTERNAL TC 1
2
Trickle_set
VREF
TEMPERATURE 3

SENSE
VIN

TEMP_ext TP_set
Overvoltage_set
OV DISABLE
VREF

Figure 53 Detailed block diagram of pulse charger.

99
Table 11 Pulse charger pin descriptions and maximum ratings

Pin Usage Maximum

specification
VIN Connection to adapter 3.6V/500mA
R_LIM Set overcurrent threshold 300µA
FC_set Set nominal battery voltage Battery specification
CLK Clock input <1KHz
PG_set Input voltage monitoring 3.6V
TEMP_ext Battery temperature monitor Battery specification
OV_set maximum battery voltage 3.3V
Trickle_set Set trickle current 20mA
TC_out Set constant current charger 3.6V
Trickle_EN Enable trickle phase 3.6V
TP_set Set trickle phase Battery specification
FC_out Set constant current charger 3.6V
FC_EN Fast charge enable 3.6V
BATT Battery connection 3.3V

Fig. 54 shows the configuration of the pulse charger in normal mode. The 3.3V

OlevinPower Li-ion single cell battery specifications are used to set the parameters for

the pulse charger during simulation. The input voltage, overvoltage, fast charge and

trickle charge thresholds are set at 3.6V, 3.3V, 2.8V and 2V respectively. The charge

current is set to 155mA and a clock frequency of 1 KHz is used during simulation. The

Li-ion battery is modeled as a 100mF capacitor in series with a 100mΩ resistance. The

simulation results are shown in Fig. 55 and 56.

100
 VIN BATT
2.67MΩ 6.2KΩ
R_LIM

OV_set
2MΩ
1.74MΩ
FC_set FC_out

2MΩ FC_EN 1MΩ

CLK
Trickle_EN
PG_set TC_out
1MΩ
Trickle_set 820KΩ
Logic 1 3.9KΩ
TEMP_ext TP_set
1.2MΩ

Figure 54 Pulse charger in normal operation connection diagram.

Figure 55 Battery charging using pulse charger.

101
Fig. 55 shows the charging of the battery going through all the charge phases. When the

battery is less than 2V, it is trickled charged with a current of 11mA. When it reaches

this voltage, point B, the charge current, 155mA, from the current limited wall adapter is

pulsed into it at 0.75D. After, the nominal voltage is reached, 2.8V, the duty cycle

changes to 0.25 to allow increase in battery charge efficiency. It is noticed that fast

charging occurs between point B and C. The user can change the nominal voltage to suit

a fast charge application if needed. Battery charging is terminated when the battery

attains full charge at point D.

Figure 56 Waveforms showing operation of pulse charger.

102
Fig. 56 shows the relevant signals of the pulse charger during operation. It is seen that up

to point A, the Trickle_EN signal is high. This shows that the battery is in a deeply

discharged state. At point B, FC_out goes high to switch to a lower duty cycle. M1_gate

shows the driving signals which enable the switch to pulse current into the battery. From

point A to B, the pulses are at 0.75D and from point B to C, they are at 0.25D. Charge

termination occurs at point C when the OV signal goes high. This means the battery has

attained full charge. During this whole charging process, the PG signal is high and it is

constantly monitored. This shows that the input voltage is suitable for charging.

3.4.1 Transformation of pulse charger to trickle charger

To setup the pulse charger to work as a trickle charger the FC_EN pin and TC_out pins

are connected to ground. The CLK pin is connected to VIN. In this configuration, the

fast charge phase is off, which means M1 will not be turned on at any time. To prove the

concept works, the battery was modeled as a 100µF capacitor in series with a 10mΩ

resistor. OV was set to 3V with an input of 3.3V. The trickle current was set at 5mA.

Fig. 57 shows operation of the pulse charger in trickle charger mode.

103
Figure 57 Operation of pulse charger as trickle charger.

In trickle charger operation mode, the signal at M1_gate is equal to VIN, which ensures

M1 is never on. As long as the battery voltage is less than 3V, the charge current is set to

5mA and the battery is charged till it reaches point B. At this point, full charge is

attained, the Trickle_EN goes low to turn off the internal current source, IT. OV goes

high to terminate charging completely.

3.4.2 Transformation of pulse charger to constant current charger

Constant current charging can be used for NiMH batteries. This configuration extends

the functionality of the pulse charger to charge another type of battery apart from the Li-

ion. To configure the pulse charger to work as a constant current charger, the

Trickle_EN and FC_out are connected to ground, while the TC_out and CLK pins are

connected to VIN. These connections force the trickle charging phase to be non-

104
operational and M1 to be always on during charging. To prove the concept works, the

battery is still modeled as a 100µF capacitor in series with a 10mΩ resistor. OV was set

to 3V with an input of 3.3V. The constant current was set at 155mA. Fig. 58 shows the

simulation results of this configuration.

Figure 58 Operation of pulse charger as constant current charger.

Fig. 58 shows the relevant signals of the pulse charger when operating as a constant

current charger. M1_gate signal shows that the main switch is on until full charge is

attained. Comparing to Fig. 56, there is no pulsing. This is because it is in constant

current mode of operation. At point A, the charger senses the battery voltage and begins

charging with a constant current of 155mA which reduces as full charge is approached.

105
At point B, battery has reached full charge, OV goes high and charging is terminated. It

is noticed that the trickle phase is off during the entire charging process.

106
4. STEP DOWN DC-DC CONVERTER AND SWITCHING NETWORK SYSTEM

For the system to support energy harvesting applications, additional blocks have to be

designed. These blocks are the step down DC-DC converter and a switching network

system. The step down converter serves as a DC supply for the energy harvesting system

(EHS) while the switching system creates a path for the battery to be charged from the

EHS. The switching network system also provides an option for the battery and the

supercapacitor to complement each other in supplying power to the load. These two

blocks together with the pulse charger form the battery charging system. This section

looks at the different types of DC-DC step down converters, the reason why a particular

topology is chosen, a detailed design procedure of the converter and simulation results

showing the operation of the converter. The flow chart detailing the operation of the

switching network system and the transistor level design is discussed. Finally, simulation

results are presented. The EHS is considered to be a black box, providing a certain

output current and voltage, and requiring a certain DC supply voltage. The specifications

of the EHS are presented in table 12.

Table 12 Specifications of the EHS

Parameter Specification
Supply voltage /current 1.8V/10mAmax
Output voltage/current 3.3V/200µAmax

107
4.1 Step down DC-DC converter

Energy harvesting systems with no start up circuit need an initial DC supply to kick start

the system [77]. This is because the voltage harvested, by the energy harvesters, is very

small and is not able to power up the CMOS circuits in the system. Hence, an initial DC

voltage is needed to kick start the CMOS circuits into operation. This initial DC supply

is provided by a battery but in most cases, the battery voltage is too high for the

operation of the system and needs to be reduced to a level which can be used by the

EHS. The circuit that performs this step down function is known as a step down DC-DC

converter. Step down DC-DC converters are used in almost any portable electronic

device. They are required because different sub circuits in devices may require different

voltage levels for operation. During circuit operation, the battery voltage is likely to

diminish, hence to allow constant operation of sub circuits during diminishing battery

voltage level, step DC-DC converters provide output voltage regulation, which allows

the supply voltage to remain constant under different load conditions and varying battery

voltage.

The main types of step down DC-DC converters are linear regulators, switched inductor

and switched capacitor converters [78]. Linear regulators, e.g. low-dropout (LDO)

regulator, are very popular due to its simplicity and low cost. However, they are

inefficient because the power transistor, Tp, used to drop the input voltage, Vin, to the

required level, Vo, is always on and at high load currents, heat dissipation is high and

hence, power consumption goes up. The LDO is shown in Fig. 59.

108
 VIN
L

Vref _ Ts
Tp C
+ DC LOAD

Vo
R1
Buck converter

LDO R2

Figure 59 Linear regulator (LDO) and switch inductor converter (buck).

Switched inductor converters, such as buck converters, offer high efficiency at increased

cost and complexity. Increased cost is due to additional external components, i.e.

inductors and capacitors, which take up board space, and complexity due to the control

loop needed to achieve good output voltage regulation. The main switch, Ts, is not

always on, but switching constantly depending on output voltage levels to ensure

reduced power consumption. The buck converter is shown in Fig. 59.

Switched capacitor converters combine the benefits of linear regulators and switched

inductor converters, i.e. simplicity and high efficiency respectively. A switched

capacitor converter consists of a series of switch networks and capacitors, which can be

connected in different configurations, to obtain different output voltages from one input

voltage.

109
Table 13 shows the comparison between the different step down converters in terms of

area and power. For the purposes of in this thesis, a switched capacitor converter is

chosen.

Table 13 Comparison of different step down converters

Converter type Area Quiescent power

LDO small high

Buck huge low

Switched capacitor moderate low

4.1.1 Switched capacitor converter

For smaller board space due to few external components, lower power consumption due

to the switching action of the converter and low current application, the switched

capacitor converter is chosen over the other step down DC-DC converters. It combines

the benefits of the switched inductor converter and the linear regulator, making it an

ideal choice for the step down function in this project. There are also no magnetic

components, hence less noisy.

The switched capacitor converter can be designed to have different gain configurations

[79] – [80]. Based on the supply voltage specification for the energy harvesting system,

the switched capacitor circuit should be able to step the battery voltage down from 3V or

3.3V to 1.8V. To do this, the converter makes use of two flying capacitors, C1 and C2,
110
which is used to store and transfer energy to the output capacitor, C out. By detecting the

output voltage, the control loop in Fig 60 is able to generate pulses at Ø1 and Ø2 to

produce a gain setting of 2/3.

Vo

Vout SENSE Ø1
Ø1 Ø2 Control
C1 Cout
Vbatt
Ø2

Ø2 Ø1

Ø1
C2

Ø1
Ø2

Figure 60 Switched capacitor converter.

In Fig.61, at the end of Ø2, the flying capacitors, C1 and C2, are connected in series

together, and in parallel with Cout. During Ø1, they are connected in parallel, and in

series with Cout. These series-parallel configurations are made possible by the control

loop which controls the different clock phases, i.e. Ø1 and Ø2.

111
 Vbatt
Vo Vo
Vbatt C1 Cout Cp Cout

C2
At Ø1

Vo
Vo

C1 Cout Cs Cout

C2 At Ø2

nT nT+T/2 nT+T

Figure 61 Operation of switched capacitor converter during different clock phases.

The analysis of this switched capacitor converter based on charge conservation [78] is

presented in equations 31 to 43.

At the end of Ø2, i.e. time nT, capacitors Cs and Cout have the following charge

equations.

𝑛 𝑛

𝑛 𝑛

At the end of Ø1,

(𝑛 ) (𝑛 )

112
 (𝑛 ) ( (𝑛 ))

The theory of the law of charge conservation states if two capacitors are connected

together, the total charge on the parallel combination is equal to the sum of the original

charge on the capacitors [81]. Based on this theory, using equations 31 to 34,

(𝑛 ) (𝑛 ) 𝑛 𝑛

(𝑛 ) ( (𝑛 )) 𝑛 𝑛

(𝑛 ) 𝑛

Therefore,

(𝑛 ) 𝑛

(𝑛 ) 𝑛

At time nT+T, using the law of conservation at Ø2,

𝑛
𝑛 (𝑛 ) (𝑛 ) 𝑛

𝑛 𝑛
( ) ( )

If ( )
and 𝑛 ( )
, then

𝑛 𝑛 𝑛

At nT+2T,

113
 𝑛 𝑛 𝑛

𝑛 𝑛( 𝑛 𝑛 ) 𝑛 𝑛 𝑛

At nT+3T,

𝑛 𝑛 𝑛 𝑛 𝑛

At nT+KT, as K varies from 0 to infinity

𝑛 𝑛 𝑛 𝑛 𝑛 𝑛 𝑛

𝑛
𝑛 ( ) 𝑛 𝑛
𝑛

where 𝑛 ( )
.

Hence as ,

𝑛 ( )
𝑛

Substituting m and n into equation 42,

𝑛
( )
( )
( ( ) )

Then, if

The efficiency, η, of the switched capacitor circuit [78] is approximated as,

114
This efficiency is maximized when the battery is at 2.7V for an output voltage of 1.8V.

If the battery voltage is more than 2.7V, the efficiency of the converter reduces and the

output voltage is obtained with an error. For the purposes in this thesis, the battery

nominal voltage cannot exceed 3V and hence at that voltage, the converter can output a

voltage with an error that does not exceed twelve percent.

4.1.1.1 Transistor-level implementation

Fig. 62 shows the transistor level implementation of the switched capacitor converter

with a gain of 2/3. There are two 1µF flying capacitors and seven switches, five of which

are PMOS and the other two, NMOS. During Ø1, P1, P2, P4 and P5 are on and N1, N2

and P3 are off. The vice versa occurs during Ø2. This switching action of these

combinations of switches, which are driven by control signals, Ø1 and Ø2, produce an

output gain of 2/3.

115
 Vo
Ø1 Ø2

P1 P3
Cout Rs1
Vbatt C1 +
P4 Rs2 Vref Clock
-
Ø1
N1
P2
Ø1

C2
Ø1
P5

N2

Ø1
control

Ø1

Ø2

Figure 62 Transistor level implementation of switched capacitor converter.

Both the NMOS and PMOS switches are designed to have low on resistance (ron) in

order to reduce voltage drop and improve efficiency during operation. Equation 45

details how the PMOS switches are sized.

If ron<1Ω, then gds < 1Ω-1.

| |
𝐿

If Vin=3V and gds<1Ω-1, then

( )
𝐿

Ø1 and Ø2 are generated by the control loop, which consists of an output voltage

detector and a non-overlapping clock. The output voltage detector consists of a

comparator and a resistor feedback network. The resistors detect the voltage level of the

116
output and the comparator outputs a high or low logic signal depending on whether the

detected output voltage is less than or equal to 1.8V. The output of the comparator is

combined with a clock signal using a NOR gate to produce a signal to control the non-

overlapping clock generator. A non-overlapping clock is used to ensure that Ø1 and Ø2

do not have the same logic value at the same time.

To determine the values of the resistors, Rs1 and Rs2, the following equation is used,

𝑅
( )
𝑅

If Vo is 1.8V,

𝑅 𝑅

Fig. 63 shows the operation of the switched capacitor circuit at no load conditions.

Figure 63 Operation of switched capacitor converter with Vbatt=3V at no load.

117
It is seen from Fig. 63, that with a clock frequency of 1MHz, the control signal to the

non-overlapping clock generator is continuously pulsing until the output reaches 1.8V

for a battery voltage of 3V at no load conditions. Cout gradually charges during this

period till it achieves 1.8V.

Fig. 64 shows the variation of the output voltage with load current. As load current

increases, the converter’s output voltage drops, indicating that it is able to source a load

current less than 30mA. Since the application is for low load currents, this does not pose

a problem because the output voltage remains constant when operating under the

required specifications of the EHS, i.e. a maximum of 10mA.

Figure 64 Converter’s output voltage variation with load current.

118
Fig. 65 shows the variation of the output voltage of the converter under varying battery

voltage. As voltage decreases, the output voltage also decreases. The converter’s output

voltage will remain stable as long as the battery voltage is above 2.7V.

Figure 65 Converter output variation with varying battery voltage.

The efficiency of the converter is shown in Fig. 66. As load current increases, the

converter’s efficiency starts decreasing. This is because the converter cannot support the

current demand, i.e. above 20mA, and the output voltage drops, causing efficiency to

also decrease.

119
 Efficiency vs load current
100

80
Efficiency (%)

60

40

20

0
0 5 10 15 20 25 30 35
Load Current(mA)

Figure 66 Efficiency of switched capacitor converter at 3V input.

4.2 Switching network system

The current sourcing capability of the EHS is not enough to meet the power

requirements of the pulse charger. The specifications of the EHS, shown in table 12,

indicate that, it can provide a maximum of 200µA at the output. The pulse charger needs

at least 1mA to operate. To prevent reduced efficiency, the switching network system is

designed to bypass the pulse charger and enable battery charging directly from an EHS.

It also has the option to enable the supercapacitor, being charged by the EHS, to deliver

short term peak current to a load when needed without affecting the battery, thereby

enabling both the battery and supercapacitor to complement each other. This block

120
ensures that the EHS and the pulse charger operate independently from each other and

can each be used as a standalone device.

A battery-supercapacitor combination allows peak current pulses to be delivered to a

load whenever needed and at the same time sustain long term load power demand. The

supercapacitor is the main power supply to the load while the battery serves as a backup.

When AC/DC input or USB is not available, the battery can be charged from the EHS

and when the supercapacitor is not able to support the load, the battery can provide

power to the load. This switching network being part of the battery charging system

provides options to make the system adaptable to load and input power conditions.

4.2.1 Implementation of the switching network system

The switching network will only be connected to the battery when AC/DC input or USB

is not available. When the block determines that these inputs are not available, the EHS

is connected to battery to charge it but it they are available, the EHS is connected to the

supercapacitor. Depending on load conditions, there is an option for the supercapacitor

or battery to be connected to the load. The flow chart summarizing the operation of the

switching network system is presented in Fig. 67.

121
 EHS present

Yes Yes

Vsc>OV Vbatt<OV

Yes
Yes

EHS charge
battery

No

No Supply load

EHS charge
Supercap

Figure 67 Flow chart showing conceptual idea of switching network system.

The conceptual block diagram of the switching network is presented in Fig. 68. The

voltage detector is implemented using a clocked comparator and a series of resistor

feedback networks. Depending on battery and supercapacitor voltage conditions, the

EHS charges the battery or the supercapacitor.

122
 Voltage detector

charge
EHS charge
+
_

Figure 68 Conceptual block diagram of switching network system.

A clocked comparator, shown in Fig. 69, is used in order to minimize power

consumption since the battery serves as its supply voltage. This is to ensure minimal

current drain from the battery to the comparator. The comparator is clocked at a

frequency of 1MHz.

MP1 MP2 MP3 MP4

OUT

MN7
OUT
IN+ IN-
MN1 MN2
Static latch
IN CLK

MN4 MN3 MN5 MN6

Preamplifier Regenerative Latch

Figure 69 Clocked comparator.

123
The comparator consists of a preamplifier stage, regenerative latch and a static latch, to

serves as a memory to hold the values for a whole clock period. The static latch also

serves as an interface to convert the analog signals into digital ones since connecting the

analog outputs directly can cause the system to be unstable [82]. The regenerative latch

(MN5, MN6, MP3, MP4 and MN7) amplifies the analog signals to full rails while the

preamplifier stage (MN1, MN2, MP1 and MP2) amplifies the input signal to improve the

sensitivity of the comparator [83].

The differential pair transistors, MN1 and MN2, are sized for minimum noise, mismatch

reduction and moderate gain. To reduce parasitics, the transistors sizes used for the

regenerative latch are kept minimum. The gain, Avp, for the preamplifier stage and that

for the regenerative stage, Ar, are respectively

where gm is transconductance and R is the output impedance.

To size the pre-amplifier based on low average power consumption, if=1 is chosen and

an Ids of 500nA. Using equation 6,

The dimensions of the transistors can be calculated using equation 7.

( )
𝐿 √ 𝑖

The dimensions of the transistors used in the comparator are presented in table 14.

124
Table 14 Clocked comparator transistor dimensions

Transistor W/L(µm)
MN1 (4/2)*2
MN2 (4/2)*2
MN3 (4/2)*4
MN4 (4/2)*4
MN5 (4/0.4)*1
MN6 (4/0.4)*1
MN7 (4/0.4)*1
MP1 (4/2)*6
MP2 (4/2)*6
MP3 (4/0.4)*1
MP4 (4/0.4)*1

Fig. 70 shows the output of the clocked comparator. As the signal crosses the reference

voltage, the output logic switches between one and zero depending on whether the signal

is greater or less than the reference voltage.

Figure 70 Operation of clocked comparator.

125
The switching network block enables switching between input charging source and the

two storage elements depending on input source availability and load conditions. The

transistor level implementation is presented in Fig. 71.

PG A

Vref -
C +
-
+
EHS
OUTPUT Load
control
A + Rb1
Super Rsc1 B batt
Cap -
Rsc2 Rb2

+ -

Vref

Figure 71 Transistor level implementation of switching network system.

From Fig. 71, switch A is turned on, when the supercapacitor voltage is below a set

threshold, e.g. 3V, to allow the supercapacitor to be charged. When the battery is not

fully charged, but the supercapacitor is, switch C turns on to allow the battery to be

charged from the EHS. This depends on whether the AC/DC input is available or not

(PG input), if it is not available, charging from the EHS will proceed or vice versa.

Signal B serves to connect the supercapacitor or battery to the load based on which

storage element is ideal for the particular application. The comparators are included to

126
prevent overcharging of the battery and the supercapacitor. Sizing of the switches are

based on a small ron by using equation 45.

If ron<1Ω, then gds < 1Ω-1. Therefore, for a Vin=3.3V and gds<10Ω-1,

( )
𝐿

To determine the values of the resistors, Rsc1, Rsc2, Rb1 and Rb2, the following equation is

used,

𝑅
( )
𝑅

If Vbatt,sc is 3V,

𝑅 𝑅

It should be noted that the EHS operates independently of the pulse charger as long as

AC/DC or USB inputs are available.

Fig. 72 shows the operation of the switching network system when the pulse charger is

off. The priority of the EHS is to charge the supercapacitor to 3V. During this period,

switch A (SW_A) is on and all other switches are off. When the supercapacitor is fully

charged and the system detects that the battery voltage is below the overvoltage

threshold, the EHS proceeds to charge the battery, turning on switch C (SW_C). Once

the supercapacitor voltage falls below 3V, the EHS goes back to charging it. This is seen

by observing the switching action of switch A. The system is designed in such a way that

127
none of the switches are on at the same time to isolate each power source to provide

reverse current protection.

Figure 72 Operation of switching network system.

128
 5. EXPERIMENTAL RESULTS

This section seeks to provide results from the designed IC operating under real life

conditions. The IC was designed using the IBM 0.18µm process with the area of the die

being 4mm2. It was packaged using an open cavity 8mm by 8mm Quad Flat No-Lead

(QFN). The experimental results of the proposed pulse charger indicating proof of

concept, the step down DC-DC converter and the switching network system are

presented. Comparison of previous works versus the proposed pulse charger in this

thesis is made. A micrograph of the designed chip is shown in Fig. 73.

Converter

2mm

Switch
networK Pulse charger

2mm

Figure 73 Micrograph of user programmable battery charging system IC.

129
Fig. 74 shows the instruments used in the testing and measurement of the proposed

charging system. It provides a pictorial representation of how the experimental setup

looks like.

Oscilloscope

Voltage source

Signal generator
Proposed
battery charging
system
Multimeter
Li-ion/NiMH

Figure 74 Pictorial representation of experimental setup.

5.1 Proposed pulse charger

Since pulse charging is applicable to NiMH, Li-ion and Lead acid batteries. To prove the

proposed pulse charger works in normal mode, it is used to charge a 2500mAh NiMH

battery. A current limited input voltage source of 3.3V/300mA is used as the source. A

clock signal of 1KHz is used. The schematic of the test bench used to verify the

operation of the proposed pulse charger is shown in Fig. 75.

130
 oscilloscope
3.3V
VIN BATT
120KΩ 6.2KΩ 2500mAh NiMH
R_LIM

OV_set
100KΩ
150KΩ
FC_set FC_out

Clock(1KHz)
180KΩ FC_EN 1MΩ
CLK oscilloscope
Trickle_EN
PG_set TC_out
100KΩ
Trickle_set 67KΩ
3.9KΩ
TEMP_ext TP_set
100KΩ

Figure 75 Test bench to verify operation of proposed pulse charger.

Fig. 76 shows the charging of the 2500mAh NiMH battery using the proposed pulse

charger with the above setup. It took about 85 minutes to attain a nominal voltage of

2.5V.

131
 Pulse charging a NiMH
2.5

2
battery voltage (V)

1.5

1
0 10 20 30 40 50 60 70 80 90
charge time (min)

Figure 76 Test result verifying operation of proposed pulse charger in normal mode.

To verify that the pulse charger can be reconfigured to operate as a trickle charger, it was

used to charge a 3V Panasonic Li-ion coin cell. It can be observed from Fig. 77 that,

operating in trickle charge mode takes a very long time for the battery to attain full

capacity. In this case, the Li-ion coin cell took about 400 minutes to attain a maximum

voltage of 2.92V. It can therefore be used as an overnight charger.

132
 Trickle charging

2.5
battery voltage (V)

1.5

1
0 50 100 150 200 250 300 350 400 450

charge time (min)

Figure 77 Trickle mode charging of Li-ion cell using reconfigured pulse charger.

The pulse charger is reconfigured to operate as a constant current charger and used to

charge a 2000mAh NiMH rechargeable battery to its nominal voltage of 2.5V when the

constant current was set to 300mA, which is about 15% of the required full charge

current. It is noticed from Fig. 78 that to charge to the nominal voltage of 2.5V, it took

about 220 minutes. More current could have been used to charge the NiMH but with

maximum constraints on the allowable charge current to prevent damage to the IC, the

charge current was limited to 300mA.

133
 Constant current charging

2.4

2.2
battery voltage (V)

1.8

1.6

1.4

1.2
0 50 100 150 200 250
charge time (min)

Figure 78 Constant current charging using reconfigured pulse charger.

5.2 Switched capacitor converter

The schematic of the test bench used to verify operation of the switched capacitor

converter is shown in Fig. 79. The converter is tested for output variation during input

voltage variation and efficiency. The resistors, 50KΩ and 100KΩ, are what sets the

output voltage to 1.8V. The values can be changed to obtain a different output. The

output and flying capacitors are set to 1µF. A 1MHz clock signal is used.

134
 oscilloscope
Out
Vbatt
+ 50KΩ 1µF

- Out_set

C1+ 100KΩ
1µF

C1-
C2+

1µF
Clock(1MHz)
C2-
CLK
oscilloscope
Control

Figure 79 Schematic of switched capacitor test bench.

1µF capacitors are used for the flying capacitors and output capacitors. The smaller the

output capacitance, the larger the output voltage ripples, as seen in Fig. 80. Increased

output capacitance reduces the ripple significantly as Fig. 81 shows.

Output voltage ripple (1µF)

Control

Figure 80 Control signal for clock phases and output voltage ripple of converter.
135
Figure 81 Converter's output with a 3V input and increased output capacitance.

Output voltage versus battery voltage

3

2.5

2
Output voltage

1.5

Simulation
0.5 Experimental

0
2.5 2.6 2.7 2.8 2.9 3 3.1 3.2 3.3
Battery voltage

Figure 82 Output voltage variation versus input voltage variation.

136
Fig. 82 shows the variation of the output voltage as input voltage varies. Experimental

results are very similar to simulated results. Fig. 83 shows the efficiency of the

converter. As load current increases, the efficiency decreases, but since the converter is

for low current application, it operation will be about 90% efficient at an input voltage of

3V.

Efficiency vs load current

100

90

80

70

60
Efficiency (%)

50

40

30
Simulation
Experiment
20

10

0
0 2 4 6 8 10 12
Load current (mA)

Figure 83 Efficiency of converter.

5.3 Switching network system

To test the operation of the switches in the switch network system, a schematic of the

test bench used is shown in Fig. 84. The maximum voltages of the battery and

137
supercapacitor (SC) are to 3V. This is done by setting the resistances to 150KΩ and

100KΩ. A current limited input voltage source of 3.3V/200µA is used to characterize the

EHS. The supercapacitor is set to 10µF while the battery is emulated using a 300µF

capacitor. SW_A is the switch between the EHS input and supercapacitor, SW_B is the

switch between the supercapacitor and battery, which only turns on when there is a load

present, and SW_C is the switch between the EHS input and battery.

Batt_connect
EHS input
150KΩ
+
battery
3.3V/200µA -
Set_batt_OV

SC_connect 100KΩ
150KΩ 10µF

Clock(1MHz)
Set_SC_OV

CLK 100KΩ
SW_A SW_B SW_C

Figure 84 Schematic of switching network system test bench.

138
 Control signals to switching newtork
4
Switch C
Switch A
3.5 Switch B

2.5
Control signals (V)

1.5

0.5

-4 -3 -2 -1 0 1 2 3 4
-6
Time (sec) x 10

Figure 85 Operation of switching network for supercapacitor and battery < 3V.

Fig. 85 shows the operating signals to the switches during operation of the switching

network system. The EHS is charges the supercapacitor intermittently, each time to 3V

and switches to charge the battery (SW_C) until it attains a voltage of 3V. After this

happens, the EHS is connected constantly to the supercapacitor (SW_A). The signals to

the switches are slewed, mainly because of the large input capacitances of the switches.

Signal B is high mainly because there is no load attached to the system. Fig. 86 presents

the view from an oscilloscope.

139
Figure 86 Operating signals of switching network system.

5.4 Comparison with previous works

The proposed charging system in this thesis varies in comparison with previous works

[56] - [60]. In [56] - [58] in table 15, a pre-charge phase is not indicated in the design.

This is because an assumption was made that the battery will not go into a pre-charge

state. [59] - [60] have a pre-charge phase but it is not user programmable. In the

proposed charging system designed in this thesis, this phase can be programmed by an

external resistor to provide a maximum current of 20mA. Options to charge the battery

from a USB network, wall adapter or an EHS provides flexibility, compared to the other

previous works which can charge a battery from either a wall adapter or USB.

140
Table 15 Comparison of proposed charger with previous works

VFPCS DVVPCS[ PLL LTC4052 This work

[57] 58] based /LTC1730
[56] [59],[60]
Input source AC/DC AC/DC AC/DC AC/DC AC/DC or
USB or
EHS
Battery type Li-ion Li-ion Li-ion Li-ion Li-ion &
NiMH

Pre-charge - - - Predefined User set

phase trickle
current
Over-voltage - - - Predefined User set
protection maximum
cell voltage
Over - - - Present present
temperature

Charge current Average Average Average User User

current current current defined defined
(wall (wall
adapter adapter
dependent) dependent)
Implementation software software discrete analog analog

Modulation PFM- PWM- - Predefined PWM-user

type predefined predefined set

Cell maximum 4.2V 4.2V 4.2V 4.2V 3.3V

voltage

User No No No No Yes
reconfigurable

141
Also, [56] - [58] were implemented using software or discrete electronic components

and pulses voltage into the battery. This work is implemented on silicon and pulses

charge current into the battery similar to what is in [59] - [60]. With [56] - [58] charging

a 4.2V battery, a supply voltage of 5V was used. In this work, the maximum battery

voltage is 3.3V and hence, the input voltage range is limited to a maximum 3.6V to

prevent oxide damage. Also, in all the previous works listed above, the modulation types

are all predefined and the system designed for a Li-ion battery. The user does not have

control over it to change charge profile while in this work, the user has control of all

thresholds and the control of the charging algorithm. The system designed in this thesis,

as compared to the previous work is completely versatile. It has been verified that it can

be reconfigured as a constant current charger to charge a NiMH battery and as a trickle

charger to serves as an overnight charger and possible for a battery to be charged via an

EHS.

142
 6. CONCLUSIONS

The design of a user programmable battery charging system has been presented in this

thesis. The performance of this charging system has been verified through simulation

and further proved through experimental results. The concept of the pulse charger

operation ensures that it can be reconfigured to be used in different applications. The

step down DC-DC converter can provide a stable output voltage for low current

applications, at voltages above the nominal battery voltage, with high efficiency values.

The switching network system further increases the number of applications in which the

charging system can be used, by proving the concept that a battery can be charged via an

energy harvesting system.

In summary, this proposed charging system could easily be used in many applications,

ensuring low cost and convenience. Not purchasing different types of chargers for

different applications can reduce cost and ensure that, at any time and place, a battery

could be charged using a wall outlet, USB or an energy harvesting system depending on

which one is available. This provides convenience for users. Further research on this

charging system should focus on making the pulse charger adaptable to various

conditions. Increasing the efficiency of the step down DC-DC converter, by

implementing different gain configurations, to match changes in varying battery voltage

should also be looked at. More emphasis should be placed on expanding the application

of the switching network system to account for battery undervoltage and varying load

and input conditions.

143
 REFERENCES

[1] Global Industrial Analysts, “Consumer Batteries: A Global Strategic Business

Report,” [Online]. Available:

http://www.prweb.com/releases/consumer_batteries/primary_secondary/prweb86

05940.htm, August 29, 2011.

[2] Stan D’Souza, “Microchip: Transformerless Power Supply,” [Online]. Available:

http://www.jimfranklin.info/microchipdatasheets/91008b.pdf, February 11, 2011.

[3] John Woodgate, “Line Filter Capacitors,” [Online]. Available:

http://my.execpc.com/~endlr/line-filter.html, February 11, 2011

[4] Sebastian Anthony, “How USB Charging Works, or How to Avoid Blowing Up

Your Smartphone,” [Online] Available:

http://www.extremetech.com/computing/115251-how-usb-charging-works-or-

how-to-avoid-blowing-up-your-smartphone, January 24, 2012.

[5] Engineering Edge, “USB Cables/Interface Overview” [Online]. Available:

http://www.engineersedge.com/computer_technology/usb_cables.htm, February

10, 2013.

[6] Universal Serial Bus Technology, “Battery Charging v.1.2 Spec and Adopters

Agreement,” [Online]. Available:

http://www.usb.org/developers/devclass_docs/BCv1.2_011912.zip, February 10,

2013.

144
[7] Kema Inc., “Market Evaluation for Energy Storage in the United States,” [Online].

Available: http://www.copper.org/about/pressreleases/pdfs/kema_report.pdf,

January, 2012.

[8] David Connolly, “A Review of Energy Storage Technologies for Integration of

Fluctuating Renewable Energy,” [Online]. Available:

http://dconnolly.net/files/A%20Review%20of%20Energy%20Storage%20Techn

ologies.pdf, October 11, 2010.

[9] Jim Eyer and Garth Corey, “Energy Storage for Electricity Grid: Benefits and Market

Potential Assessment Guide,” Sandia Report [Online]. Available:

http://sgstage.nrel.gov/sites/default/files/resources/energy_storage.pdf, February,

2010.

[10] IEC Market Strategy Board, “Electrical Energy Storage,” [Online]. Available:

http://www.iec.ch/whitepaper/pdf/iecWP-energystorage-LR-en.pdf, December,

2011.

[11] Electropaedia, “Fuel Cell,” [Online]. Available:

http://www.mpoweruk.com/fuel_cells.htm, February 11, 2013.

[12] Office of Basic Energy Sciences, DOE, “Basic Research Needs for Electrical

Energy Storage,” [Online]. Available: http://web.anl.gov/energy-storage-

science/publications/EES_rpt.pdf, July 2007.

[13] H.F Gibbard, “Nickel Metal Hydride Battery Applications,” Proceedings of the

Ninth Annual Battery Conference on Applications and Advances, January, 1994.

145
[14] Electropaedia, “Battery Applications,” [Online]. Available:

http://www.mpoweruk.com/applications.htm, February 11, 2013.

[15] Texas Instruments Inc., “Using NiMH and Li-ion Batteries in Portable

Applications,” [Online]. Available: http://www.ti.com/lit/an/slua015/slua015.pdf,

February 26, 2013.

[16] Interstate Powercare, “Motive Power Batteries,” [Online]. Available:

http://interstatepowercare.com/PDF/pc_io_manual_2007.pdf, February 26, 2013.

[17] Grigorii L Soloveichik, “Battery Technologies for Large-Scale Stationary Energy

Storage” Annual Review of Chemical and Biomolecular Engineering, Vol. 2, no.

5, pp. 503-527,2011.

[18] H. Oman, “Aerospace and Military Battery Applications,” IEEE Aerospace and

Electronic Systems Magazine, Vol.17, no. 10, October 2002.

[19] The Boston Consulting Group, “Batteries for Electric Car, Challenges,

Opportunities and Outlook to 2020,” [Online]. Available:

http://www.electricdrive.org/index.php?ht=a/GetDocumentAction/id/27906,

January, 2010.

[20] Dan Leistikow, “An Update on Advanced Battery Manufacturing,” [Online].

Available: http://energy.gov/articles/update-advanced-battery-manufacturing,

October 16, 2012.

[21] East Penn Manufacturing Co. Inc.,“Battery Applications,” [Online]. Available:

http://www.dekabatteries.com/assets/base/0149.pdf, January 31, 2007.

146
[22] West Marine, “Selecting a Marine Storage Battery,” [Online]. Available:

http://www.westmarine.com/webapp/wcs/stores/servlet/WestAdvisorView?langI

d=-1&storeId=11151&page=Selecting-a-Marine-Storage-

Battery#.URlMm2cbBSQ, February 11, 2013.

[23] Department of Defense, “Unmanned Systems Integrated Roadmap: FY2011-2036,”

[Online]. Available:

http://www.defenseinnovationmarketplace.mil/resources/UnmannedSystemsInte

gratedRoadmapFY2011.pdf, February 11, 2013.

[24] Energizer, “Battery Internal Resistance,” [Online]. Available:

http://data.energizer.com/PDFs/BatteryIR.pdf, December, 2005.

[25] Learning about Electronics, “Battery Internal Resistance,” [Online]. Available:

http://www.learningaboutelectronics.com/Articles/Battery-internal-resistance,

February 11, 2013.

[26] Energizer, “Energizer L91,” Energizer [Online]. Available:

http://data.energizer.com/PDFs/l91.pdf, February 11, 2013.

[27] Jaycar Electronics,“Battery Terms and What They Mean.” [Online]. Available:

http://www1.jaycar.com.au/images_uploaded/battglos.pdf, September 11, 2012.

[28] Green Batteries, “Glossary of Battery Terms.” [Online]. Available:

http://www.greenbatteries.com/batteryterms.html, September 11, 2012

[29] Corrosion Doctors, “Self-Discharge of Batteries,” [Online]. Available:

http://corrosion-doctors.org/Batteries/self-compare.htm, February 12, 2013

147
[30] Battery University, “Global Battery Markets.” [Online]. Available:

http://batteryuniversity.com/learn/article/global_battery_markets, February 25,

2013

[31] Davide Andrea, “Battery Management Systems for Large Lithium-Ion Battery

Packs,” 1st ed. Artech House Inc., 2011

[32] Infiniti Research Limited, “Global Lithium Battery Market 2011-2015,” [Online].

Available: http://www.reportlinker.com/p0962689-summary/Global-Lithium-

Battery-Market.html, August, 2012.

[33] Pike Research, “Revenue for Lithium Ion Battery Market Set to Grow by 700% by

2017,” [Online]. Available: http://www.prnewswire.com/news-releases/revenue-

for-lithium-ion-battery-market-set-to-grow-by-700-by-2017-142516205.html,

March 13, 2012.

[34] Electropaedia, “Battery Chargers and Charging Methods,” [Online]. Available:

http://www.mpoweruk.com/chargers.htm, October 6, 2012.

[35] Yuh-Shyan Hwang, Shu-Chen Wang, Fong-Cheng Yang and Jiann-Jong Chen,

“New Compact CMOS Li-ion Battery Charger Using Charge-Pump Technique

for Portable Applications,” IEEE Transactions on Circuits and Systems, Vol. 54,

no. 4, pp. 705-712, April 2007.

[36] Rhino-Charge Engineering, “Pulse Technology,” [Online]. Available:

http://www.tstonramp.com/~rhinocharge/pulsetech_faq.html, February 23, 2007.

[37] Scott Dearborn, “Charging Lithium-Ion Batteries: Not all Charging Systems Are

Created Equal,” [Online]. Available:

148
 http://www.microchip.com/stellent/groups/designcenter_sg/documents/market_c

ommunication/en028061.pdf, September 12, 2006.

[38] J. A. Paradiso and T. Starner, “Energy Scavenging for Mobile and Wireless

Electronics,” IEEE Pervasive Computing, Vol. 4, no. 1, pp. 18-27, January, 2005.

[39] S. Chalasani and J.M. Conrad, “A Survey of Energy Harvesting Sources for

Embedded Systems,” IEEE Southeastcon, pp. 442 – 447, April, 2008.

[40] Murugavel Raju, “Energy Harvesting,” [Online]. Available:

http://www.ti.com/corp/docs/landing/cc430/graphics/slyy018_20081031.pdf,

November 10, 2008.

[41] M.R. Mhetre, N.S. Nagdeo and H.K. Abhyankar, “Micro Energy Harvesting for

Biomedical Applications: A Review,” 3rd International Conference on

Electronics Computer Technology (ICECT), Vol. 3, pp. 1-5, April 2011.

[42] Vytautas Bielinskas, “Efficiency of Solar Energy Harvesting,” [Online]. Available:

http://construction21.eu/articles/h/efficiency-of-solar-energy-harvesting.html,

October 3, 2012.

[43] National Instruments, “Using Energy Harvesting Devices with National Instruments

Wireless Sensor Networks (WSN),” [Online]. Available:

http://www.ni.com/white-paper/12128/en#toc2, September 28, 2012.

[44] Pai H. Chou and Sehwan Kim, ‘Techniques for Maximizing Efficiency of Solar

Energy Harvesting Systems,” [Online]. Available:

http://www.ece.uci.edu/~chou/icmu10.pdf, February 17, 2010.

149
[45] Jeff Gruetter ,“Solar Energy Harvesting,” [Online]. Available:

http://cds.linear.com/docs/en/article/SolarEnergyHarvesting.pdf, October 19,

2010.

[46] Voltaic, “Amp Solar Charger,” [Online]. Available:

http://www.voltaicsystems.com/amp.shtml, October 11, 2012.

[47] Simone Dalola, Marco Ferrari, Vittorio Ferrari, Michele Guizzetti, Daniele Marioli

and Andrea Taroni, “Characterization of Thermoelectric Modules for Powering

Autonomous Sensors,” IEEE Transactions on Instrumentation and Measurement,

Vol. 58, no. 1, pp. 99-107, January 2009.

[48] Nicholas S. Hudak and Glenn G. Amatucci, “Small-scale Energy Harvesting

through Thermoelectric, Vibration, and Radio Frequency Power Conversion,”

Journal of Applied Physics, vol. 103, no. 10, February. 2008.

[49] M. Kishi, H. Nemoto, T. Hamao, M. Yamamoto, S. Sudou, M. Mandai, and S.

Yamamoto, “Micro-thermoelectric Modules and their Application to

Wristwatches as an Energy Source,” Proceedings 18th International Conference

on Thermoelectrics, pp. 301–307, August 1999.

[50] Gumballtech, “Energy Harvesting: Charge your Wearable Technology with Heat,

Sun, and Vibration,” [Online]. Available:

http://www.gumballtech.com/2012/02/21/energy-harvesting-charge-your-

wearable-technology-with-heat-sun-and-vibration/, February 21, 2012.

150
[51] T.A. Smith, J.P. Mars and G.A. Turner, “Using Supercapacitors to Improve Battery

Performance,” Power Electronics Specialists Conference, Vol. 1, pp. 124-128,

2002.

[52] A. Burke, “Ultracapacitors: Why, How, and Where is the Technology”, Journal of

Power Sources, Vol. 91, no. 1, pp. 37-50, November 2000,

[53] Marin S. Halper and James C. Ellenbogen, “Supercapacitors: A Brief Overview,”

[Online]. Available:

http://wwwsrv1.mitre.org/work/tech_papers/tech_papers_06/06_0667/06_0667.p

df, March 5, 2008.

[54] Illinois Capacitor, Inc., “Supercapacitors,” [Online]. Available:

http://www.illinoiscapacitor.com/pdf/papers/supercapacitors.pdf, January 13,

2012.

[55] Jun Li, Edward Murphy, Jack Winnick and Paul A. Kohl, “The Effects of Pulse

Charging on Cycling characteristics of Commercial Lithium-ion Batteries,”

Journal of Power Sources, Vol. 102, no. 1-2, pp. 302-309, December 2001.

[56] L. R. Chen, “ PLL-Based Battery Charge Circuit Topology,” IEEE Transactions on

Industrial Electronics, Vol. 51, no. 6, pp. 1344–1346, December 2004.

[57] L. R. Chen, “A Design of an Optimal Pulse Charge System by Frequency Varied

Technique,” IEEE Transactions on Industrial Electronics, Vol. 54, no. 1, pp.

398–405, February 2007.

151
[58] L. R. Chen, “Design of Duty-Varied Voltage Pulse Charger for Improving Li-ion

Battery Charging Response,” IEEE Transactions on Industrial Electronics, vol.

56, no. 2, pp. 480–487, February 2009.

[59] Linear Technology, “LTC4052-4.2-Lithium-ion Battery Pulse Charger with

Overcurrent Protection,” [Online]. Available:

http://www.linear.com/product/LTC4052-4.2, January 1, 2008.

[60] Linear Technology, “LTC1730- Lithium-ion Battery Pulse Chargers with

Overcurrent protection,” [Online]. Available:

http://www.linear.com/product/LTC1730, January 1, 2008.

[61] Maxim Integrated, “MAX1879-Simple, Efficient, 1-Cell Li+ Pulse Charger.”

[Online]. Available:

http://www.maximintegrated.com/datasheet/index.mvp/id/2544, May 30, 2001.

[62] International Warehouse, “A123-3.3VB Lithium Ion 2300mAh Single Cell

Battery,” [Online]. Available:

http://www.hobbyking.com/hobbyking/store/__16575__A123_3_3VB_Lithium_

Ion_2300mAh_Single_Cell_Battery.html, January 11, 2013

[63] Fairchild Semiconductor, “CMOS Schmitt Trigger-A Uniquely Versatile Design

Component,” Fairchild Semiconductor Application note, no. 140, June 1975.

[64] I. Chang, A.A. Abidi, and C.R. Viswanathan, “Flicker Noise in CMOS

Transistors from Subthreshold to Strong Inversion at Various Temperatures,”

IEEE Transactions on Electronic Devices, Vol. 41, no. 11, pp. 1965-1971,

November, 1994.

152
[65] Yuan Taur and Tak H. Ning, “Fundamentals of Modern VLSI Devices,” 1st. ed.,

Cambridge University Press, ch. 3, pp. 120-128, October1998.

[66] E.A. Vittoz, “The Design of High-Performance Analog Circuits on Digital CMOS

Chips,” IEEE Journal of Solid-State Circuits, Vol. 20, no. 3, pp. 657-665, June

1985.

[67] M. Filanovsky and H. Baltes “CMOS Schmitt Trigger design”, IEEE Transactions

on Circuits and Systems, Vol. 41, no. 1, pp. 46-49, January 1994.

[68] Munish Kumar, Parminder Kaur and Sheenu Thapar, “Design of CMOS Schmitt

Trigger,” International Journal of Engineering and Innovative Technology

(IJEIT), Vol. 2, no. 1, July 2012.

[69] F. B. Diniz, L. E. P. Borges and B. de. B. Neto, “A Comparative Study Of Pulsed

Current Formation for Positive Plates of Automotive Lead Acid Batteries,”

Journal of Power Sources, Vol. 109, no. 1, pp. 184–188, June 2002.

[70] Technical Document Center, “Recharging the Unchargeable,” [Online]. Available:

http://techdoc.kvindesland.no/radio/psu/20061103160627551.pdf, November 3,

2006.

[71] Marian Stofka, “Rectangular-waveform generator produces 25 and 75% duty

cycle.” EDN Design Ideas, no. 5, pp. 74-75, May 2010.

[72] Anico, “Charging at High and Low Temperatures,” [Online]. Available:

http://www.anico.hu/anico.php?o=tudod-

e/toltes%20kulonb%20homersekleteken_angol, February 26, 2013.

153
[73] Vadim V. Ivanov and Igor M. Filanovsky, “Operational Amplifier Speed and

Accuracy Improvement: Analog Circuit Design with Structural Methodology,”

1st ed., Springer, ch. 6, pp. 114–115, March, 2004.

[74] Lonne Mays, “Advanced ICs Facilitate Application-Specific Configuration of Li-

ion Battery Chargers,” Electronics world, Vol. 115, pp. 16-18, July 2009.

[75] Yoshiyasu Saito, Kiyonami Takano and Akira Negishi, “Thermal Behaviors of

Lithium-ion Cells during Overcharge,” Journal of Power sources, Vol. 97-98,

pp. 693-696, July 2001.

[76] Takahisa Ohsaki, Takashi Kishi, Takashi Kukobi, Norio Takami, Nao Shimura et

al, “ Overcharge Reaction of Lithium-ion Batteries”, Journal of Power Sources,

Vol. 146, no. 1-2, pp. 97-100, August 2005.

[77] I. Doms, P. Merken, R. Mertens, and C. Van Hoof, “Integrated Capacitive Power-

Management Circuit for Thermal Harvesters with Output Power 10 to1000μW,”

IEEE International Solid State Circuits Conference- Digest of Technical Papers,

pp. 300–301, February, 2009.

[78] Mengzhe Ma, “Design of High Efficiency Step-Down Switched Capacitor DC/DC

Converter,” M.Sc. thesis, Oregon State University, May, 2003.

[79] Linear Technology, “LTC3250 - High Efficiency, Low Noise, Inductorless Step-

Down DC/DC Converter,” Linear Technology [Online]. Available:

http://www.linear.com/product/LTC3250, September 19, 2009.

154
[80] Linear Technology, “LTC1503 - High Efficiency Inductorless Step-Down DC/DC

Converter,” Linear Technology [Online]. Available:

http://www.linear.com/product/LTC1503, January 30, 2010.

[81] Walt Kester, Brian Erisman and Gurjit Thandi, “Switched Capacitor Voltage

Converters,” [Online]. Available: http://www.analog.com/static/imported-

files/tutorials/ptmsect4.pdf, March 26, 2009.

[82] Ahmad Shar, “Design of a High-Speed CMOS Comparator,” M.Sc. thesis,

Linköping Institute of Technology, November, 2011.

[83] Shubhara Yewale and Radheshyam Gamad, “Design of Low Power and High Speed

CMOS Comparator for A/D Converter Application,” Wireless Engineering and

Technology, Vol. 3, pp. 90-95, April, 2012.

155