Bidirectional Asynchronous Generator and Battery Interface ...
Transcript of Bidirectional Asynchronous Generator and Battery Interface ...
Bidirectional Asynchronous Generator and
Battery Interface Circuit
by
Adél Coetzer
Thesis presented in partial fullment of the requirements forthe degree of Master of Engineering (Electrical) in the
Faculty of Engineering at Stellenbosch University
Supervisor: Prof. H.D.T. Mouton
Co-supervisor: Mnr. G.R. Turner
December 2016
Declaration
By submitting this thesis electronically, I declare that the entirety of the workcontained therein is my own, original work, that I am the sole author thereof(save to the extent explicitly otherwise stated), that reproduction and pub-lication thereof by Stellenbosch University will not infringe any third partyrights and that I have not previously in its entirety or in part submitted it forobtaining any qualication.
Date: . . . .December 2016
Copyright © 2016 Stellenbosch UniversityAll rights reserved.
i
Stellenbosch University https://scholar.sun.ac.za
Abstract
In some instances additional power is required to support an electrical load onboard an aircraft. A ram air turbine (RAT) generator can be used to supplythe necessary power in conjunction with a storage unit such as a battery orsupercapacitors. The RAT is driven by the airow surrounding the sub-sonicaircraft and acts as the main power supply. The storage unit is charged whenthe RAT is supplying excess power or when the load is disconnected.
This thesis presents the design and implementation of a prototype circuitthat demonstrates the interface between a generator and a battery of elec-trochemical cells to an electrical load. To simulate the airow through theturbine an asynchronous machine is employed to turn the shaft of an asyn-chronous generator. The prototype circuit is designed such that the battery isconnected to a dc bus through a bidirectional dc-dc converter and the asyn-chronous machine interfaces with the dc bus through a bidirectional inverter.The load is connected between the positive and negative dc bus rails.
The dc-dc converter circuit functions to regulate the dc bus voltage byutilising a double control loop strategy with an inner and outer control loop.The inverter functions as a synchronous rectier and the power delivered to thedc bus from the generator is controlled as a function of the converter's batteryvoltage. The simulated and measured test results of the complete system arecompared and presented.
ii
Stellenbosch University https://scholar.sun.ac.za
Opsomming
In sommige gevalle benodig vliegtuie addisionele krag om 'n elektriese las vankrag te voorsien. 'n Klein lugaangedrewe turbine kan gebruik word om dieaddisionele krag te voorsien in samewerking met 'n stooreenheid soos 'n batteryof superkapasitor. Die stooreenheid word herlaai wanneer die lugaangedreweturbine oortollige krag lewer aan die stelsel of wanneer die las ontkoppel word.
Hierdie tesis bespreek die ontwerp en implementering van 'n prototipestroombaan wat die werking tussen 'n generator en battery demonstreer. 'nAsinkroon motor word gebruik om die lugvloei deur die turbine te simuleer endie dryfas van die asinkroon generator aan te dryf. Die battery word gekoppelaan 'n gs bus deur middel van 'n bidireksionele gs-gs omsetter en die generatorword gekoppel aan die gs bus deur middel van 'n bidireksionele ws-gs omsetter.Die las word gekoppel tussen die positiewe en negatiewe gs bus lyne.
Die gs-gs omsetter funksioneer om die gs busspanning te reguleer deurmiddel van 'n dubbelle beheerlus strategie wat bestaan uit 'n binneste enbuitenste beheerlus. Die ws-gs omsetter funksioneer as 'n sinkroon gelykrigterwaar die krag wat aan die gs bus gelewer word, beheer word as 'n funskie vandie gs-gs omsetter se battery spanning. Die gesimuleerde en gemete resultatevir die volledige stelsel word vergelyk en voorgelê.
iii
Stellenbosch University https://scholar.sun.ac.za
Acknowledgements
I would like to express my sincere gratitude to the following people and organ-isations:
My supervisor, Prof H.D.T. Mouton, and my co-supervisor, GeoreyTurner, for their continuous support, advice and guidance throughoutthis project.
The CSIR for providing the required lab facilities and in particular Ge-orey Turner for his assistance in the lab.
Adrian Adams from the CSIR, for allocating the funding required forthis project.
Prof. M.J. Kamper for his guidance on the modelling of induction ma-chines.
Kushal Ramdas from the CSIR, for his assistance with Altium Designer.
Christo Swanepoel for his continuous encouragement, love and supportthroughout the duration of this project.
My family and friends for their love and support.
iv
Stellenbosch University https://scholar.sun.ac.za
Contents
Declaration i
Abstract ii
Opsomming iii
Acknowledgements iv
Contents v
List of Figures ix
List of Tables xiii
Nomenclature xiv
1 Introduction 1
1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Thesis Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . 21.3 Thesis Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2 Literature Review 5
2.1 Ram Air Turbines as Generators . . . . . . . . . . . . . . . . . . 52.1.1 Energy Management Strategies . . . . . . . . . . . . . . 52.1.2 Existing Power Generation Topologies . . . . . . . . . . 6
2.2 DC-DC Converters . . . . . . . . . . . . . . . . . . . . . . . . . 72.2.1 Pulse-Width Modulation . . . . . . . . . . . . . . . . . . 72.2.2 Bidirectional Converter Topology . . . . . . . . . . . . . 9
2.3 DC-AC Inverters . . . . . . . . . . . . . . . . . . . . . . . . . . 102.3.1 Three-Phase Inverters . . . . . . . . . . . . . . . . . . . 102.3.2 Inverter Control Schemes . . . . . . . . . . . . . . . . . . 12
2.4 Asynchronous Machines . . . . . . . . . . . . . . . . . . . . . . 122.4.1 Principle of Operation . . . . . . . . . . . . . . . . . . . 132.4.2 Equivalent Circuit . . . . . . . . . . . . . . . . . . . . . 162.4.3 Dynamic d-q Model . . . . . . . . . . . . . . . . . . . . . 18
v
Stellenbosch University https://scholar.sun.ac.za
CONTENTS vi
2.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3 DC-DC Converter Design and Implementation 21
3.1 Basic Circuit Operation . . . . . . . . . . . . . . . . . . . . . . 213.1.1 Inductor Design . . . . . . . . . . . . . . . . . . . . . . . 253.1.2 Bus Capacitor Design . . . . . . . . . . . . . . . . . . . . 273.1.3 MOSFET Power Loss and Heat Sink Design . . . . . . . 293.1.4 Soft-start Circuit . . . . . . . . . . . . . . . . . . . . . . 323.1.5 Isolated Gate-Drive Circuitry . . . . . . . . . . . . . . . 36
3.2 PWM Controller Design . . . . . . . . . . . . . . . . . . . . . . 383.2.1 Carrier Waveform Generation . . . . . . . . . . . . . . . 393.2.2 Controller Plant . . . . . . . . . . . . . . . . . . . . . . . 403.2.3 Current Control Loop . . . . . . . . . . . . . . . . . . . 413.2.4 Current and Voltage Limiting Circuit . . . . . . . . . . . 493.2.5 Voltage Control Loop . . . . . . . . . . . . . . . . . . . . 50
3.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
4 DC-DC Converter Simulation and Test Results 56
4.1 Test Set-Up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 564.2 Start-Up Response . . . . . . . . . . . . . . . . . . . . . . . . . 57
4.2.1 Gating Signal Implementation . . . . . . . . . . . . . . . 604.3 Step Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
4.3.1 Inductor Ripple Current . . . . . . . . . . . . . . . . . . 634.3.2 Gating Signals . . . . . . . . . . . . . . . . . . . . . . . . 65
4.4 Eciency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 664.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
5 Asynchronous Machine Modelling 68
5.1 Machine Parameter Estimation . . . . . . . . . . . . . . . . . . 685.1.1 Stator Resistance . . . . . . . . . . . . . . . . . . . . . . 695.1.2 Leakage Reactance . . . . . . . . . . . . . . . . . . . . . 695.1.3 Rotor Resistance . . . . . . . . . . . . . . . . . . . . . . 715.1.4 Magnetising Reactance . . . . . . . . . . . . . . . . . . . 715.1.5 Core Resistance . . . . . . . . . . . . . . . . . . . . . . . 72
5.2 Machine Characteristics . . . . . . . . . . . . . . . . . . . . . . 725.2.1 Torque-Speed Relationship . . . . . . . . . . . . . . . . . 725.2.2 Current-Speed Relationship . . . . . . . . . . . . . . . . 73
5.3 Dynamic Machine Model . . . . . . . . . . . . . . . . . . . . . . 755.3.1 Modelling Results . . . . . . . . . . . . . . . . . . . . . . 75
5.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
6 Inverter Design and Implementation 78
6.1 Basic Circuit Operation . . . . . . . . . . . . . . . . . . . . . . 786.1.1 Generator Voltage and Current . . . . . . . . . . . . . . 85
Stellenbosch University https://scholar.sun.ac.za
CONTENTS vii
6.1.2 IGBT Power Loss and Heat Sink Design . . . . . . . . . 876.1.3 Over-Temperature Protection . . . . . . . . . . . . . . . 916.1.4 Isolated Gate-Drive Circuitry . . . . . . . . . . . . . . . 936.1.5 Overvoltage Protection . . . . . . . . . . . . . . . . . . . 96
6.2 PWM Controller Design . . . . . . . . . . . . . . . . . . . . . . 976.2.1 Inverter Plant . . . . . . . . . . . . . . . . . . . . . . . . 986.2.2 Digital Signal Processor Interfacing . . . . . . . . . . . . 1016.2.3 Current Control Loop . . . . . . . . . . . . . . . . . . . 1066.2.4 Voltage Control Loop . . . . . . . . . . . . . . . . . . . . 112
6.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
7 Inverter Simulation and System Integration Test Results 118
7.1 System Integration . . . . . . . . . . . . . . . . . . . . . . . . . 1187.1.1 Test Set-Up . . . . . . . . . . . . . . . . . . . . . . . . . 1197.1.2 Start-Up Procedure . . . . . . . . . . . . . . . . . . . . . 119
7.2 PWM Generation and Gate Signals . . . . . . . . . . . . . . . . 1207.2.1 Reference Waveforms . . . . . . . . . . . . . . . . . . . . 1207.2.2 Triangular Carrier Waveform . . . . . . . . . . . . . . . 1237.2.3 Switching Signals . . . . . . . . . . . . . . . . . . . . . . 1237.2.4 IGBT Gate Signal . . . . . . . . . . . . . . . . . . . . . 124
7.3 Current Simulation and Test Results . . . . . . . . . . . . . . . 1257.3.1 Inverter Phase Currents . . . . . . . . . . . . . . . . . . 1257.3.2 Bus Current . . . . . . . . . . . . . . . . . . . . . . . . . 1307.3.3 Battery Current . . . . . . . . . . . . . . . . . . . . . . . 132
7.4 Bus Current and Battery Voltage Relationship . . . . . . . . . . 1337.4.1 Simulated Response . . . . . . . . . . . . . . . . . . . . . 1337.4.2 Measured Response . . . . . . . . . . . . . . . . . . . . . 134
7.5 Safety Features Testing . . . . . . . . . . . . . . . . . . . . . . . 1367.5.1 Over-Temperature Protection . . . . . . . . . . . . . . . 1367.5.2 Overcurrent Protection . . . . . . . . . . . . . . . . . . . 1367.5.3 Overvoltage Protection . . . . . . . . . . . . . . . . . . . 137
7.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
8 Conclusions 138
8.1 Thesis Conculsions . . . . . . . . . . . . . . . . . . . . . . . . . 1388.2 Recommendations and Future Work . . . . . . . . . . . . . . . . 139
List of References 141
Appendices 146
A Complete Circuit Schematics 147
A.1 DC-DC Converter Circuit Schematics . . . . . . . . . . . . . . . 147A.1.1 Power Stage Schematics . . . . . . . . . . . . . . . . . . 147
Stellenbosch University https://scholar.sun.ac.za
CONTENTS viii
A.1.2 Control Stage Schematics . . . . . . . . . . . . . . . . . 150A.2 Inverter Circuit Schematics . . . . . . . . . . . . . . . . . . . . 152
A.2.1 Power Stage Schematics . . . . . . . . . . . . . . . . . . 152A.2.2 Control Stage Schematics . . . . . . . . . . . . . . . . . 154
B Printed Circuit Boards 157
B.1 DC-DC Converter Printer Circuit Boards . . . . . . . . . . . . . 157B.1.1 Power Stage PCB . . . . . . . . . . . . . . . . . . . . . . 157B.1.2 Controller PCB . . . . . . . . . . . . . . . . . . . . . . . 160
B.2 Inverter Printer Circuit Boards . . . . . . . . . . . . . . . . . . 161B.2.1 Power Stage PCB . . . . . . . . . . . . . . . . . . . . . . 161B.2.2 Controller PCB . . . . . . . . . . . . . . . . . . . . . . . 162
Stellenbosch University https://scholar.sun.ac.za
List of Figures
1.1 Block digram of a bidirectional battery and machine interface circuit. 1
2.1 Half-bridge converter circuit. . . . . . . . . . . . . . . . . . . . . . . 82.2 PWM signal (green line) generated using a control signal (red line)
and sawtooth carrier waveform (blue line). . . . . . . . . . . . . . . 82.3 Bidirectional dc-dc converter circuit. . . . . . . . . . . . . . . . . . 92.4 Three-phase voltage source inverter connected to a three-phase load. 102.5 PWM voltages generated in three-phase inverter applications. . . . 112.6 Ideal three-phase, two pole induction machine. . . . . . . . . . . . . 132.7 Single-phase equivalent circuit of an induction motor. . . . . . . . . 172.8 Dynamic d-q model of a squirrel cage induction motor for the (a)
q-axis component and (b) d-axis component. . . . . . . . . . . . . . 19(a) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19(b) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.1 Bidirectional dc-dc converter circuit. . . . . . . . . . . . . . . . . . 223.2 Current and voltage waveforms describing the operation of the con-
verter when the batteries are being charged at maximum current. . 233.3 Current waveforms describing the operation of the converter when
the batteries are sourcing maximum current. . . . . . . . . . . . . . 253.4 Waveforms indicating the capacitor current and voltage when the
batteries are being charged at the maximum current. . . . . . . . . 283.5 Pre-charge soft-start circuit implementation. . . . . . . . . . . . . . 333.6 Dead-time implementation. . . . . . . . . . . . . . . . . . . . . . . 353.7 Post-charge soft-start circuit implementation. . . . . . . . . . . . . 363.8 Power section of the isolated gate-drive circuit. . . . . . . . . . . . 363.9 Block diagram of the MOSFET gate-drive circuit. . . . . . . . . . . 373.10 Block diagram of converter's control circuit. . . . . . . . . . . . . . 393.11 Circuit outline of the triangular carrier waveform generation. . . . . 403.12 Bidirectional dc-dc converter circuit used for controller design. . . . 413.13 Current control loop block diagram. . . . . . . . . . . . . . . . . . . 423.14 Inductor current sensing circuit. . . . . . . . . . . . . . . . . . . . . 433.15 Bode plot of current controller without compensation. . . . . . . . . 443.16 Pole-zero compensation amplier. . . . . . . . . . . . . . . . . . . . 45
ix
Stellenbosch University https://scholar.sun.ac.za
LIST OF FIGURES x
3.17 Current controller bode plot. . . . . . . . . . . . . . . . . . . . . . . 483.18 Current control loop with additional gain factor KB. . . . . . . . . 483.19 Bifurcation diagram used for stability analysis. . . . . . . . . . . . . 493.20 Postive and negative active clamping circuit. . . . . . . . . . . . . . 493.21 Voltage and current control loop block diagram. . . . . . . . . . . . 523.22 Simplied voltage and current control loop block diagram. . . . . . 523.23 Reduced voltage and current control loop block diagram. . . . . . . 523.24 Bus voltage feedback circuit. . . . . . . . . . . . . . . . . . . . . . . 533.25 Bode plot of the voltage controller without compensation. . . . . . 533.26 Bode plot of voltage controller under load. . . . . . . . . . . . . . . 55
4.1 Converter prototype circuit test set-up. . . . . . . . . . . . . . . . . 574.2 Simulated dc bus voltage and inductor current during start-up. . . 584.3 Measured dc bus voltage and inductor current during start-up. . . . 594.4 Measured inductor current during start-up. . . . . . . . . . . . . . . 604.5 Measured oscillator signal used in gate-drive circuitry. . . . . . . . . 604.6 Measured gating signal from the controller and MOSFET gate for
Q2 during start-up. . . . . . . . . . . . . . . . . . . . . . . . . . . . 614.7 Simulated step response of dc bus voltage and inductor current. . . 624.8 Measured step response of the dc bus voltage under ac coupling. . . 634.9 Simulated inductor current. . . . . . . . . . . . . . . . . . . . . . . 644.10 Measured inductor current without a load connected. . . . . . . . . 644.11 Measured inductor current with a 2.1 kW load connected to the dc
bus rails. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 654.12 Measured controller gating signals forQ1 andQ2 under steady-state
no-load conditions. . . . . . . . . . . . . . . . . . . . . . . . . . . . 654.13 Measured inductor current and ac coupled battery voltage under
load. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
5.1 Per-phase equivalent circuit of an induction machine. . . . . . . . . 685.2 Approximate per-phase equivalent circuit of an induction machine. 705.3 Reduced equivalent circuit of an induction machine under no-load. . 715.4 Torque vs rated speed characteristic curve. . . . . . . . . . . . . . . 735.5 Stator current vs rated speed characteristic curve. . . . . . . . . . . 745.6 Three-phase stator voltages and currents at zero slip. . . . . . . . . 765.7 Stator and rotor currents at -5 % slip. . . . . . . . . . . . . . . . . 76
6.1 Bidirectional converter and inverter interface circuit. . . . . . . . . 796.2 PWM signal generation block diagram. . . . . . . . . . . . . . . . . 806.3 Inverter phase voltage and current for one half-bridge circuit. . . . . 816.4 Switch and diode currents for S1 and S2 over a single angular rotation. 836.5 Instantaneous and average dc bus current. . . . . . . . . . . . . . . 846.6 Instantaneous and fundamental line and phase voltages. . . . . . . 856.7 Instantaneous line current. . . . . . . . . . . . . . . . . . . . . . . . 86
Stellenbosch University https://scholar.sun.ac.za
LIST OF FIGURES xi
6.8 Average switching current approximation. . . . . . . . . . . . . . . 896.9 Temperature sensing circuit for the IGBT module. . . . . . . . . . . 916.10 PWM signal and error signal integration. . . . . . . . . . . . . . . . 926.11 Block diagram of the IGBT gate-drive circuit. . . . . . . . . . . . . 946.12 Block diagram of the inverter control circuit. . . . . . . . . . . . . . 976.13 Bidirectional inverter circuit. . . . . . . . . . . . . . . . . . . . . . . 996.14 Phase angle over generator operating frequency range. . . . . . . . 1006.15 Shaft coupling with magnetic pick-up for speed feedback. . . . . . . 1026.16 Magnetic pick-up interfacing with the dsPIC. . . . . . . . . . . . . 1036.17 Three-phase reference waveform generation with the dsPIC. . . . . 1056.18 DC bus current control loop block diagram. . . . . . . . . . . . . . 1066.19 DC bus current sensor feedback circuit. . . . . . . . . . . . . . . . . 1086.20 Bode plot of the open-loop response of the current control loop
without the compensator present. . . . . . . . . . . . . . . . . . . . 1106.21 Active integrator compensation circuit. . . . . . . . . . . . . . . . . 1106.22 Bode plot of the open-loop response of the complete current control
loop. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1116.23 Tracking back anti-windup scheme with an integral controller. . . . 1126.24 Voltage and current control loop block diagram. . . . . . . . . . . . 1136.25 Battery voltage feedback circuit. . . . . . . . . . . . . . . . . . . . 1146.26 Reduced voltage and current control loop block diagram. . . . . . . 1156.27 Bode plot of the open-loop frequency response of the voltage control
loop without compensation. . . . . . . . . . . . . . . . . . . . . . . 1156.28 Bode plot of the open- and closed-loop frequency response of the
voltage control loop. . . . . . . . . . . . . . . . . . . . . . . . . . . 116
7.1 Test set-up for generator and battery interface circuit. . . . . . . . 1197.2 Measured modulator reference waveforms at an increasing genera-
tor shaft speed. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1217.3 Measured modulator reference waveforms at a shaft frequency of
25 Hz. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1217.4 Measured modulator reference waveforms at a shaft frequency of
50 Hz. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1227.5 Measured modulator reference waveforms after the operating shaft
speed range was exceeded. . . . . . . . . . . . . . . . . . . . . . . . 1227.6 Measured triangular carrier waveform. . . . . . . . . . . . . . . . . 1237.7 Measured PWM switching signals for one inverter half-bridge. . . . 1247.8 Measured controller and gate PWM switching signals. . . . . . . . . 1247.9 Simulated inverter phase currents at a 3.3 A dc bus current set-point.1267.10 Measured inverter phase currents at a measured dc bus current of
3.3 A. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1267.11 Simulated inverter phase current ripple. . . . . . . . . . . . . . . . 1277.12 Measured inverter phase current ripple. . . . . . . . . . . . . . . . . 128
Stellenbosch University https://scholar.sun.ac.za
LIST OF FIGURES xii
7.13 Simulated modulator reference voltage and phase current for oneinverter half-bridge. . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
7.14 Measured modulator reference voltage and phase current for oneinverter half-bridge. . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
7.15 Measured modulator reference voltage and phase current at a shaftfrequency of 34 Hz. . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
7.16 Simulated dc bus current during generator start-up. . . . . . . . . . 1307.17 Measured dc bus current at the maximum set-point current. . . . . 1317.18 Measured dc bus current whilst charging the batteries. . . . . . . . 1327.19 Measured battery current whilst charging the batteries. . . . . . . . 1327.20 Simulated mean dc bus current and battery voltage. . . . . . . . . . 1347.21 Measured mean dc bus current vs battery voltage. . . . . . . . . . . 135
A.1 Schematic for the top-half of the dc-dc converter circuit. . . . . . . 148A.2 Schematic for the bottom-half of the dc-dc converter circuit. . . . . 149A.3 Control schematic for the dc-dc converter circuit. . . . . . . . . . . 151A.4 Schematic for the power stage of the inverter circuit. . . . . . . . . 153A.5 Schematic for the inverter digital control circuit. . . . . . . . . . . . 155A.6 Schematic for the inverter analog control circuit. . . . . . . . . . . . 156
B.1 PCB for the top-half of the dc-dc converter circuit producing +175 V.158B.2 PCB for the bottom-half of the dc-dc converter circuit producing
-175 V. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159B.3 Control PCB of the dc-dc converter. . . . . . . . . . . . . . . . . . 160B.4 PCB for the power stage of the inverter circuit. . . . . . . . . . . . 161B.5 PCB for the digital inverter control circuit. . . . . . . . . . . . . . . 162B.6 PCB for the analog inverter control circuit. . . . . . . . . . . . . . . 163
Stellenbosch University https://scholar.sun.ac.za
List of Tables
1.1 Bidirectional interface system parameters . . . . . . . . . . . . . . . 3
3.1 BSB DB12-40 battery specications. . . . . . . . . . . . . . . . . . 223.2 Converter circuit parameters. . . . . . . . . . . . . . . . . . . . . . 293.3 IRFP264 MOSFET parameter values. . . . . . . . . . . . . . . . . . 293.4 Pre-charge soft-start circuit component values. . . . . . . . . . . . . 333.5 Post-charge soft-start circuit component values. . . . . . . . . . . . 353.6 Component values used in oscillator gate-drive circuit. . . . . . . . 373.7 TLP250 characteristic values. . . . . . . . . . . . . . . . . . . . . . 373.8 Triangular waveform generator component values. . . . . . . . . . . 403.9 Component values for the current sense circuit. . . . . . . . . . . . 433.10 Component values for the current loop compensator. . . . . . . . . 473.11 Component values for the voltage loop compensator. . . . . . . . . 543.12 Stability analysis of the voltage control loop at various mean in-
ductor current operating points. . . . . . . . . . . . . . . . . . . . . 55
5.1 Three phase induction machine characteristic values. . . . . . . . . 695.2 Calculated machine parameters. . . . . . . . . . . . . . . . . . . . . 725.3 Machine parameters used for modelling. . . . . . . . . . . . . . . . 75
6.1 FS30R06W1E3 IGBT module parameter values. . . . . . . . . . . . 876.2 ACPL-333J gate driver characteristic values. . . . . . . . . . . . . . 946.3 Parameter values of the dsPIC30F4013 DSP. . . . . . . . . . . . . . 1016.4 Component values for the magnetic pick-up and dsPIC interfacing
circuit. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1036.5 Allegro ACS756 Hall eect sensor characteristics. . . . . . . . . . . 1076.6 Component values for the current feedback circuit. . . . . . . . . . 108
xiii
Stellenbosch University https://scholar.sun.ac.za
Nomenclature
Variables
fs switching frequency . . . . . . . . . . . . . . . . . . . . . . [ kHz ]
idc instantaneous dc bus current . . . . . . . . . . . . . . . . . [A ]
iL instantaneous inductor current . . . . . . . . . . . . . . . . [A ]
ma modulation index . . . . . . . . . . . . . . . . . . . . . . . [ ]
Abbreviations
ac alternating current
ADC analog-to-digital converter
DAC digital-to-analog converter
dc direct current
DSP digital signal processor
emf electromotive force
IC integrated circuit
IGBT insulated-gate bipolar transistor
LUT lookup table
mmf magnetomotive force
MOSFET metal-oxide-semiconductor eld-eect transistor
MPPT maximum power point tracking
NTC negative temperature coecient
PCB printed circuit board
PI proportional-integral
PWM pulse-width modulation
RAT ram air turbine
rms root mean square
VSD variable-speed drive
VSI voltage source inverter
xiv
Stellenbosch University https://scholar.sun.ac.za
Chapter 1
Introduction
1.1 Background
In order to support an electrical load on board an aircraft, additional power isrequired. For this application a ram air turbine (RAT) generator can be usedto supply the necessary power in conjunction with a storage unit such as abattery or supercapacitors. The RAT is driven by the airow surrounding thesub-sonic aircraft and acts as the main power supply providing average powerto a load. Generally the storage unit is only used to supply power duringtransient conditions [1]. If however the RAT is unable to supply the requiredpower due to a lack of airow, the storage unit supplies the necessary poweruntil it is depleted. The storage unit is charged when the RAT is supplyingexcess power.
A prototype circuit that demonstrates the interface between an air-poweredturbine generator and a battery of electrochemical cells to an electrical loadon board an aircraft is proposed. To simulate the airow through the turbinean asynchronous machine is employed to turn the shaft of an asynchronousgenerator. A speed-varying inverter will be used to control the voltage andfrequency applied to the stator windings of the driving machine, thus simulat-ing varying airow conditions. A deep-cycle lead-acid battery will be used asthe storage unit. The layout of the proposed circuit is shown in Fig. 1.1.
BidirectionalSwitched-Mode
dc-dcConverter
L
O
A
D
Bidirectional3-Phase
Switched-ModeInverter
M
Figure 1.1: Block digram of a bidirectional battery and machine interfacecircuit.
1
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 1. INTRODUCTION 2
The battery is connected to a dc bus through a non-isolated bidirectional dc-dcconverter and the asynchronous machine interfaces with the dc bus through anon-isolated bidirectional inverter. The interface is required to transfer up to2 kW of electrical power between the battery and asynchronous machine. Theload is connected between the positive and negative dc bus rails. The proposednominal dc bus voltage is 350 V rail-to-rail. A reasonably high bus voltage isused to reduce the current within the dc bus, thus reducing self-heating whileimproving the eciency of the power distributed to the load.
The dc-dc converter is bidirectional to facilitate extended periods of re-generative braking (during which energy is returned to the battery). Theconverter is required to be current-controlled and voltage-regulated. As suchthe direction of current to or from the batteries will be regulated as a functionof the dc bus voltage.
The inverter circuit will function as a synchronous rectier to facilitate theow of current from the asynchronous machine to the dc bus whilst airborne.
1.2 Thesis Objectives
The thesis objectives can be divided into three parts. The rst objective is todesign, manufacture and test the bidirectional non-isolated current-controlledvoltage-regulated switched-mode dc-dc converter that is required to interfacea bank of batteries to a dc bus. The bidirectional transfer of up to 2 kWelectrical power between a nominally 350 V dc bus and a bank of batteries isto be demonstrated. The eciency of the converter is to be determined. Aclosed-loop control circuit is required for the design of the converter and willbe analysed before manufacturing the prototype circuit to ensure the stabilityof the design.
Secondly, the design, manufacture and testing of a bidirectional non-isolatedcurrent-regulated switched-mode inverter is required to interface a commer-cially available 3-phase asynchronous machine to a dc bus. A bidirectionaltransfer of up to 2 kW electrical power between a 3-phase asynchronous ma-chine and a nominally 350 V dc bus is to be demonstrated. The shaft ofthe inverter-connected asynchronous machine will be coupled to an identical3-phase asynchronous machine controlled by an o-the-shelf variable-speeddrive (VSD) to function as the air-driven turbine. Low-speed commerciallyavailable 3-phase asynchronous machines (400 V AC, 3 kW, 3000 rpm) willbe employed. A closed-loop control circuit will be designed and analysed toensure the closed-loop stability of the design. Table 1.1 shows a summary ofthe system parameters.
The last objective is to integrate and test the inverter, dc bus and dc-dcconverter interface. A bidirectional transfer of up to 2 kW of electrical powerbetween the inverter, converter and the load is to be demonstrated. Thebattery current due to the 2 kW load at the output of the dc-dc converter
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 1. INTRODUCTION 3
Table 1.1: Bidirectional interface system parameters
Parameter ValueSystem power rating (bidirectional) 2 kWDC bus voltage operating range 340 V - 360 VNominal dc bus voltage 350 VMachine speed rating 3000 rpmMachine power rating 3 kWMachine voltage rating 400 V AC
should be shown to progressively decrease to zero as the mechanical powerapplied to the shaft of the inverter-connected machine is progressively raised,until such point as the power delivered by the inverter (less losses) equalsthat consumed by the load. Similarly, the battery current should be shown tofurther decrease (below zero) as the mechanical power applied to the shaft ofthe inverter-connected machine is progressively raised beyond that consumedby the load. Thus, the batteries should discharge while the batteries sourcepower to all or a fraction of the load's requirement, and should charge whilethe asynchronous machine sources power in excess of that consumed by theload.
1.3 Thesis Outline
The structure of the thesis is briey described below in terms of the contentprovided in each of the chapters.
Chapter 2 provides background information on the relevant literatureused in this thesis. Previous work regarding the use of ram air turbines asgenerators is discussed. Bidirectional converter and inverter topologies areinvestigated along with pulse-width modulation schemes that can be used tocontrol these circuits. The operation and development of an equivalent circuitfor an asynchronous machine are also presented.
Chapter 3 describes the basic circuit operation of the bidirectional dc-dcconverter circuit. A detailed design of the control circuit used for the converteris presented along with the frequency domain closed-loop stability analysis ofthe system.
Chapter 4 presents the time domain analysis of the dc-dc converter. Thesimulation results are compared to the measured results of the prototype cir-cuit, which was manufactured from the specications given in Chapter 3.
Chapter 5 derives the asynchronous machine parameters from the equiv-alent machine model and the datasheet information. The machine model issimulated and the results are compared to the given datasheet information inorder to conrm the credibility of the model. The generator operation of the
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 1. INTRODUCTION 4
machine is also investigated.Chapter 6 describes the basic circuit operation of the inverter functioning
as a synchronous rectier. A detailed design of the control loops used tocontrol the inverter switching action is presented. Frequency domain analysisis used to conrm the closed-loop stability of the system. The implementationof additional safety features is discussed.
Chapter 7 details the integration of the inverter and dc-dc converter cir-cuit. The time domain simulation of the inverter circuit is compared with themeasured results from the prototype circuit. The dc-dc converter's bidirec-tional capability is investigated. The test results used to conrm the successfulintegration between the converter and inverter circuits are presented. Lastly,the safety features' testing procedures and results are given.
Chapter 8 concludes the thesis by presenting the overall system simulationand test results. Future work and recommendations are also considered.
Stellenbosch University https://scholar.sun.ac.za
Chapter 2
Literature Review
This literature review gives background information on some of the key ele-ments of the proposed project. The focus is on the use of ram air turbines asgenerators, the development of converter and inverter circuits and the mod-elling and operation of three-phase asynchronous machines.
2.1 Ram Air Turbines as Generators
Ram air turbine systems are most commonly used to supply emergency poweron-board aircraft when the main power supply fails but can also be used tosupply continuous power whilst airborne. The turbine usually drives an electricgenerator connected to a dc bus and is supported by a battery or other storagedevice to form a hybrid system [1]. Some of the key requirements of sucha system include high reliability, adequate load regulation and high overallsystem eciency.
2.1.1 Energy Management Strategies
Since the RAT speed will vary during ight, controlled converters are gener-ally required. These converters operate either to form a controlled interfacebetween the machine and the supply bus, or to regulate the machine excitation,or both [2]. In [1] and [3] two energy management strategies are investigatedto interface a RAT and a storage device: the conventional strategy and a dualstrategy.
The conventional strategy makes use of a current-controlled dc-dc converterto interface the dc bus and the storage unit. The RAT is used as a voltagesource and the excitation current of the generator is controlled to maintain aconstant dc bus voltage at the output of the rectier. The RAT is thus onlyused to supply the required average power to the load while the storage unitsupplies peak and transient power.
5
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 2. LITERATURE REVIEW 6
In the dual-control strategy, the RAT's generator excitation voltage is con-trolled in order to regulate the current supplied to the dc bus, while the storagedevice with its dc-dc converter is used to regulate the dc bus voltage. This al-lows the use of maximum power point tracking (MPPT) but only if the powerdemanded by the load and storage device exceeds the maximum power that canbe supplied by the RAT. MPPT techniques are explained in detail in [4, 5, 6].The overall objective is to provide the optimal rotor speed with respect to theratio between the blade-tip speed and wind speed [7]. This will produce themaximum amount of machine torque and power. The dual strategy optimisesenergy transfer and is therefore the preferred control strategy.
MPPT is very popular in wind-power applications where the power deliv-ered by wind-driven generators are fed back into a grid or other storage unitwith a large storage capacity. In the case of a ram air turbine, the RAT is onlyused to supply the power required by the load and to charge the storage deviceto its recommended capacity. Thereafter, no more power must be deliveredto the system since there is no place to dump the excess power. If MPPT isused in RAT applications, the system must be designed such that it will exitMPPT mode if the load power requirement is less than the RAT's maximumpower.
2.1.2 Existing Power Generation Topologies
Various approaches have been investigated to generate power from a ram airturbine and is summarised in this subsection. In 2005 P.H. Mellor publisheda paper [8] on generating power from a RAT using a brushless permanent-magnet (PM) generator and an integrated xed-ratio gearbox. The machinewas controlled to produce optimal torque at low speeds, while at high speedsthe peak line voltage was limited to the desired dc bus voltage. A resolver tomeasure the degrees of shaft rotation was used for feedback along with machinecurrent measurements. An insulated-gate bipolar transistor (IGBT) moduleand a digital signal processor (DSP) based controller was used to control thegenerator. This design showed an overall system eciency of 84 %.
A dierent approach was used by P. Bolognesi in [2] where the generatorfor the RAT was a hybrid-excited dc machine consisting of both permanentmagnets and eld coils. The armature of the dc machine is directly connectedto the dc bus, while the excitations of the eld coils are controlled by means ofa chopper circuit to keep the dc bus voltage within the desired voltage range.Extensive modelling showed the validity of such a system.
In [1] the use of a hybrid system comprising of supercapacitors and a syn-chronous generator are proposed to generate power on board an aircraft. Abidirectional current-controlled dc-dc converter is used to interface with thedc bus and supercapacitors. To reduce current ripple and increase the ap-parent switching frequency, an interleaved converter was used consisting oftwo IGBT modules with triple buck-boost branches where the PWM signals
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 2. LITERATURE REVIEW 7
for each branch are 120 degrees phase shifted. Proportional-integral controllersare used to regulate the dc bus voltage. The stator currents of the synchronousgenerator are rectied through a three-phase diode-bridge rectier to producepower to the load. Both the conventional and dual energy-management strate-gies were tested and shown to have nearly equivalent dynamic behaviour.
Similar work was done by X. Roboam in [3] with a bigger focus on thedierence between the two energy management strategies. The only dierencewas shown to be that the dual strategy allows faster dc bus regulation andmaintains better system stability during transient operation.
A lithium-ion battery and RAT interface is presented in [9]. The turbineis connected to a synchronous generator through a speed multiplier. A dioderectier, connected to the dc bus, is used to rectify the three-phase statorvoltages. A bidirectional dc-dc converter is used for interfacing the dc bus andLi-Ion battery. Again, both the conventional and dual strategies were tested.The dc bus voltage ripple on the conventional strategy was shown to be lessthan for the dual strategy, while the dual strategy showed slightly better dcbus regulation.
2.2 DC-DC Converters
The proposed circuit for this thesis requires a switched-mode bidirectionaldc-dc converter to interface with the dc bus and battery storage device. Theintention is to control the converter current in order to regulate the dc busvoltage. In switched-mode converters this is achieved by controlling the onand o time of the switches. Dierent switching schemes exist but the mostwidely used scheme is pulse-width modulation [10] which employs switchingat a constant frequency.
2.2.1 Pulse-Width Modulation
Pulse-width modulation (PWM) varies the duty cycle of converter's switchesat a high frequency to obtain the desired average output voltage or current[10]. An example of a circuit that uses PWM to control the switches is shownin Fig. 2.1, where insulated-gate bipolar transistors (IGBTs) are used as theswitches. The signal to determine the state of a switch (on or o) is obtainedby comparing the control signal to a repetitive waveform with a constant fre-quency, referred to as the carrier waveform as shown in Fig. 2.2. The carrierwaveform can either be a sawtooth or triangular waveform [10]. The dier-ence between the desired and actual output is passed through a compensationamplier to generate the required control signal [11].
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 2. LITERATURE REVIEW 8
−
Vd
2
+
−
Vd
2
+
S1
+
-
voS2
Figure 2.1: Half-bridge converter circuit.
t
1
−1
t
Vd
2
−Vd
2
Figure 2.2: PWM signal (green line) generated using a control signal (red line)and sawtooth carrier waveform (blue line).
The duty cycle of a switch with constant control signal vcontrol is expressed by
D =tonTs
=vcontrol
Vc
, (2.1)
where Ts is the switching period which corresponds to the carrier waveformfrequency and Vc is the peak value of said carrier waveform.
PWM strategies can be divided between naturally sampled PWM andregular sampled PWM. Naturally sampled PWM is the earliest and moststraightforward strategy which compares a low-frequency control signal to ahigh-frequency carrier waveform and is used in analog controllers, as shown inFig. 2.2. Regular sampled PWM is used in digital control systems where thecontrol signal is sampled at a regular interval and the duty cycle is adjustedin proportion to the value of the sample.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 2. LITERATURE REVIEW 9
2.2.2 Bidirectional Converter Topology
An in-depth study on the dierent switched-mode dc-dc converter topologiesare given in [12]. For a bidirectional converter the most commonly used topol-ogy consists of a half-bridge circuit that can either sink or source current fromor to the dc bus [13], [14]. A simple bidirectional half-bridge circuit is shownin Fig. 2.3 with metal-oxide-semiconductor eld-eect transistors (MOSFETs)as switches. Alternatively, IGBTs can be used as switches and should includefreewheeling diodes across them.
Vbat
LiL vm
Q2
Q1
C
+
−
Vdc
Figure 2.3: Bidirectional dc-dc converter circuit.
From rst principles it is known that the average voltage across an induc-tor under steady state conditions must equal zero [11]. Thus, the followingstatement with respect to Fig. 2.3 can be made:
Vbat = vm, (2.2)
where vm is the average mid-point voltage which can be calculated as
vm =
(t1(on)Ts
Vdc +t2(on)Ts
0
), (2.3)
where t1(on) and t2(on) is the on-time of switches Q1 and Q2, respectively. From(2.2) and (2.3) the relationship between the dc bus voltage and the batteryvoltage is thus
Vdc =Vbatt1(on)
Ts
=Vbat
D1
, (2.4)
where D1 is the duty cycle of Q1. The duty cycle of Q2 is
D2 =t2onTs
= 1−D1 = 1− Vbat
Vdc
, (2.5)
such that Q1 and Q2 is never on simultaneously. For ideal switches the aboveequation would suce to prevent short-circuit conditions. However, in real
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 2. LITERATURE REVIEW 10
life switches require a nite amount of time to change state. By introducingdead-time, cross conduction can be avoided. The dead-time inserts a smalltime delay before a switch is turned on in order to allow the other switch tocompletely switch o rst [15].
2.3 DC-AC Inverters
For the proposed project a bidirectional three-phase switched-mode ac-dc in-verter circuit is required to interface an asynchronous generator and dc bus.Two topologies exist for inverter circuits, namely current source inverters(CSIs) and voltage source inverters (VSIs). CSIs are only used in very highpower applications. For this literature review only VSIs will be consideredsince they are by far the most commonly used inverter topology [11].
2.3.1 Three-Phase Inverters
The basic topological structure of an inverter circuit is a half-bridge. A three-phase inverter requires three half-bridges where each half-bridge pole-pointconnects to one of the output load terminals as shown in Fig. 2.4. The switchescan either be IGBTs with freewheeling diodes or MOSFETs with built-in diodesto support bidirectional power ow [15].
idc
C
+Vdc
−Vdc
S1
S2
vAo
S3
S4
vBoiBo
S5
S6
vCoiCo
iAo
3-PhaseLoad
Figure 2.4: Three-phase voltage source inverter connected to a three-phaseload.
The principle of operation is similar to that of the circuit described in Section2.2.2 where the top and bottom switch of each phase leg are switched com-
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 2. LITERATURE REVIEW 11
plementary to each other. Considering the half-bridge circuit with switchesS1 and S2, when S1 is switched on the pole voltage vAo gets pulled high to+Vdc and when S2 is on vAo gets pulled down to −Vdc. By controlling theon-time of the switches the fundamental voltage across the load terminals canbe controlled. A PWM technique is thus generally also used in VSI circuits tocontrol the switches [16].
t
t
t
t
fA(t) fB(t) fC(t)
vAo
vBo
vABM
√3
M
Figure 2.5: PWM voltages generated in three-phase inverter applications.
Three-phase PWM is described in Figure 2.5. It shows three control sinusoidsfA(t), fB(t) and fC(t) indicated in red, green and black respectively, where eachwaveform is controlling one half-bridge. The three waveforms are phase shiftedby 120 degrees relative to each other to generate the required fundamentalthree-phase voltage waveforms at the load terminals. A triangular waveformis used as the carrier waveform and is indicated in blue. The two PWMvoltage waveforms vAo and vBo represent the two pole voltages from Fig. 2.4,respectively. If the load is connected in delta conguration, the voltage appliedacross one of the three load windings will equal the dierence between the twoinverter pole voltages vAo and vBo. The resulting PWM waveform is shown asvAB in Fig 2.5. If vAB is put through a low pass lter the resulting waveformis a sinusoid with a amplitude
√3 times greater than that of the fundamental
pole voltage waveforms indicated in Fig 2.5.The required dead-time to prevent a short-circuit between the dc bus rails
varies with the power rating of the inverter circuit. Typically the requireddead-time to ensure safe operating conditions is well below 1 µs [15].
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 2. LITERATURE REVIEW 12
2.3.2 Inverter Control Schemes
VSIs can be classied as either current- or voltage-controlled. A CCVSI hasthe advantage of inherent overcurrent protection.
Three main control schemes exist for inverters that drive ac machinesnamely scalar control, vector control and direct torque control [17]. Scalarcontrol techniques have been widely used in industry and are easy to imple-ment [16]. Volt/Hertz control is a scalar control technique where the ratiobetween the stator voltage and frequency is kept constant to ensure a con-stant air gap ux close to rated ux [18]. Volt/Hertz control is thus used involtage-controlled voltage-source inverters (VCVSIs). Scalar control gives in-ferior performance compared to vector control schemes in terms of their speedresponse. However, if the speed of the machine is not required to changerapidly, scalar control provides more than adequate performance [18].
Vector control allows an ac machine to be controlled as if it were a sepa-rately excited dc motor by transforming its phase currents into a synchronouslyrotating reference frame [19]. The two resulting currents are dc currents, onerepresenting the torque producing component of the stator current while theother corresponds to the ux producing component of the stator current. Bycontrolling these two currents the machine ux and torque can thus be con-trolled. Vector control, also known as eld-oriented control, provides excellentperformance but with the drawback of high complexity. Vector control schemesrequire a powerful microcontroller or DSP [16]. With vector control the innercontrol loop is always current controlled allowing it to function as a CCVSI.Dierent forms of eld-oriented control (FOC) exist such as indirect FOC [20]and direct FOC [21], [22].
Direct torque control (DTC) schemes control both the torque and statorux of the machine whereby the required ux and torque is compared to therespective estimated values to generate two error voltages. The error voltagesare constrained to lie between an upper and lower limit respectively, thusproviding hysteresis-band control [23]. A detailed description of DTC is givenin [17]. Direct torque control schemes have shown similar performance to thatof vector control schemes [16]. The combination of DTC and vector controlschemes also exist and is presented in [24].
2.4 Asynchronous Machines
This project will make use of two asynchronous machines, one to functionas a generator, the other to drive the shaft of the generator. The operationof asynchronous machines is investigated in order to model and develop thecontrol circuit for the inverter controlling the generator.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 2. LITERATURE REVIEW 13
2.4.1 Principle of Operation
Asynchronous machines, also known as induction machines, are the most com-monly used type of ac machines in industry. The machine consists of a statorand rotor winding that are magnetically coupled to transfer energy betweenmechanical and electrical systems [19]. The rotor of an induction machine canbe classied as either a wound rotor or a squirrel-cage rotor. Wound rotor in-duction machines are fairly uncommon, thus for this study only squirrel-cagerotor induction machines are investigated. A squirrel-cage rotor is constructedsuch that the windings on the rotor are conducting bars short-circuited ateach end by conducting rings. Both the stator and rotor cores are made fromlaminated ferromagnetic material.
The fundamental principle of induction machines is that of a rotatingsinusoidally-distributed magnetic eld within the air gap of the machine. Thisis achieved by applying balanced sinusoidal three-phase voltages to the statorwindings and in return a synchronously rotating magnetic eld is produced.A summary of the analytical derivation [16] follows below.
−a
a
b
−b−c
c
Axis ofphase a
Axis ofphase c
Axis ofphase b
θe
Figure 2.6: Ideal three-phase, two pole induction machine.
Figure 2.6 shows an ideal three-phase, two-pole induction machine where thephase windings for the stator are indicated by concentric circular coils. Themachine can either be connected in delta or wye conguration. The three-phasewindings are 120 degrees displaced from each other (sinusoidally distributed)and the air gap in the machine is considered uniform.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 2. LITERATURE REVIEW 14
When balanced sinusoidal three-phase voltages are applied to the windings theinstantaneous currents are
ia = Imaxcos(ωet), (2.6)
ib = Imaxcos(ωet− 120°), (2.7)
ic = Imaxcos(ωet+ 120°), (2.8)
where Imax is the maximum stator winding current, ωet is the frequency of theapplied excitation voltage and t denotes time. Each of the phase windings willindependently induce a sinusoidally distributed magnetomotive force (mmf)wave since
F = Ni, (2.9)
where F is mmf and N is the number of turns in a phase winding. Theresulting mmf expressions are given by
Fa = Niacos(θ), (2.10)
Fb = Nibcos(θ − 120°), (2.11)
Fc = Niccos(θ + 120°), (2.12)
where θ is the spatial angle. The total mmf at angle θ is given as
F(θ) = Fa(θ) + Fb(θ) + Fc(θ)
= Niacos(θ) +Nibcos(θ − 120°) +Niccos(θ + 120°).(2.13)
Substituting (2.6) through (2.8) in (2.13) results in
F(θ, t) = NImax [cos(ωet)cos(θ) + cos(ωet− 120°)cos(θ − 120°)
+ cos(ωet+ 120°)cos(θ + 120°)] . (2.14)
After simplifying (2.14) the expression becomes
F(θ, t) =3
2NImaxcos(ωet− θ). (2.15)
The air gap mmf wave from (2.15) is shown to be sinusoidally distributed,rotating at synchronous speed ωs where
ωs =2
polesωe. (2.16)
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 2. LITERATURE REVIEW 15
For a two-pole machine the angular frequency of the applied electrical excita-tion ωe is equal to the synchronous speed of the air gap mmf ωs. In revolutionsper minute the synchronous speed can be expressed by
ns =
(120
poles
)fe, (2.17)
since
fe =ωe
2π, (2.18)
where fe is the applied electrical frequency. The time-varying mmf wave willproduce an electric eld in accordance with Faraday's law:∮
c
E · ds = − d
dt
∫S
B · da. (2.19)
Equation (2.19) states that the induced electromotive force (emf) around aclosed contour is equal to the negative rate of change of the magnetic ux φpassing through an enclosed area [25] since
φ =
∫S
B · da. (2.20)
The rotor winding forms the closed contour and links the core ux N times,thus (2.19) reduces to
e = −Ndφ
dt= −dλ
dt, (2.21)
where λ is the ux linkage of the winding. The emf e will cause an inducedcurrent to ow in the rotor winding. When the current-carrying winding isexposed to a magnetic eld, torque is produced to rotate the rotor winding.This is referred to as the starting torque. If the torque on the load attachedto the shaft of the motor is less than the starting torque, the rotor will startto turn [16]. The rotor will turn in the same direction as that of the rotatingmmf wave in accordance with Lenz's law, which states that an induced emfproduces a current that will ow in the direction to oppose the change whichproduced it [25].
If the rotor were to rotate at synchronous speed, no induction would takeplace and hence no torque would be produced. At any other speed, rotorcurrent is induced and asynchronous torque is developed, hence the nameasynchronous machine. The dierence between the synchronous speed ns andthe rotor shaft speed nr is called the slip speed and is commonly referred toas the slip of the motor. Slip s in per unit is
s =ns − nr
ns
=ωs − ωr
ωs
=ωsl
ωs
, (2.22)
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 2. LITERATURE REVIEW 16
where ωsl is the slip frequency. Slip is generally expressed as a percentage ofthe synchronous speed. From (2.22) the rotor speed is
nr = (1− s)ns, (2.23)
or
ωr = (1− s)ωs. (2.24)
While the rotor is stationary, the slip is 1 per unit and the rotor now behavesthe same as a secondary winding on a transformer. The rotor frequency isthe same as the stator frequency until the rotor starts to turn. The relativemotion between the stator ux and rotor is responsible for the induced emf inthe rotor windings and the frequency of the induced emf in the rotor is
fr = sfe, (2.25)
where fe is the electrical frequency of the applied currents in the stator. If theper unit slip value is between 1 and 0 the asynchronous machine is operatedas a motor and hence electrical power is converted to mechanical power. From(2.24) a per unit slip value smaller than 0 indicates the rotor shaft speedis greater than the synchronous stator speed. To achieve this, torque hasto be applied to the shaft of the induction machine and thus the inductionmachine now functions as a generator. The stator terminals are then used asa voltage source to supply electrical power [26]. Applications include the useof an induction generator driven by a wind turbine and connected to a powersystem to provide renewable energy. This thesis will use an induction machineoperated at negative slip in order to supply power to a load.
2.4.2 Equivalent Circuit
A transformer functions by transferring power exclusively to windings throughinduction. An induction machine can thus be seen as a transformer witha rotating secondary winding [27]. Due to the similarity between inductionmachines and transformers, the model of an induction machine is similar tothat of a transformer. A single-phase equivalent circuit of a squirrel-cageinduction motor is shown in Fig. 2.7.
The stator side is modelled as an inductor Lls in series with a resistor Rs
to represent the leakage impedance. The air gap model consists of two shuntcomponents, a core-loss resistance Rc and a magnetizing inductance Lm. In-duction machines are usually operated at low slip and hence the frequency ofthe induced rotor emf is also low. Due to this, the core loss of the rotor mag-netic circuit is often ignored [27]. The rotor is modelled as leakage inductanceLlr in series with rotor resistance Rr. The rotor quantities are however referredto the stator side assuming a turns ratio of 1:1. The impedance of the rotorreferred to the stator is
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 2. LITERATURE REVIEW 17
−Vs
+
Rs
Is
Lls Llr
IrRr
SLmRc
Figure 2.7: Single-phase equivalent circuit of an induction motor.
Zr = Rr + jωslLlr. (2.26)
After substituting (2.22) in the above equation and simplifying, (2.26) becomes
Zr =Rr
s+ jωsLlr. (2.27)
The rotor reactance is now dened in terms of the stator frequency and canbe analysed as part of the stator equivalent circuit as shown in Fig. 2.7.
From the equivalent circuit the necessary power equations can be deter-mined. For a three-phase motor the input power is
Pin = 3VsIscos(θ), (2.28)
where θ is the phase dierence between the applied stator voltage Vs and thecurrent Is owing through the stator winding. The stator winding copper lossis
Pstator = 3Is2Rs. (2.29)
The power transferred across the air gap is
Pgap =3Ir
2Rr
s. (2.30)
The rotor copper loss is
Protor = 3Ir2Rr. (2.31)
The electromechanical power developed by the motor is thus
Pmech = Pgap − Protor = 3Ir2Rr
(1− s
s
)= (1− s)Pgap. (2.32)
The electromechanical torque produced can be calculated since mechanicalpower is equal to angular velocity times torque [26]. With reference to (2.24)the mechanical torque expression becomes
Tmech =Pmech
ωr
=Pgap
ωs
= 3Ir2 Rr
sωs
. (2.33)
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 2. LITERATURE REVIEW 18
The current owing in the rotor is thus responsible for creating the requiredmotor torque. The torque applied to the shaft of the motor will equal themechanical torque minus the friction, windage and other rotational losses.This equivalent circuit and derived equations of an induction machine can beused to model and predict the behaviour of induction machines.
2.4.3 Dynamic d-q Model
For a dynamic model of a three-phase machine, three equivalent circuits arerequired, one for each phase. In order to simplify and reduce the total numberof dierential equations, a variable transformation rst formulated by R.H.Park in the 1920's can be applied to the equivalent circuit. The Park trans-formation refers stator variables to a synchronously rotating reference framewith respect to the rotor. The transformation has since been generalised suchthat machine variables are referred to a frame of reference that is rotatingat an arbitrary angular velocity. The preferred transformation to analyse thedynamic behaviour of a balanced three phase machine is where both the statorand rotor variables are transformed to a synchronously rotating dq0 referenceframe that turns with the rotating magnetic eld. This transformation turnsthe voltage and current ac quantities into dc quantities allowing better analysisof the machine's transient responses [19].
The dq0 transformation transforms the three-phase abc system into two ro-tating components, namely the direct (d) and quadrature (q) axis. The q-axisis aligned with the a-axis and precedes the d-axis by 90°. A third component,the zero-sequence component, is also included to complete the transformationof the three phase variables. However, under balanced three phase conditions,no zero-sequence components are present [26]. The formulation of the dq0transform is explained in [28] and is given by
ξqd0 = K(φ)ξabc, (2.34)
where
K(φ) =2
3
cosφ cos(φ− 2π3) cos(φ+ 2π
3)
sinφ sin(φ− 2π3) sin(φ+ 2π
3)
12
12
12
. (2.35)
The inverse dq0 transform is used to transform the variables back to the abcreference frame and is given by
ξabc = K−1(φ)ξqd0, (2.36)
where
K−1(φ) =
cosφ sinφ 1cos(φ− 2π
3) sin(φ− 2π
3) 1
cos(φ+ 2π3) sin(φ+ 2π
3) 1
. (2.37)
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 2. LITERATURE REVIEW 19
The angle φ is equal to ωet for the synchronously rotating reference framewhere all variables are referred to the stator side.
The dynamic d-q model [29] of the equivalent circuit analysed in Section2.4.2 is shown in Fig. 2.8. The core-loss resistance is omitted for this model.Note, the rotor side is short-circuited as would be the case for a squirrel-cageinduction machine.
−vqs
+
Rs
iqs ωeλds
+ − Lls Llr
iqr (ωe − ωr)λdr
− +
RrLm λqrλqs
(a)
−vds
+
Rs
ids ωeλqs
− + Lls Llr
idr (ωe − ωr)λqr
+ −
RrLm λdrλds
(b)
Figure 2.8: Dynamic d-q model of a squirrel cage induction motor for the (a)q-axis component and (b) d-axis component.
From Fig. 2.8 the following dierential equations can be deduced:
vqs = Rsiqs +d
dtλqs + ωeλds, (2.38)
vds = Rsids +d
dtλds − ωeλqs, (2.39)
vqr = 0 = Rriqr +d
dtλqr + (ωe − ωr)λdr, (2.40)
vdr = 0 = Rridr +d
dtλdr − (ωe − ωr)λqr, (2.41)
where the ux linkages can be written as:
λqs = Llsiqs + Lm(iqs + iqr), (2.42)
λds = Llsids + Lm(ids + idr), (2.43)
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 2. LITERATURE REVIEW 20
λqr = Llriqr + Lm(iqs + iqr), (2.44)
λdr = Llridr + Lm(ids + idr). (2.45)
From [16] the derived torque expression for the developed motor torque in d-qform becomes:
Te =3
2Lm(iqsidr − idsiqr). (2.46)
In variable frequency applications the rotor speed ωr is not a constant and canbe related to the torque by
Te = TL + Jdωr
dt, (2.47)
where TL is the load torque and J is the rotor inertia.Equations (2.38) to (2.47) provide the complete model of an induction
machine and will be used for modelling purposes.
2.5 Summary
This chapter gave the foundation of the research covered in this thesis. Pre-vious work regarding the use of ram air turbines as generators was presented.Bidirectional converter and inverter topologies were investigated, along withpulse-width modulation schemes that are used to control both inverter andconverter circuits. The operation and development of the equivalent circuitfor an asynchronous machine were also presented.
Stellenbosch University https://scholar.sun.ac.za
Chapter 3
DC-DC Converter Design and
Implementation
This chapter describes the design of a non-isolated bidirectional current-controlledvoltage-regulated switched-mode dc-dc converter. To facilitate the regulationof the dc bus voltage, two control loops (an inner and an outer) are employed.A detailed design of the control circuitry is given and the closed-loop stabilityis investigated. The design of a two-stage soft-start circuit to limit the inrushcurrent from the battery to the dc bus during start-up is presented as well asthe design of the gate-drive circuitry.
3.1 Basic Circuit Operation
The dc bus voltage is required to be regulated in the range of 340 V to 360 V(350 V nominal), regardless of the direction of current ow through the bat-tery. The dc bus is divided between a positive (+175 V) and negative (-175 V)rail to achieve a nominal rail-to-rail voltage of 350 V. Each rail section has itsown bidirectional converter and is designed to function independently usingidentical controllers. The chosen converter topology is two bidirectional half-bridge circuits, one on top of the other, as shown in Fig. 3.1 with the midpointgrounded.
Twenty 12 V, 40 Ah rated deep-cycle lead-acid batteries are divided be-tween the two rail sections, each using ten batteries indicated by Vbat. Table 3.1shows the characteristic values of a single BSB DB12-40 battery with a oatinglife expectancy of 12 years. The internal battery resistance, equivalent seriesresistance of the inductor and the resistance of the switches are all accountedfor by one representative resistor in the converter denoted Rin in Fig. 3.1.
For the duration of this thesis small letter variables refer to instantaneousvalues while capital letter variables refer to average values.
If the average inductor current IL is positive, current is owing from the dcbus into the batteries, thus charging the batteries. A negative average induc-
21
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 22
Rin
Vbat
LiL vm
id2Q2
iQ2
id1Q1
iQ1
ic
C
+Vdc
idc
Rin
Vbat
LiLvm
id3Q3
iQ3
id4Q4
iQ4
C
−Vdc
Figure 3.1: Bidirectional dc-dc converter circuit.
Table 3.1: BSB DB12-40 battery specications.
Parameter Description ValueVb Rated battery voltage 12 VQAh Charge capacity 40 AhVbf Maximum oat use charge voltage 13.8 VVbc Maximum cycle use charge voltage (10 A max) 14.8 VRb Internal resistance at full charge 9.5 mΩ
tor current indicates current owing from the batteries into a load connectedbetween the dc bus rails. To ensure the 2 kW bidirectional power ow spec-ication is met, the maximum and minimum average inductor current IL isset to ±10 A corresponding to ±2.4 kW of power at the battery's side of theconverter. Also, the recommended maximum charging current of the batteriesis given in the datasheet as 10 A.
Discrete MOSFETs are used as the switches for Q1 through Q4 and theswitching frequency is set to 40 kHz.
With reference to Fig. 3.1, Fig. 3.2 shows the inductor voltage and currentfor the top-half of the converter circuit when the batteries are being chargedat a maximum mean current of 10 A. When switch Q1 is on (and Q2 is o)the voltage across the inductor is vL(max) = Vdc-Vbat = 55 V. Similarly, when
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 23
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
·10−5
−200
−100
0
100
v L(V
)
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
·10−5
89101112
i L(A
)
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
·10−5
0
100
200
v Q1(V
)
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
·10−5
0
100
200
v Q2(V
)
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
·10−5
0
5
10
15
i Q1(A
)
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
·10−5
0
5
10
15
Time (s)
i d2(A
)
Figure 3.2: Current and voltage waveforms describing the operation of theconverter when the batteries are being charged at maximum current.
switch Q2 is on (and Q1 is o) the inductor voltage is vL(min) = -Vbat = -120 V.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 24
Since
vL = LdiLdt
, (3.1)
the inductor current changes linearly with time. However in order that theinductor conduct a current that is on average (typically averaged over 5 to 10switching cycles) constant, the time averaged voltage applied to the inductormust be zero. That is to say VL = 0. Thus
t1(on)(Vdc − Vbat) + t2(on)(0− Vbat) = 0, (3.2)
where t1(on) is the on-time of switch Q1 and similarly t2(on) is the on-time ofswitch Q2. During individual switching cycles, however, the inductor currentramps up linearly through the MOSFET ofQ1 from its minimum value (iL(min))to its maximum value (iL(max)) during the on-time of switchQ1. The on-time ofa switch with respect to the total switching period is more commonly referredto as the duty cycle and for Q1 the duty cycle is
D1 =t1(on)Ts
=Vbat
Vdc
=120
175= 0.6857, (3.3)
as derived in Section 2.2.2. Similarly the inductor current ramps down fromits maximum to its minimum value during the on-time of Q2. Thus the dutycycle of Q2 is given by
D2 = 1−D1 =t2(on)Ts
=Vdc − Vbat
Vdc
=55
175= 0.3143. (3.4)
The voltage across the drain and source of switches Q1 and Q2 are shown inFig. 3.2. Due to the direction of current when the mean inductor current ispositive, current can only ow through the MOSFET of Q1 and through thediode of Q2. These current waveforms are also shown in Fig. 3.2 where iQ1
denotes the MOSFET current through Q1 and id2 denotes the diode currentthrough Q2 as indicated in Fig. 3.1.
The bottom-half of the converter will function in a similar manner wherethe duty cycle of switch Q4 equals that of Q1 and the duty cycle for Q3 is thesame as Q2's. The inductor current and voltage waveforms also remain thesame.
If the mean inductor current of 10 A were to change to -10 A, the basiccircuit operation stays the same, however current will ow in the oppositedirection, thus supplying current to the dc bus. Due to the change of currentdirection, current will now ow through the diode of Q1 and through theMOSFET of Q2 as shown in Fig. 3.1 and Fig. 3.3. The voltage across theinductor and switches will remain the same as before. The duty cycles of theswitches also remain the same.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 25
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
·10−5
−12−11−10−9−8
i L(A
)
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
·10−5
0
5
10
15
i d1(A
)
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
·10−5
0
5
10
15
Time (s)
i Q2(A
)
Figure 3.3: Current waveforms describing the operation of the converter whenthe batteries are sourcing maximum current.
3.1.1 Inductor Design
Both inductors labelled L in Fig. 3.1 are wounded from Litz wire with a3.8 mm2 total cross-sectional area (120 strands of 0.2 mm diameter enam-elled copper wire) wound over an Arnold A-866142-2 powdered-iron toroidalcore. The current density in the wire at a maximum average inductor cur-rent of 10 A is 2.65 A/mm2. The inductor is designed to have a peak-to-peakcurrent ripple of approximately 22.5 % of the absolute maximum value of themean battery current, thus 2.25 A.
Since the voltage applied across the inductor is constant while either Q1 orQ2 is on, then, from (3.1) and (3.2):
vL∆t = L∆iL
= vL(max)t1(on)
= −vL(min)t2(on).
(3.5)
Rearranging and substituting (3.3) and (3.4) into the above equation yields
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 26
∆iL =vL(max)D1Ts
L
= −vL(min)D2Ts
L
=37.7Ts
L,
(3.6)
where vL(max) and vL(max) were already determined as 55 V and -120 V. Witha peak-to-peak inductor current ripple of 2.25 A and given that
Ts =1
fs=
1
40 kHz= 25 µs, (3.7)
the value of the inductor L is thus calculated from (3.6) and (3.7) as 420 µH.The inductance rating, also known as the AL value of a core, is used to de-termine the amount of turns required to achieve a specic inductance for thatparticular core. The AL value for the Arnold A-866142-2 core is given in thedatasheet as 142 nH/turn2. The required number of turns is thus:
N =
√L
AL
= 54.32 ≈ 55. (3.8)
To ensure the inductor core will not saturate, additional calculation are needed.From Ampere's Law [26] it is known that
Hc =NILle
, (3.9)
where Hc is the average magnetic eld intensity in the core and le is the meanux path (core) length. Converting (3.9) into Oersted by multiplying with theconversion factor of 4π
1000, results in a magnetic eld intensity of
Hc =4πNIL1000 le
=(0.4)π(55)(10)
19.612= 35.24 Oe. (3.10)
The following relationship applies provided the inductor core is not saturated:
Bc = µHc, (3.11)
where Bc is the magnetic core ux density and µ the permeability given as125 in the datasheet. The resulting ux density in Kilogauss is thus 4.41 KG.When comparing these values to the datasheet showing the B-H curve of theA-866142-2 core in Oersted and Kilogauss, it can be seen that the core will onlysaturate for a magnetic eld intensity greater than 100 Oe. The magnetic corewill thus not saturate for the designed conditions. The B-H curve is howevernot completely within the linear approximation range of the permeability atthis point.
The graph indicates a true ux density of approximately 3.4 KG resultingin a permeability closer to 90 when a mean current of 10 A ows through the
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 27
55 core windings which will inuence the amount of inductor current ripple.Inductance can also be calculated as [26]
L =N2µAc
lc, (3.12)
where Ac is the cross-sectional area of the core. From the above equation itcan be seen that a decrease in permeability (with all other variables remainingconstant) will cause a decrease in inductance. A lower inductance with thesame amount of voltage applied across it for the same duration of time willultimately result in a larger current ripple as deduced from (3.6). The currentripple at a mean inductor current of zero or close to zero will thus have apeak-to-peak value equal to 2.25 A as designed for, however at a mean inductorcurrent of 10 A the inductor ripple will rise to a value closer to 3 A.
3.1.2 Bus Capacitor Design
The assumption is made that the mean current IQ1 owing through switch Q1
equals the mean dc bus current Idc, which is either sourced by another circuitor sunk by a resistive load connected between the dc bus rails. Thus,
Idc = IQ1 = ILD1, (3.13)
where IL is the average inductor current and D1 is dened in (3.3). The meandc bus voltage is assumed constant at 175 V, hence only the ripple componentof iQ1 ows through capacitor C. At a maximum average inductor currentof 10 A, the corresponding currents iQ1 and ic are shown in Fig. 3.4. Thecapacitor current ic is negative simply due to the positive reference directionchosen for the current as indicated in Fig 3.1.
By denition current is equal to the rate at which charge ows, hence:
ic =dq
dt, (3.14)
where q denotes electric charge. In integral form the above equation becomes
q =
∫icdt. (3.15)
From Fig. 3.2 and Fig. 3.3 it can be seen that the maximum magnitude of thecapacitor current ic will remain the same whether the mean inductor currentis positive or negative. From (3.15) the net charge ∆q, while either switch Q1
or Q2 is on, is calculated as the time integral of the capacitor current ic duringthat time period. With reference to Fig. 3.4 and while Q1 is o and Q2 is on,the net charge is calculated as an area where
∆q = IQ1D2Ts = (ILD1)D2Ts = 53.88 µC. (3.16)
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 28
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
·10−5
0
5
10
15
i Q1(A
)
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
·10−5
−5
0
5
10
i c(A
)
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
·10−5
175
Time (s)
v dc(V
)
Figure 3.4: Waveforms indicating the capacitor current and voltage when thebatteries are being charged at the maximum current.
From rst principles it is known that charge is also equal to capacitance timesvoltage, hence:
q = Cvc, (3.17)
where vc is the voltage across the capacitor. The ripple component on the dcbus voltage is chosen as 30 mV (approximately 0.02 % of the mean dc busvoltage). The value of the capacitor required at the dc bus is thus calculatedas 1796 µF from (3.16) and (3.17).
Two parallel connected 1000 µF, 200 V Panasonic electrolytic capacitorswere chosen for each of the dc bus capacitors C shown in Fig. 3.1. The resultingvoltage ripple on the dc bus as shown in Fig. 3.4 is thus 27 mV. The rmsvalue of the capacitor current ic shown in Fig. 3.4 was calculated in MATLABas 4.54 A. Since two parallel connected capacitors are used, the rms currentthrough each of them is 2.27 A. The datasheet gives the maximum operatingrms ripple current as 4.85 A for one capacitor. The rms ripple current is thuswell within the allowed operating range.
The nal component values used in the bidirectional dc-dc converter circuitis given in Table 3.2.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 29
Table 3.2: Converter circuit parameters.
Parameter ValueVdc 175 VVbat 120 VRin 1 ΩL 419 µHC 2000 µF
3.1.3 MOSFET Power Loss and Heat Sink Design
The switches used for Q1 through Q4 are International Rectier's IRFP264MOSFETs. Table 3.3 shows the characteristic values of the chosen MOSFETs.
Table 3.3: IRFP264 MOSFET parameter values.
Parameter Description ValueVDSS Drain-to-source breakdown voltage 250 VRDS(on) Static drain-to-source on-resistance (VGS = 10 V) 75 mΩID Maximum continuous drain current (Tc = 25°) 38 AQg Maximum total gate charge 210 nCtc(on) Turn-on (rise) time 99 nstc(off) Turn-o (fall) time 92 ns
The calculation of MOSFET power losses for half-bridge topologies are de-scribed in detail in [30] and [31] where the diodes are assumed ideal. For thebidirectional dc-dc converter circuit two cases have to be investigated, one forpositive inductor current and one for negative inductor current.
CASE 1: Positive Inductor Current
Current can only ow from the dc bus into the batteries when switch Q1 isclosed since the diode in Q1 is reversed biased. Switching as well as conductionlosses are thus present. Due to the direction of the inductor current, currentwill however ow through the freewheeling diode of Q2 causing diode losses.
When a MOSFET transitions from on to o and similarly from o to on,there is a small time period (tc(on) and tc(off)) during which the switch hasa voltage across it as well as a current owing through it. This results inswitching losses. In Fig. 3.2 the current through Q1 is shown. The absolutemaximum and minimum inductor current is 11.125 A (iL(max)) and 8.875 A(iL(min)) respectively, which corresponds to an average maximum current of10 A with 2.25 A of ripple. The average switching losses for Q1 are thuscalculated as [31]
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 30
Pc(on) =1
2VdciL(min)tc(on)fs, (3.18)
and
Pc(off) =1
2VdciL(max)tc(off)fs. (3.19)
The total switching losses for Q1 are thus
Pswitch =1
2Vdcfs
(iL(min)tc(on) + iL(max)tc(off)
)= 6.66 W.
(3.20)
The conduction losses occur due to the on-resistance of the switches and aregiven by [31]
Pcond = RDS(on)I2Q(rms). (3.21)
When switch Q1 is turned on the voltage across the inductor changes almostinstantaneously from −Vbat to Vdc−Vbat while the current through the inductorrises gradually from iL(min) to iL(max) at a rate of
diLdt
= m =Vdc − Vbat
L=
VL(max)
L, (3.22)
The rms value of the current through Q1 is calculated as [31]
I2Q1(rms) =1
Ts
∫ Ts
0
i2Q1dt
=1
Ts
∫ D1Ts
0
(iL(min) +mt
)2dt
=1
Ts
[1
3m
(iL(min) +mt
)3]D1Ts
0
=1
3mTs
[(iL(min) +mD1Ts
)3 − i3L(min)
]= 68.9 A2,
(3.23)
where D1 was calculated in (3.3). The total losses for Q1 are thus
PQ1 = Pswitch + Pcond = Pswitch +RDS(on)I2Q1(rms)
= 6.66 W+ 5.16 W = 11.82 W.(3.24)
The power loss due to the forward voltage of the MOSFET's freewheelingdiode is
PD = VSDId, (3.25)
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 31
where Id is the mean current owing through the diode over one switchingperiod. The graph in the IRFP264 MOSFET datasheet shows a forward diodevoltage (VSD) of approximately 0.8 V at 10 A of current. The mean diodecurrent of Q2 is calculated as
Id2 = ILD2 = (10)(0.3143) = 3.143 A. (3.26)
Thus, the diode conduction power loss in Q2 is
PD2 = VSDId2 = (0.8)(3.143) = 2.51 W. (3.27)
The total power losses for the top-half of the converter circuit whilst chargingthe batteries are
PT = PQ1 + PD2 = 14.33 W. (3.28)
The total power losses of the bottom-half circuit are however also equal to PT
since the losses in Q4 are equal to that of Q1 and similarly the losses of Q3 areequal to that of Q2 due to the identical circuit operation described earlier. Thetop- and bottom-half of the converter circuit will be manufactured on separateprinted circuit boards, each with their own heat sink. The power losses for thetwo sections are thus not added together.
CASE 2: Negative Inductor Current
For the case where current will ow in order to charge the dc bus, the switcheswill operate dierently. Current will now ow through the diode of Q1 (causingdiode conduction losses) and through the MOSFET of Q2 (causing switchingas well as conduction losses).
First consider the losses in Q2. The switching losses in Q2 are identicalto the switching losses calculated for Q1 in (3.20) in the previous case (seecurrent waveforms in Fig. 3.3). The total switching losses for Q2 are thus also6.66 W.
The MOSFET conduction losses will dier from the previous case since theduty cycle of Q2 is dierent to that of Q1. Equation 3.22 describing the rateof change of the inductor current through Q2 is written as
diLdt
= m =−Vbat
L=
VL(min)
L. (3.29)
The rms current through Q2 is thus
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 32
I2Q2(rms) =1
Ts
∫ Ts
0
i2Q2dt
=1
Ts
∫ D2Ts
0
(iL(min) +mt
)2dt
=1
3mTs
[(iL(min) +mD2Ts
)3 − i3L(min)
]= 31.56 A2,
(3.30)
where iL(min) is -8.875 A and D2 is 0.3143. From (3.21) and (3.30) the con-duction loss for Q2 is calculated as 2.37 W resulting in a total power loss of9.03 W for Q2.
From (3.26) the mean current Id2 through the diode of Q1 is 6.857 A atan absolute mean inductor current of 10 A and duty cycle D1. Substitutingthe value for Id2 into (3.25), the diode conduction loss for Q1 is calculated as5.49 W. The total power losses for the top-half of the converter circuit whilstcharging the dc bus are
PT = PD1 + PQ2 = 14.52 W. (3.31)
The total power losses for the bottom-half of the converter circuit will be thesame as for the top-half.
From the two cases the most power is dissipated when current is owingfrom the batteries into the dc bus. The power losses from the operation of theswitches are converted into heat and thus an adequate heat sink is required.The heat sink is designed for the worst case scenario, thus the maximum totalpower of the two cases are used in the heat sink design which is 14.52 W. Therise in junction temperature due to the power dissipated in the MOSFETs anddiodes are given by [11]
∆Tj = PT (Rθjc +Rθcs +Rθsa) , (3.32)
where Rθjc, Rθcs and Rθsa are the junction-to-case, case-to-sink and sink-to-ambient thermal resistances of the MOSFETs. The assumed maximum rise injunction temperature is 50°C. From the datasheet Rθjc and Rθcs are given as0.45°C/W and 0.24°C/W respectively. The sink-to-ambient thermal resistanceis calculated as 2.75°C/W from (3.32). The size of the aluminium heat sinkavailable is 190 mm × 150 mm with a base thickness of 5 mm which givesa sink-to-ambient thermal resistance of 0.37°C/W [32], which is more thanadequate.
3.1.4 Soft-start Circuit
A two-stage soft-start circuit comprising pre- and post-charge mechanisms [33]are used to increase the dc bus voltage gradually to the desired 350 V rail-to-rail
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 33
voltage.First consider the top-half of the converter circuit of Fig. 3.1 used to gener-
ate the +175 V bus voltage. Initially the dc bus voltage is at ground potential.Without any soft-start mechanism in place, as soon as the circuit is switchedon, in-rush currents will charge the dc bus to the equivalent battery voltagethrough the MOSFET's freewheeling diodes. The pre-charge circuit shownin Fig. 3.5 is used to limit the battery in-rush current to 5 A by means offour parallel-connected 100 Ω resistors which are placed in series with theconverter's battery. At such time as the bus capacitor C is charged to theequivalent battery potential (≈120 V), the resistors are short-circuited by arelay to eectively remove the resistors from circuit. Table 3.4 show the valuesof the components used in the pre-charge circuit.
LM311R1
VC
D1
VE
D2
C1
R2
+Vdc
R3
VC
Rp
LiL
Half-Bridge
Converter
+Vdc
VbatRg
Q2
VE
Figure 3.5: Pre-charge soft-start circuit implementation.
Table 3.4: Pre-charge soft-start circuit component values.
Component Description ValueVC - 12 VVE - 0 VRp - 25 ΩQb 2N2219A NPN BJT Ic = 0.8 ARg - 10 kΩD1 Zener diode 5.6 VD2 1N4148 switching diode VF = 0.72 VC1 - 10 µFR1 - 1 kΩR2 - 220 kΩR3 - 12 kΩ
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 34
The relay action is controlled by an LM311 comparator. A 5.6 V Zener diodeis used as the reference voltage at the non-inverting input of the comparator,as shown in Fig. 3.5. A resistor divider circuit (R2 and R3) is used to scaledown the dc bus voltage to the inverting input of the comparator. Oncethe scaled dc bus voltage exceeds the reference voltage, the inverting outputof the comparator will change to a high state. This allows current to owthrough the LM311's output transistor, thus closing the relay switch to createa short-circuit across the four parallel connected power resistors represented byRp. Capacitor C1 was added to introduce a small time delay in order to avoidany false state changes at the comparator's output during start-up. With thegiven resistor values the relay will trip at a measured dc bus voltage of
Vdct =
(R2 +R3
R3
)Vref = 108.3 V, (3.33)
which is low enough to avoid the relay from tripping while the converter issupplying power to a load, since the load can cause the battery voltage todrop slightly.
As seen in Fig. 3.5 the pre-charge circuit is connected in series between theconverter's batteries and inductor. The rest of the converter circuit remainsunchanged and is shown in Fig. 3.5 as a block diagram. The pre-charge cir-cuit receives power from one of the ten 12 V batteries. To accommodate therequired current to turn the relay switch on and o properly, an additionalNPN BJT was added. Together with the LM311's output BJT they form aDarlington pair [34] where Rg is used to drain the capacitance.
A similar pre-charge circuit is used for the bottom-half of the convertercircuit to gradually charge the dc bus to -120 V. The only dierence is thescaled negative dc bus voltage feeds into the non-inverting input of the com-parator and a negative 5.6 V reference voltage is used at the inverting input.The complete circuit schematic is shown in Appendix A.1.1.
The next stage of the soft-start circuit is used to charge the dc bus from theequivalent battery potential to the respective +175 V and -175 V rails. Thepost-charge mechanism is designed to limit current spikes once the switchesstart switching. To achieve this, the on-times of the switches have to graduallyincrease thus allowing the post-charge mechanism to control the duty-cycle ofthe switches by adjusting the dead-time.
At the instant Q1 is switched o, there is a small but nite time delaybefore switch Q2 is switched on, namely the dead-time. During this timeperiod both switches remain o. The practical implementation of dead-timefor switches Q1 and Q2 are demonstrated in Fig. 3.6. The modulator referencesignal ve (which is generated by the control circuit) is indicated in black andthe triangular carrier waveform vc is indicated in green in Fig. 3.6. If nodead-time were implemented, both switches will have a duty cycle of 50 %. Anoset voltage Voffs is subtracted from ve to generate the modulator reference
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 35
Voltage ve1
vcve2ve
Time
Voltage vpwm1
vpwm2
Figure 3.6: Dead-time implementation.
voltage ve1 for switch Q1 and similarly Voffs is added to ve to generate themodulator reference voltage ve2 for Q2.
The on-time of the switches are determined as follows: if ve1 is greater thanvc, Q1 is on and if ve2 is less than vc, Q2 is on. The resulting PWM signalsfor Q1 and Q2 are shown in the bottom graph of Fig. 3.6. The implementeddead-time is seen to have reduced the duty cycles of both switches. Thedead-time is adjusted by changing the magnitude of the oset voltage Voffs.
The post-charge circuit functions by initially setting Voffs to a value largerthan the magnitude of the carrier waveform, thus both switches will have a dutycycle of zero. After the dc bus has reached the equivalent battery potential,the oset voltage is progressively ramped down over a period of approximately10 s to a minimum value, thus limiting the on-time and duty cycles of bothswitches, but allowing them to start switching.
The ramping down of the oset voltage is controlled by a capacitor andresistor circuit as shown in Fig. 3.7 with the component values given in Ta-ble 3.5.
Table 3.5: Post-charge soft-start circuit component values.
Component ValueVRef 5 VC2 470 µFD3 VF = 0.72 VR4 170 kΩR5 1.8 kΩR6 5.6 kΩ
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 36
VRef
C2
D3
R4
R5
R6
−
+
LM358N
R5
R6
Voffs
Figure 3.7: Post-charge soft-start circuit implementation.
As the voltage across C2 increases when charged through R5 and R6, Voffs
decreases until the minimum value of 0.44 V is reached. The minimum valuecorresponds to a dead-time of 1 µs and was chosen as such to eliminate anyMOSFET shoot-through currents. Resistor R4 is adjustable to increase ordecrease the dead-time.
3.1.5 Isolated Gate-Drive Circuitry
An isolated gate-drive circuit is employed to communicate the switching signalsgenerated by the converter's control circuit to the gate of each MOSFET.Galvanic isolation between the switches and the control circuit is essentialto prevent damage to the control circuit due to the large voltage excursionsencountered by the MOSFETs.
Each isolated gate-drive circuit comprises an optocoupler [11], a simplepush-pull oscillator circuit (oscillating at approximately 300 kHz), a smalltoroidal transformer (with split primary and secondary windings), and twofast rectier diodes. The oscillator supplies power to the optocoupler via thetransformer and fast rectier diodes as shown in Fig. 3.8, with the correspond-ing component values given in Table 3.6, while the optocoupler communicatesthe switching signals to the gate of a MOSFET.
S1 S2
R1 R1
C1 C1R2
VE
VC
T1
1:1
D1
D1
C2
VSC
VSE
Figure 3.8: Power section of the isolated gate-drive circuit.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 37
Table 3.6: Component values used in oscillator gate-drive circuit.
Component Description ValueVC Battery supply voltage 12 VVE Battery supply voltage 0 VS1 PNP BC556 BJT Ic = 100 mAR1 - 220 ΩC1 - 100 pFR2 - 1 MΩD1 BAT85 Schottky diode IF = 200 mAC2 - 100 µFVSC Isolated supply voltage 12 VVSE Isolated supply voltage 0 V
An EPOS B64290 toroid was wounded with 10 turns of thin enamel copper wireon both sides to construct the transformer T1 with a turns ratio of one-to-one.The TLP250 was used for the optocoupler as it is equipped with under-voltagelockout protection. The characteristics of the TLP250 is given in Table 3.7.
Table 3.7: TLP250 characteristic values.
Parameter Description ValueVCC Supply voltage 10-35 VIO Output current ±1.5 AViso Isolation voltage 2500 Vtp Propagation delay time 150 nsCM Common mode transient immunity 5000 V/ µs
OscillatorPowerCircuit
VSC
VSE
PWMSignal
Rp
Anode
Cathode
TLP250
Vout
RG
RPD
D
Figure 3.9: Block diagram of the MOSFET gate-drive circuit.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 38
Four gate-drive circuits as shown in the block diagram of Fig. 3.9 are required,one for each MOSFET. The value of the resistor Rp was chosen as 220 Ω tolimit the input current to the TLP250 to approximately 15 mA. A pull-downresistor RPD of 4.7 kΩ is used to drain the gate capacitance once the MOSFETswitches o. The gate resistor was chosen as 27 Ω. The maximum outputvoltage of the optocoupler with the described power circuit is 12 V, hence themaximum current the optocoupler is ever required to source is
IO =Vout
Rg
=12
27= 0.44 A. (3.34)
The datasheet of the TLP250 optocoupler species a maximum output currentof 1.5 A. The optocoupler will thus be able to source the above mentionedcurrent to the gate of the MOSFET.
The maximum voltage across any of the MOSFETs is the dc bus voltageof 175 V. With the turn-o time given as 92 ns in the MOSFET datasheetand the turn-on time as 99 ns, the minimum transient immunity required bythe optocoupler is 1902 V/µs. The 5000 V/µs rated transient immunity of theoptocoupler is thus sucient.
A detailed schematic showing the isolated gate drivers, pre-charge soft-startcircuit and complete converter circuit for both positive and negative dc busgeneration can be found in Appendix A.1.1.
3.2 PWM Controller Design
A block diagram of the two identical controllers (one for each half-bridge con-verter) is shown in Fig. 3.10. Two control loops are used; an inner currentcontrol loop which is enclosed by an outer voltage control loop. The controlleris structured such that the inner control loop measures and compares the meaninductor current IL to a set-point voltage vseti generated by the outer controlloop. The set-point voltage corresponds to a current reference where 1 V rep-resents 2 A of current. The output of the voltage controller vrefi is limited tovseti as shown in Fig. 3.10 to eectively limit the average inductor current. Theouter control loop measures and compares the dc bus voltage vdc to a xedset-point Vdc
∗ corresponding to a bus voltage of positive or negative 175 Vrespectively.
The current control loop is responsible for generating the PWM signals thatfeed the gates of the MOSFETs on the converter circuit. The output of thecurrent controller produces an modulator reference signal ve that is used forcomparison with the carrier waveform to generate the required PWM signals.The reference signal is limited as seen in Fig. 3.10 to ensure the amplitude ofthe reference signal never exceeds the amplitude of the carrier waveform. Itthus limits the duty cycle of the switches to a value below 100 %. The PWM
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 39
Vdc∗ Σ
+
−vfbv
VoltageController
vrefi
Limiter
vseti
Σ+
−vfbi
CurrentController
Limiter
PlantiL Bus
Plantvdc
CurrentSensor
VoltageSensor
ModulatorCircuitInner Loop
vε ve
Figure 3.10: Block diagram of converter's control circuit.
generation circuit is denoted in Fig. 3.10 as the modulator circuit and will bedescribed in more detail in the subsections that follow.
Frequency domain analysis is employed to design the voltage and currentcontrollers. However, a time domain analysis of the circuit is also requiredto linearise the system and to provide the necessary dierential equations toimplement a time domain simulation of the converter and controller in thenext chapter.
3.2.1 Carrier Waveform Generation
A 40 kHz triangular waveform is chosen as the carrier waveform to generatethe required PWM signals. A 4 MHz crystal-controlled square-wave oscillatoris scaled down to 40 kHz using two Hex D-type ip-ops, ve ip-ops fromeach in order to divide by 100. The resulting square-wave is integrated togenerate a triangular waveform. The dc component is removed (to prevent theintegration of a constant) by capacitively coupling the output of the integratorcircuit to the gain-stage op-amp (an LF353), which is also used to buer theoutput waveform. The circuit outline is shown in Fig. 3.11 and the componentvalues are given in Table 3.8. The triangular waveform at the output has apeak-to-peak value of 11 V.
The integrator is in actual fact a low-pass lter with a low corner frequencyof
fc =1
2πR3C2
= 159.2 Hz, (3.35)
to limit the dc gain while still functioning as an integrator at 40 kHz. Theproblem with using a pure integrator (by removing R3) is without any dcfeedback, the output voltage will tend to drift until the op-amp saturates dueto the op-amp's internal oset and bias currents. By adding R3, dc feedbackis provided and the output voltage is stabilised [35].
The not-gates used are provided by a 74HC14 hex inverting Schmitt trig-ger integrated circuit (IC) package containing 6 inverting buers. Both the
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 40
C1 C1
R1
X1
4 MHz
÷ 10
400 kHz
÷ 10
40 kHz
R2
C2R3
−
+
LF353N
R2C3
R3
C2
R4−
+
LF353N
R5
vtri
40 kHz
Figure 3.11: Circuit outline of the triangular carrier waveform generation.
Table 3.8: Triangular waveform generator component values.
Component ValueX1 4 MHzR1 10 MΩC1 22 pFR2 22 kΩR3 1 MΩC2 1 nFC3 100 nFR4 1.2 kΩR5 4.7 kΩ
74HC14 and 74HC174 IC packages require a supply voltage of 5 V.
3.2.2 Controller Plant
For the analysis and design of the controller only the top-half of the convertercircuit is used as shown in Fig. 3.12. The same control circuit is implemented inthe bottom-half of the converter. The plant portion of the converter includesthe current sense resistors used by the current control loop to generate thecurrent feedback measurement and is denoted by Rcs in Fig. 3.12.
With reference to Fig. 3.12 the equation describing the inductor current asa function of the mid-point voltage vm is given by
vm − Vbat − iL (Rin +Rcs) = LdiLdt
. (3.36)
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 41
Rin
Vbat
LiL vm
Q2
Q1ic
C
Rcs
+Vdc
idc
Figure 3.12: Bidirectional dc-dc converter circuit used for controller design.
The dierential equation for the inductor current is thus
diLdt
=vm − Vbat − iL (Rin +Rcs)
L, (3.37)
where
vm = s1vdc. (3.38)
The variable s1 dictates the state of switch Q1 where s1 = 1 indicates theswitch is on and s1 = 0 indicates Q1 is o. The mid-point voltage is thus ateither +Vdc or zero depending on the amplitude of the modulator referencevoltage with respect to the carrier waveform. If the reference voltage ve attime t is larger than the carrier waveform vtri, s1 = 1, otherwise s1 = 0.
For frequency domain analysis the eect of the battery voltage is omit-ted since it is considered a constant dc value. Hence, the mean voltage Vm
measured at the mid-point between switches Q1 and Q2 is given by
Vm = (sL+Rin +Rcs)IL, (3.39)
where IL is the mean battery and inductor current. Solving (3.39) the forwardtransfer function of the controller plant Gp(s), describing the mean inductorcurrent as a function of the mid-point voltage, is given by
Gp(s) =ILVm
=1
sL+Rin +Rcs
. (3.40)
3.2.3 Current Control Loop
The current control loop, shown in Fig. 3.13, functions to regulate the meaninductor current owing to and from the batteries by controlling the switchingactions of the MOSFETs. The time-varying modulator reference voltage ve is
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 42
compared (by means of a high-speed AD790JN comparator from Analog De-vices) to the triangular waveform described in Section 3.2.1, where the outputof the comparator provides the PWM switching signal.
IL∗
(Vseti)Σ
+
−Vfbi
Gc(s) KR Kdc Gp(s)
H(s)
ILVe D1
Figure 3.13: Current control loop block diagram.
A linear model of the modulator circuit is assumed, where the duty cycle ofthe PWM signal D is calculated as the ratio of the mean modulator referencevoltage Ve to the amplitude of the carrier waveform. Assuming the triangularcarrier waveform has a minimum value of 0 V and a maximum value of VR,the duty cycle D is:
D =Ve
VR
. (3.41)
To account for the linear PWM model constant KR is introduced in Fig. 3.13.From (3.41) the constant KR is expressed as the ratio of the duty cycle to themodulator reference voltage, thus
KR =D1
Ve
=1
VR
, (3.42)
where D1 is the duty cycle of switch Q1. With reference to Fig. 3.12 the meanmid-point voltage Vm is related to the duty cycle (as proven in Section 2.2.2)by
Vm = D1Vdc =Ve
VR
Vdc. (3.43)
For the design of the current controller the dc bus voltage is assumed constant,which is reasonable given that the current control loop is specically designedto respond signicantly faster than the voltage control loop. Constant Kdc isthus 175 V. From (3.43) the mid-point voltage is the duty cycle multiplied byKdc.
Ten parallel connected 0.68 Ω metal lm resistors were used for currentsensing. The feedback voltage Vfbi provided to the inverting input of thesummation in Fig. 3.13, is thus the product of the inductor current IL and thefeedback transfer function H(s). The current-sensing feedback circuit is shownin Fig. 3.14 with the corresponding component values given in Table 3.9.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 43
Rcs
IL
R1R2
R1
−
+
LF353N
Vfbi
R2
Figure 3.14: Inductor current sensing circuit.
Table 3.9: Component values for the current sense circuit.
Component ValueRcs 68 mΩR1 4.7 kΩR2 34.5 kΩ
A dierential signal from the current sense resistors feed an LF353N op-ampthrough resistors to add the required gain. The transfer function H(s) is thusgiven by
H(s) =Vfbi
IL=
R2
R1
Rcs. (3.44)
A gain of 7.34 (R2
R1) is used such that at a maximum current of 10 A the
feedback signal will be 5 V. Similarly in the time domain the current loopfeedback voltage is given by
vfbi =R2
R1
RcsiL. (3.45)
With reference to Fig. 3.13, the open-loop transfer function is
Gol(s) = Gc(s)KRKdcGp(s)H(s), (3.46)
where Gc(s) is the compensation amplier. The frequency domain analysisof Gol(s) without the transfer function of the compensation amplier presentis shown in Fig. 3.15. In adherence with the Nyquist stability criterion, theopen-loop transfer function with compensation is required to have both posi-tive phase and gain margins in order for the closed-loop transfer function tobe stable (not oscillate). The phase margin should, preferably, be at least 30°[36]. Furthermore, the unity gain crossover frequency of the open-loop transferfunction should be chosen according to the desired performance of the system,but is generally designed to be at around one-fth of the system's switchingfrequency or lower [37].
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 44
−40
−20
0
20
Magnitude
(dB)
100 101 102 103 104 105−90
−45
0
Frequency (Hz)
Phase
(deg)
Figure 3.15: Bode plot of current controller without compensation.
As seen from Fig. 3.15 the phase margin is positive, the unity crossover fre-quency is at 2.99 kHz and the dc gain is 17.8 dB. The corner frequency is at390 Hz. A compensation amplier will increase the bandwidth as well as thedc gain. Ideally the dc gain should strive to innity thus an additional pole isrequired.
In order to improve the bandwidth as well as the dc gain a PI or pole-zerocompensator [38] is required. As the name suggests, the last mentioned consistsof a pole-zero pair, where the pole and zero are placed to achieve the requiredclosed-loop stability of the system. The pole-zero and PI compensators are verysimilar, the only benet of using a pole-zero compensator is the additional polepresent at high frequencies. This allows a roll-o of -40 dB per decade belowthe 0 dB which helps to improve the ltering of switching noise. A pole-zerocompensation amplier as shown in Fig. 3.16 is chosen as the compensatorGc(s) for the current control loop due to the good transient response and loadregulation that it provides.
The dierential equations describing the pole-zero compensator are derivedby assuming the amplier is an inverting amplier with the non-inverting inputgrounded. The inverting input voltage is dened as vfbi − vseti. From rstprinciples it is known that the currents owing into the amplier inputs sumto zero, thus
vfbi − vsetiR1
+ C1dvedt
+ C2dvc2dt
= 0, (3.47)
where the voltage across capacitor C1 is the reference voltage ve. The voltage
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 45
−
+
LF353N
ve
R1vfbi
C1
R2C2
R1
vseti
C1
R2
C2
Figure 3.16: Pole-zero compensation amplier.
across C2 is dened as vc2 and can also be expressed by
vc2 = ve − ic2R2 = ve − C2R2dvc2dt
. (3.48)
Solving the above dierential equation yields
dvc2dt
=ve − vc2C2R2
. (3.49)
Substituting (3.49) into (3.47) and rearranging results in the dierential equa-tion for the modulator reference voltage ve where
dvedt
=vseti − vfbi
C1R1
+vc2 − veR1R2
. (3.50)
Similarly for the frequency domain analysis the transfer function Gc(s) is de-rived below. Firstly,
Zf =
(R2 +
1
sC2
)|| 1
sC1
=
(sC2R2 + 1
sC2
)|| 1
sC1
=1
sC2
sC2R2+1+ sC1
=sC2R2 + 1
s2C1C2R2 + s(C1 + C2),
(3.51)
where Zf is the combined amplier feedback impedance. The transfer functionfor Fig. 3.16 is thus
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 46
Gc(s) =Ve
Vseti − Vfbi
=Zf
R1
=sC2R2 + 1
R1 [s2C1C2R2 + s(C1 + C2)],
(3.52)
and in standard form
Gc(s) =sC2R2+1C1+C2
sR1
(sC1C2R2
C1+C2+ 1
)=
1
R1(C1 + C2)
sC2R2 + 1
s(
sC1C2R2
C1+C2+ 1
) .
(3.53)
The dc gain in decibels is
Gcgain = 20log10
[1
sR1(C1 + C2)
]. (3.54)
The two time constants are
τ1 = C2R2, (3.55)
and
τ2 =C1C2R2
C1 + C2
, (3.56)
with corresponding corner frequencies of
f1 =1
2πτ1=
1
2π
(1
C2R2
), (3.57)
and
f2 =1
2πτ2=
1
2π
(C1 + C2
C1C2R2
). (3.58)
As seen from (3.53), the pole-zero compensator has an integrator pole whichwill give the required innite dc gain. The rst corner frequency f1 should beless than or equal to the corner frequency of the open-loop response (390 Hz)shown in Fig. 3.15 in order to increase the bandwidth. The inductor currentripple will impose a ripple voltage on the reference signal Ve. To ensure theripple on the reference voltage is ltered out properly (to produce a constantreference voltage), the desired bandwidth for the current controller is at around
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 47
one tenth of the switching frequency, thus at 4 kHz. The second corner fre-quency should be below the 0 dB line thus at a value equal to or higher than4 kHz. Substituting (3.57) into (3.58) gives
f2 =
(C1 + C2
C1
), (3.59)
where the desired corner frequency for f1 is chosen as 100 Hz and for f2 as5 kHz, resulting in a relationship between C1 and C2 of
C2 = 49C1. (3.60)
The value chosen for C1 is 10 nF and by taking the above equation into account,the closest value for C2 is 470 nF. From (3.57) the calculated value of R2 is3.4 kΩ. The bandwidth of the controller will change from 2.99 kHz to 4 kHzby adjusting the gain factor Gcgain. Since the roll-o of the controller is at-20 dB/decade when it crosses the 0 dB line the amount of gain required tochange the controller's bandwidth is calculated as
∆Gcgain = 20log10(4000)− 20log10(2990) = 2.53 dB. (3.61)
From (3.54) and (3.61) the value of R1 is calculated as 3.99 kΩ. After test-ing and simulation some of the values had to be adjusted slightly. The nalcomponent values are given in Table 3.10.
Table 3.10: Component values for the current loop compensator.
Component ValueR1 3.3 kΩR2 4.7 kΩC1 10 nFC2 470 nF
The closed-loop transfer function of the complete current controller shown inFig. 3.13 is
Gcl(s) =ILVseti
=Gc(s)KRKdcGp(s)
1 +Gc(s)KRKdcGp(s)H(s). (3.62)
To ensure the stability of the system with the compensation amplier present,the open-loop response of the current controller is investigated. Fig. 3.17 showsthe Bode plot of both the open- and closed-loop responses. The open-loopresponse shows a phase margin of 53.8 degrees and a unity gain crossoverfrequency of 3.2 kHz. The closed-loop response shows a dc gain of 6.04 dBcorresponding to a static loop sensitivity of 2.0 A of battery current per volt of
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 48
excitation applied to the set-point input Vseti. Thus, at an absolute maximumset-point voltage of 5 V, the battery current will rise to 10 A as desired.
−100
−50
0
50
100
Magnitude
(dB)
100 101 102 103 104 105−180
−135
−90
−45
0
Frequency (Hz)
Phase
(deg)
Open-loopClosed-loop
Figure 3.17: Current controller bode plot.
In Fig. 3.17 the gain margin is shown to be innite as the phase plot of theopen-loop response never crosses the -180 degree line. Since an innite gainmargin cannot be realised in practice, a bifurcation test [39] was undertakento determine a more realistic prediction of the gain margin. A gain variableKB was added to the compensation circuit as shown in Fig. 3.18 such that thetransfer function for the compensation amplier changed to KBGc(s).
Vseti Σ+
−Vfbi
Gc(s) KB Kd Gp(s)
H(s)
IL
Compensator
Ve VeB
Figure 3.18: Current control loop with additional gain factor KB.
The gain variable KB was linearly increased from one to 40 in a MATLABSimulink model of the converter circuit and controller. The result of the bi-furcation test is shown in Fig. 3.19. An additional gain of 22 was added to thecompensation circuit, where the output reference voltage VeB feeds the PWM
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 49
comparator, before the output voltage Ve of the controller Gc(s) was observedto bifurcate. This indicates a more realistic gain margin of 26.8 dB. A positivegain and phase margin suggest that the current control loop is stable.
0 5 10 15 20 25 30 35−0.5
0
0.5
1
KB
Ve(V)
Figure 3.19: Bifurcation diagram used for stability analysis.
3.2.4 Current and Voltage Limiting Circuit
An active clamp circuit [35] shown as a limiter in Fig. 3.10 is used to limitthe absolute maximum of the mean current demanded from the battery. Thisaction is realised by clamping the range of the set-point current dictated by thevoltage control loop to -10 A and +10 A. The clamp circuit in Fig. 3.20 showsboth positive and negative voltage clamping where the reference voltages canbe made adjustable.
−
+
LF353N
Rlim
Vin
−Vref
−
+
LF353N
+Vref
Vout
Figure 3.20: Postive and negative active clamping circuit.
Since the current control loop has a static loop sensitivity of 2 A per voltof excitation at the set-point input, the values for -Vref and -Vref are set toplus and minus 5 V respectively, to correspond to the current limit of -10 Ato +10 A. The input voltage Vin is the reference set-point vrefi produced bythe voltage control loop while the output voltage Vout is the limited set-pointvoltage vseti for comparison with vfbi. The resistor used for Rlim is 470 Ω andthe two diodes are BAT85 Schottky diodes.
The same circuit was used to limit the output of the compensation amplierGc(s) to -5 V and +5 V so as to clamp the range of the modulator reference
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 50
signal. With the triangular waveform at a 5.5 V amplitude (11 V peak topeak) and a maximum reference voltage of +5 V, the maximum duty cycleis limited to 95 %. Similarly the minimum duty cycle is limited to 5 % at aminimum reference signal of -5 V.
3.2.5 Voltage Control Loop
The voltage control loop functions to regulate the dc bus voltage vdc to thecorresponding voltage set-point Vdc
∗. The bus plant portion of the controllerblock diagram shown in Fig. 3.10 contains dc bus capacitors of the convertercircuit. The dc bus voltage is related to the rest of the converter circuitthrough the current owing into the capacitors and the capacitors themselves.The current owing into the capacitor as shown Fig. 3.12 is related to theinductor current by
ic = idc − s1iL, (3.63)
where s1 is again used to determine the state of switch Q1 as either on (s1 = 1)or o (s1 = 0). The current through the capacitor is also dened as
ic = Cdvcdt
= Cdvdcdt
. (3.64)
Combining (3.63) and (3.64) results in the dierential equation describing thedc bus voltage as a function of the inductor current and is given below as
dvdcdt
=idc − s1iL
C. (3.65)
With s1iL averaged over at least 5 to 10 switching cycles, s1iL is approximatelyequal to d1iL, where d1 is the instantaneous duty cycle of switch Q1. Equation(3.65) time averaged over 5-10 switching cycles now becomes
dvdcdt
=idc − d1iL
C. (3.66)
Both the inductor current iL and the duty cycle d1 are time-varying signalsresulting in a non-linear dierential equation for the dc bus voltage. For fre-quency domain analysis transfer functions are required to be linear. Thus(3.66) is linearised before applying the Laplace transform for frequency do-main analysis. A function can be linearised around a certain operating pointusing a rst order Taylor series approximation [40]. The approximation for(3.66) is thus
vdc = f (d1(0), iL(0)) +∂f
∂d1
∣∣∣d1=d1(0)iL=iL(0)
(d1(t)− d1(0))
+∂f
∂iL
∣∣∣d1=d1(0)iL=iL(0)
(iL(t)− iL(0)) ,
(3.67)
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 51
and in increment form
d∆vdc(t)
dt=
∂f
∂d1
∣∣∣d1=d1(0)iL=iL(0)
∆d1(t) +∂f
∂iL
∣∣∣d1=d1(0)iL=iL(0)
∆iL(t), (3.68)
where f is the dierential equation dvdcdt
given in (3.66). The operating point ofthe duty cycle d1 is around 0.6857 and denoted as d1(0). The operating pointfor the inductor current iL(0) is calculated as
dvdcdt
= 0 =idc − d1(0)iL(0)
C. (3.69)
Rearranging the above equation:
iL(0) =idc
d1(0). (3.70)
The operating point for the inductor current thus depends on the amountof current owing into the dc bus. Since this value will vary, the frequencyand time domain analysis are conducted at the maximum and minimum meandc bus current variation as well as at a mean dc bus current of zero. Afterevaluating (3.68) the linearised dierential equation for the dc bus voltage is
d∆vdc(t)
dt= −iL(0)
C∆d1(t)−
d1(0)
C∆iL(t). (3.71)
Multiplying both sides with C and taking the Laplace transform of the aboveequation, (3.71) becomes
− sCVdc(s) = Kd0IL(s) +Ki0D1(s), (3.72)
and after simplifying
Vdc(s) =Kd0IL(s) +Ki0D1(s)
−sC= Fp(s) [Kd0IL(s) +Ki0D1(s)] , (3.73)
where Fp(s) is used to dene the dc bus plant portion of the converter.The block diagram of the linearised control system is shown in Fig. 3.21 whereFc(s) denotes the controller for the voltage control loop. The feedback voltageVfbv shown in Fig. 3.21 is a scaled version of the bus voltage, where Ks is thescaling factor. The block diagram in Fig. 3.21 can be manipulated by movingthe branch point at IL(s) to D1(s) as shown in Fig. 3.22. By performing blockdiagram reduction the control system diagram reduces to the diagram shownin Fig. 3.23.
From Fig. 3.23 the open-loop transfer function of the voltage control loopis calculated as
Fol(s) =KsKRFc(s)Gc(s)Fp(s)
1 +KdcKRGc(s)Gp(s)H(s)[Ki0 +KdcKd0Gp(s)] , (3.74)
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 52
Vdc∗ Σ
+
−Vfbv
Fc(s)
Vseti
Σ+
−Vfbi
Gc(s) KRD1
Kdc Gp(s)IL
Kd0 Σ++
Fp(s) Vdc(s)
H(s)
Ks
Ki0
Vε
Figure 3.21: Voltage and current control loop block diagram.
Vdc∗ Σ
+
−Vfbv
Fc(s)
Vseti
Σ+
−Vfbi
Gc(s) KR
D1Kdc Gp(s) Kd0 Σ
++
Fp(s) Vdc
KdcH(s)Gp(s)
Ks
Ki0
Vε
Figure 3.22: Simplied voltage and current control loop block diagram.
Vdc∗ Σ
+
−Vfbv
Fc(s)Gc(s)KR
1 +KdcKRGc(s)Gp(s)H(s)Ki0 +KdcKd0Gp(s) Fp(s) Vdc
Ks
Vε
Figure 3.23: Reduced voltage and current control loop block diagram.
and the closed-loop transfer function as
Fcl(s) =Vdc(s)
Vdc∗ =
1Ks
Fol(s)
1 + Fol(s). (3.75)
The measured dc bus voltage is scaled down through a dierential amplierfunctioning as an attenuator as shown in Fig. 3.24 with R1 equal to 238 kΩ andR2 equal to 6.5 kΩ. The equivalent scaling factor denoted K2 in the voltagecontrol loop is the ratio between R2 and R1 which is 1
35. Thus at an ideal
positive dc bus voltage of 175 V, the equivalent feedback voltage vfbv is 5 V.The reference voltage Vdc
∗ is therefore set to 5 V.A 2 kW load is connected to the dc bus rails. The transfer function Fp(s)
given in (3.73), describing the dc bus plant portion of the system, is modelled
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 53
VGND
R1R2
+Vdc
R1
−
+
LF353N
vfbv
R2
Figure 3.24: Bus voltage feedback circuit.
with and without the load present to conrm stability throughout. With theload Rdc connected the transfer function changes to
Fp(s) =−Rdc
sCRdc + 1, (3.76)
where the minimum value of the load is 60 Ω. The open-loop frequency re-sponse without the compensation amplier Fc(s) present is shown in Fig. 3.25for the loaded as well as unloaded dc bus condition. The initial dc bus currentis assumed zero, thus from (3.70) the operating point for the inductor currentKi0 is also zero.
−200
−150
−100
−50
0
50
Magnitude
(dB)
10−1 100 101 102 103 104 105−270
−225
−180
−135
−90
−45
0
Frequency (Hz)
Phase
(deg)
UnloadedLoaded
Figure 3.25: Bode plot of the voltage controller without compensation.
As seen from Fig. 3.25 when a 60 Ω load is connected the open-loop responsedoes not have innite gain at dc. A pole-zero compensator is once again chosen
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 54
as the compensation method since it will provide an additional pole to achievean innite dc gain and also provide the required additional bandwidth for thevoltage control loop.
The dierence between the set-point voltage Vdc∗ and the scaled bus voltage
Vfbv is Vε, which is the input to the voltage control loop's compensator, Fc(s).The derivation for the two corner frequencies and dc gain component for apole-zero compensator was shown in Section 3.2.3. The rst corner frequencyf1 is chosen to be at 5 Hz which is where the bode plot shown in Fig. 3.25rolls o at -20 dB/decade. The second corner frequency f2 is chosen to be ata frequency of around 10 kHz which will be well below the 0 dB line once thedc gain is added. The bandwidth of the voltage controller should ideally be ataround one tenth of the current control loop's, thus at 320 Hz.
The compensation amplier is identical to the circuit shown in Fig. 3.16where the two inputs are now Vsetv and Vfbv and the output is Vseti. The samedierential equations thus also apply. The component values were calculated ina similar manner to that of the current controller's compensation amplier byusing equations (3.54) through (3.59). The nal component values are given inTable 3.11 with the corner frequencies at 33.8 Hz and 15.95 kHz respectively.
Table 3.11: Component values for the voltage loop compensator.
Component ValueR1 1 kΩR2 100 kΩC1 100 pFC2 47 nF
The bode plots for both the open- and closed-loop frequency responses of thevoltage control loop are shown in Fig. 3.26 at a dc bus current of zero. A dcgain of 30.9 dB is indicated, corresponding to a static loop sensitivity of 35 Vof bus voltage per volt of excitation applied to the set-point input. Thus, ata set-point value of 5 V, the bus voltage will rise to 175 V. A phase-margin,gain-margin and unity gain crossover frequency of 77.6 degrees, 20.8 dB and294 Hz are further indicated, suggesting that the voltage control loop is stableunder the specied conditions. Since a load connected to the dc bus onlyaects the controller at very low frequencies (below 50 Hz) the gain and phasemargins remains the same under no-load conditions.
The stability of the voltage control loop was tested again at the meanminimum and maximum inductor current operating points of ±10 A. Theopen-loop frequency response results are summarised in Table 3.12.
The results show the voltage control loop will remain stable within theinductor current operating range with positive gain and phase margins for all
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 3. DC-DC CONVERTER DESIGN AND IMPLEMENTATION 55
−200
−150
−100
−50
0
50
100
Magnitude
(dB)
10−1 100 101 102 103 104 105−360
−270
−180
−90
0
Frequency (Hz)
Phase
(deg)
Open-loopClosed-loop
Figure 3.26: Bode plot of voltage controller under load.
Table 3.12: Stability analysis of the voltage control loop at various meaninductor current operating points.
Ki0 10 A 0 A -10 AGain margin 24.2 dB 20.8 dB 17.1 dBPhase margin 81.4° 77.6° 73.7°Unity cross-over 319 Hz 294 Hz 270 Hz
conditions. A complete circuit schematics of the controllers used for both thetop and bottom converter are shown in Appendix A.1.2.
3.3 Summary
The detailed design of the bidirectional dc-dc converter circuit was presentedin this chapter. An overview of the basic circuit operation was given. Thedesign of the dc-dc converter's double control loop control circuit was presentedand the implementation thereof detailed. Both frequency and time domainequations were derived to describe the operation of the circuit. The frequencydomain analysis showed both the inner and outer control loops are stable. Thetime domain analysis is presented in the next chapter along with the measuredresults from the manufactured prototype circuit for comparison purposes.
Stellenbosch University https://scholar.sun.ac.za
Chapter 4
DC-DC Converter Simulation and
Test Results
In this chapter the operation of the designed dc-dc converter from the previouschapter is simulated in the time domain. This simulation is used to conrmthe stability of the converter during start-up, as well as to examine the stepresponse of the controller with both current and voltage control loops imple-mented. Various measurements are taken from the manufactured prototypecircuit in order to validate the time domain simulations and to demonstratethe practical operation of the circuit.
Since the simulation of the negative bus voltage is identical to the invertedsimulation of the positive bus voltage, only the top-half of the converter circuit(producing the positive bus voltage +Vdc) was used for simulation purposes.The simulations were implemented in MATLAB and corresponds to a simpleiterative Euler solution of the relevant dierential equations, (3.37), (3.38),(3.49), (3.50) and (3.65) from Chapter 3, that describe the behaviour of thecircuit.
4.1 Test Set-Up
The dc-dc converter test set-up is shown in Fig. 4.1. The printed circuit boards(PCBs) used for the practical implementation of the converter are presented inAppendix B.1. The two power boards responsible for generating the positiveand negative dc bus rails are mounted on two heat sinks as shown in Fig. 4.1.The control board is situated on top of the two heat sinks. A small 12 V fanis mounted on the side of the heat sinks to increase the airow, thus coolingthe heat sinks.
The batteries are located at the bottom of the trolley shown in Fig. 4.1. Thetwo motors on-top of the batteries are used with the inverter circuit and areirrelevant to the measurements taken for this chapter. The circuit is switchedon by the two circuit breakers at the front of the trolley where each circuit
56
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 4. DC-DC CONVERTER SIMULATION AND TEST RESULTS 57
Figure 4.1: Converter prototype circuit test set-up.
breaker is connected to the positive and negative point of 10 series connectedbatteries.
The load used during testing is a 60 Ω resistive load comprising 6 seriesconnected 10 Ω resistors to form a resistor bank. Each resistor is rated tohandle 500 W and thus the six series connected resistors can handle a totalpower of 3 kW.
For safety purposes during measurements, the mid-point between the top-and bottom-half of the converter was used as the grounding point for all voltageprobe connections.
4.2 Start-Up Response
The simulated dc-dc converter start-up response showing both the instanta-neous dc bus voltage and inductor current are displayed in Fig. 4.2. Thissimulation includes the modelling of the soft-start circuit as well as the dead-time of the switches. A negative inductor current indicates current owingfrom the batteries into the dc bus to charge the dc bus or power a load, whilea positive inductor current indicates current owing from the dc bus into thebatteries thus charging the batteries.
As seen from Fig. 4.2 the rst stage of the soft-start circuit, the pre-chargecircuit, is active for the rst 350 ms after start-up, during which time thebus voltage rises from zero to the equivalent battery potential of 120 V. Thein-rush current iL charging the dc bus is seen to peak at 4.8 A.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 4. DC-DC CONVERTER SIMULATION AND TEST RESULTS 58
0 0.5 1 1.5 2 2.5 3 3.5 4 4.50
50
100
150
200
v dc(V
)
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
−4
−2
0
2
Time (s)
i L(A
)
Figure 4.2: Simulated dc bus voltage and inductor current during start-up.
At approximately 1.5 s the second stage of the soft-start circuit is activated,allowing the bus voltage to rise from 120 V to 175 V over the next 0.65 seconds.When the bus voltage reaches 175 V there is a small voltage overshoot of 2 V,which is within the allowed range of 170 V to 180 V. It is thus not necessaryto implement any additional anti-windup scheme [41] to prevent the dc busvoltage from overshooting after reaching its nominal value of 175 V.
The large initial ripple voltage on the dc bus is caused by the post-chargecircuit limiting the on-time of the switches by adjusting the dead-time. Theripple is seen to decrease until it reaches a minimum peak-to-peak value of27 mV once the dead-time has reached the minimum value. At time t = 4.5 sthe dead-time has decreased suciently to allow the controller to functionnormally thus regulating the dc bus voltage to 175 V continuously withoutany signicant voltage ripple. At the same time the inductor current is alsoshown to have reached the designed current ripple of approximately 2.25 Apeak-to-peak.
Fig. 4.3 shows the measured dc bus voltage and inductor current duringstart-up. The magnitude and rate of change of the bus voltage is similar to theresult of the time domain simulation shown in Fig. 4.2. Due to the samplingrate of the oscilloscope used to obtain the measurements (1000 samples overthe chosen time period), the inductor current displayed in Fig. 4.3 is undersampled and thus aliasing occurs. In the 5 s measurement taken for Fig. 4.3the inductor current is time averaged over a period of 5 ms, thus aliases of the
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 4. DC-DC CONVERTER SIMULATION AND TEST RESULTS 59
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5−200
−150
−100
−50
0
50
100
150
200
v dc(V
)
Positive RailNegative Rail
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
−4
−2
0
2
Time (s)
i L(A
)
Figure 4.3: Measured dc bus voltage and inductor current during start-up.
inductor current are displayed as opposed to the true instantaneous inductorcurrent. The inductor current ripple under normal operating conditions cantherefore not be seen as in the simulated model.
As predicted, the bus voltage is observed to rise to the equivalent batterypotential within 350 ms of connecting the battery to the converter at whichtime an audible click is heard as the relay short-circuits the soft-start resistors.The battery potential is approximately 125 V per rail section which is slightlyhigher than the simulated potential due to the initial state of charge of thepurchased batteries. Another measurement of the start-up current iL was takenover a 1 s time period to determine a more accurate peak start-up current of4.6 A as shown in Fig. 4.4. The small current spike of 0.6 A at 350 ms occurswhen the relay switch closes.
The bus voltage is subsequently observed to rise to and regulate at 177.2 Vand -177.8 V respectively in Fig. 4.3 over a period of approximately 0.6 s as thedead-time is progressively reduced. The rail-to-rail nominal bus voltage is thus355 V. The positive and negative dc bus voltages peak at 179 V and -180 Vrespectively, according to the oscilloscope measurements. This correspond to avoltage overshoot of approximately 2 V from the regulated voltage as predictedby the time domain simulation. The positive and negative bus voltage areobserved to regulate 2 to 3 V higher than the nominally required value of plus
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 4. DC-DC CONVERTER SIMULATION AND TEST RESULTS 60
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1−5−4−3−2−101
Time (s)
i L(A
)
Figure 4.4: Measured inductor current during start-up.
and minus 175 V due to a small inaccuracy in the set-point value Vdc∗. Since
it is still within the allowed operating range, no additional adjustments weremade. After another 2.5 s the ripple on the dc bus voltage is seen to havedecreased signicantly and the mean inductor current is shown to be zero.
4.2.1 Gating Signal Implementation
The voltage waveform generated by the push-pull oscillator used to power theoptocoupler (that communicates the gating signals between the controller andMOSFET gates) is shown in Fig. 4.5. The oscillation frequency is measuredas 311 kHz. The peak-to-peak voltage is approximately 24 V corresponding toa 1:1 transformer ratio when the circuit is powered from a single 12 V battery.
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
·10−5
−10
0
10
Time (s)
Voltage
(V)
Figure 4.5: Measured oscillator signal used in gate-drive circuitry.
Both the current and voltage control loops employ integrators in their com-pensation ampliers. This causes the initial modulator reference voltage (gen-
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 4. DC-DC CONVERTER SIMULATION AND TEST RESULTS 61
erating the PWM signals) during the start-up sequence to be at a maximum,since the dc bus voltage has not yet reached the nominal value. The dutycycle of the switches are limited to 95 %. With the dead-time included themaximum duty cycle is 92 % and the resulting gating signal generated by thecomparator for Q2 is shown in Fig. 4.6 in blue. The green line shows the signalat the gate of the MOSFET.
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
·10−5
0
2
4
6
8
10
12
Time (s)
GatingSignals(V
)
ControllerMOSFET Gate
Figure 4.6: Measured gating signal from the controller and MOSFET gate forQ2 during start-up.
A similar switching signal is seen at the gate of Q3 while switches Q1 and Q4
are o completely due to the added dead-time.
4.3 Step Response
To determine the transient response of the dc-dc converter a step change in theload was implemented in both the simulated and prototype circuit. Fig. 4.7shows the predicted response of both the dc bus voltage and inductor currentto a step change in the load. The application of a 2.1 kW load was simulatedby adjusting the mean dc bus current Idc from 0 A to -6 A in the time domainsimulation. The system is tested for a load capability slightly higher thanthe required 2 kW due to the availability of the 60 Ω dummy load that wasused during testing. At time t = 0.1 s the load step response is implementedresulting in a negative average inductor current of 9.32 A and a negative voltagespike of 1.4 V. After 300 ms the simulated load is disconnected and the meancurrent is seen to return to zero while the dc bus voltage peaks at 176.4 Vbefore settling back to 175 V.
A similar response is seen when a dc current source is instantly appliedto the dc bus at time t = 0.7 s by changing the mean dc current Idc from0 A to 5.8 A (corresponding to 2.03 kW of power delivered to the converter).
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 4. DC-DC CONVERTER SIMULATION AND TEST RESULTS 62
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2173
174
175
176
177
v dc(V
)
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2
−10
0
10
Time (s)
i L(A
)
Figure 4.7: Simulated step response of dc bus voltage and inductor current.
The voltage briey peaks at 176.3 V and the mean inductor current is shownto be 8.42 A. After 300 ms when the current source is removed, the dc busvoltage falls to 173.7 V before settling back to the nominal dc bus voltage.The inductor current is seen to return to a mean value of zero. Accordingto the time domain simulation the duration of the transient responses areapproximately 30 ms. As seen from Fig. 4.7 the simulated converter providesa good transient response by settling quickly and having a maximum voltageovershoot of less than 1 % of the nominal bus voltage. This simulation alsodemonstrates the bidirectional capability of the converter.
The transient response of the prototype circuit was tested by connectingand disconnecting a 2.1 kW (60 Ω) load to the 355 V nominal dc bus. Themeasurement was taken at the positive dc bus voltage rail by using the ac cou-pled setting on the oscilloscope. As shown in Fig. 4.8 the load was connectedat time t = 0.05 s which caused a brief voltage dip of 1.32 V. The bus voltagesettled back to 177.2 V after 0.5 s at which time the load was disconnectedcausing a voltage spike with an amplitude of 0.92 V. Another 400 ms later thedc bus voltage settled back to the nominal voltage.
The transient response of the simulated and measured step responses com-pare reasonable well. The main dierence being the simulated circuit shows a
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 4. DC-DC CONVERTER SIMULATION AND TEST RESULTS 63
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1−1.5
−1
−0.5
0
0.5
1
Time (s)
ACcoupledv d
c(V
)
Figure 4.8: Measured step response of the dc bus voltage under ac coupling.
faster settling time than the measured step response.The test showing the bidirectional capability of the dc-dc converter cir-
cuit will be presented in Chapter 7 as part of the system integration testingsince the asynchronous generator is required to source current to the con-verter. The design and implementation of the inverter circuit connecting tothe asynchronous generator will be discussed in the following chapters.
4.3.1 Inductor Ripple Current
The simulated inductor current ripple is shown in Fig. 4.9 and the measuredinductor current at no-load is shown in Fig. 4.10. The simulation predicts acurrent ripple magnitude of 2.25 A peak, which is 22.5 % of absolute maximummean current that the inductor is intended to conduct at full power (sink orsource). The magnitude of the current ripple is independent of the meancurrent, provided that the bus voltage, battery voltage and inductance of theinductor remain constant. The inductor current duty cycle from Fig. 4.9 isshown to be 68.8 % which is as expected.
The measured inductor current at no-load has a peak-to-peak value of2.06 A which is slightly less than the predicted value mainly due to the higherbattery potential. The leakage inductance in the cables connecting the batter-ies to the rest of the converter circuit, as well as the leakage inductance in thepath through the series connected batteries themselves will add to the totalinductance causing the current ripple to decrease slightly.
The measured inductor current when a 2.1 kW load is connected acrossthe bus rails, is shown in Fig. 4.11. The mean inductor current is -9.59 Acorresponding closely to the predicted value of -9.32 A, where the negativesign indicates the batteries are sourcing instead of sinking current.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 4. DC-DC CONVERTER SIMULATION AND TEST RESULTS 64
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
·10−4
−1.5
−1
−0.5
0
0.5
1
1.5
Time (s)
i L(A
)
Figure 4.9: Simulated inductor current.
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
·10−4
−1.5
−1
−0.5
0
0.5
1
1.5
Time (s)
i L(A
)
Figure 4.10: Measured inductor current without a load connected.
In the converter simulation the peak-to-peak value of the current ripple staysconstant at no-load and full-load conditions since the value of the inductorand battery voltage are assumed constant. From Fig. 4.11 it can be seenthat the measured peak-to-peak current ripple increased to 3.1 A at full-load,compared to the 2.06 A at zero load. This is mainly due to the drop in thetotal battery potential from 247.2 V to 236.6 V when the load is connected.Also there is a marginal decline in the inductance of the inductor, due to thedecreased instantaneous permeability of the inductor core material at greatermagnetisation.
In both cases, the measured current is of triangular form with a duty cycleof approximately 69 % at no-load and a duty cycle of 66 % at full-load, duringwhich time switches Q1 and Q4 are turned on and the inductor current isobserved to increase linearly.
Overall, the time domain simulations and measurements for the dc-dc con-verter correspond very well. The simulation can thus be considered a useful
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 4. DC-DC CONVERTER SIMULATION AND TEST RESULTS 65
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
·10−4
−12
−11
−10
−9
−8
Time (s)
i L(A
)
Figure 4.11: Measured inductor current with a 2.1 kW load connected to thedc bus rails.
tool for predicting the behaviour of the prototype circuit.
4.3.2 Gating Signals
The measured controller gating signals for switches Q1 and Q2 under no-loadare presented in Fig. 4.12. The switches are seen to never be on simultaneouslyindicating the dead-time of 1 µs was implemented correctly. The duty cycleof switch Q1 is 69.2 % and 22.7 % for Q2.
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
·10−4
0
1
2
3
4
5
Time (s)
GatingSignals(V
)
Q1
Q2
Figure 4.12: Measured controller gating signals for Q1 and Q2 under steady-state no-load conditions.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 4. DC-DC CONVERTER SIMULATION AND TEST RESULTS 66
4.4 Eciency
A 60 Ω resistor bank was employed to load the converter to approximately2.1 kW. The converter eciency is determined by comparing the power deliv-ered to the load to that drawn from the battery. The measured instantaneousinductor current and ac coupled battery voltage are shown in Fig. 4.13. Themean battery voltage was measured using a multimeter as 236.6 V. The instan-taneous input power is also shown in Fig. 4.13 and was obtained by multiplyingthe magnitude of the inductor current with the battery voltage (after addingthe dc oset of 236.6 V), using MATLAB.
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
·10−4
−12
−11
−10
−9
−8
i L(A
)
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
·10−4
−4
−2
0
2
4
v bat(V
)ac
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
·10−4
1.82
2.22.42.62.8
·103
Time (s)
p in(W
)
Figure 4.13: Measured inductor current and ac coupled battery voltage underload.
From Fig. 4.13 the measured ac coupled battery voltage is seen to be in phasewith the measured inductor current. Since there is no phase shift between thebattery voltage and inductor current, the average inductor and battery voltagecan be used to calculate the average input power. With the average inductor
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 4. DC-DC CONVERTER SIMULATION AND TEST RESULTS 67
current measured with a LEM current probe as 9.592 A, the average inputpower is thus
Pin = VbatIL = 2269 W. (4.1)
The mean value of the instantaneous input power pin shown in Fig. 4.13 wascalculated in MATLAB as 2268.6 W conrming the average power calculationof (4.1).
The regulated dc bus voltage was measured with another multimeter as355 V rail-to-rail and the dc bus current owing into the load was measuredwith the LEM current probe as 6.01 A. The output power at the load is thus
Pout = VdcIdc = 2133 W, (4.2)
resulting in a converter eciency of
η =Pout
Pin
= 94.02%. (4.3)
The total power loss is calculated as
Ploss = Pin − Pout = 136 W. (4.4)
The total switching and conduction losses were calculated as 14.33 W in (3.28)for one half-bridge converter, thus 28.66 W for the complete converter circuit.The current sense resistors of 68 mΩ is responsible for a power loss of 6.3 Wat a mean inductor current of 9.6 A. The remaining power loss of a 101 W ismainly due to the battery's internal resistance, the inductor's equivalent seriesresistance and the resistance within the (long) cables connecting the batteriesto the dc-dc converter circuit.
4.5 Summary
The time domain simulation of the dc-dc converter circuit was presented in thischapter along with the measured test results. The test set-up was describedin detail. The simulated start-up response of the dc-dc converter comparedextremely well to the measured start-up response. The transient response ofthe prototype circuit was tested by implementing a step change in the loadconnected to the converter's dc bus terminals. The simulated and measuredtransient responses compared reasonable well. The dc-dc converter circuitdemonstrated its ability to regulate the dc bus voltage under no-load andloaded conditions, whilst meeting the design specications. The prototypecircuit showed a high eciency and overall satisfactory performance.
Stellenbosch University https://scholar.sun.ac.za
Chapter 5
Asynchronous Machine Modelling
An asynchronous machine is used as the generator that connects to the invertercircuit functioning in rectier mode. A model of the asynchronous machine isrequired in order to design the controller for the inverter and generator. Thischapter derives all the necessary machine parameters for a computer simulatedmachine model. The machine model is demonstrated and compared with thetrue machine characteristics.
5.1 Machine Parameter Estimation
The asynchronous machine used is a 3 kW three-phase squirrel-cage rotorinduction motor. The motor will however be operated as a generator. Theequivalent circuit for an asynchronous machine (also known as an inductionmachine) was described in detail in the literature review Section 2.4.2. Theequivalent per-phase circuit diagram is repeated here as shown in Fig. 5.1 forconvenience.
−Vs
+
Rs
Is
jXls jXlr
Ir
Rr
Rr
s(1− s)jXmRc
Figure 5.1: Per-phase equivalent circuit of an induction machine.
The equivalent circuit parameters can be obtained by measurements, no-loadtesting and a locked-rotor test. The datasheet provides most of the no-load andlocked-rotor measurements required for the parameter estimation calculations.Table 5.1 provides a summary of the machine characteristics, as given in thedatasheet.
68
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 5. ASYNCHRONOUS MACHINE MODELLING 69
Table 5.1: Three phase induction machine characteristic values.
Parameter Rated ValueShaft power 3 kWFrequency 50 HzPoles 2Slip 4.33 %Voltage 230/400 V (ac rms)Current 10.6/6.07 ALocked rotor current 70.7/40.7 ANo-load current 4.87/2.8 AFull load torque 9.99 NmLocked rotor torque 230 %Breakdown torque 280 %
Due to the available bus voltage of ±175 V provided by the dc-dc converter,the induction machine will be connected in delta (low-voltage conguration).The machine slip is denoted by a small letter s for the duration of this chapter.
5.1.1 Stator Resistance
The stator resistance Rs shown in Fig. 5.1 was measured with a multimeteras 2.1 Ω by measuring the resistance between one of the stator terminals andneutral, while the machine was connected in wye conguration.
5.1.2 Leakage Reactance
The leakage reactances Xls and Xlr shown in Fig. 5.1 are usually calculatedas one entity Xe and approximated by
Xls = 0.4 Xe, (5.1)
and
Xlr = 0.6 Xe, (5.2)
for a class B motor [26] such as the motor used for this project. An approximatemodel of the per-phase equivalent circuit [27] of an induction machine is usedto determine the leakage reactances as shown in Fig. 5.2. The approximation isvalid since the core resistance Rc and impedance of the magnetising inductanceXm are much larger than the stator leakage impedance Rs + jXls.
The leakage reactances are calculated from the expression for breakdowntorque which is derived below. First consider the expression for the real powerdeveloped at the shaft of the machine as given in (2.32) and repeated here:
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 5. ASYNCHRONOUS MACHINE MODELLING 70
−Vs
+ Is
Rs jXls jXlr
Ir
Rr
Rr
s(1− s)jXmRc
Figure 5.2: Approximate per-phase equivalent circuit of an induction machine.
Pd = 3Ir2Rr
(1− s
s
). (5.3)
The rotor current Ir is calculated as
Ir =Vs
Re + jXe +Rr(1−s)
s
, (5.4)
where
Re = Rs +Rr. (5.5)
Substituting (5.4) into (5.3) results in
Pd =3V 2
s Rr(1−s)
s
R2e +X2
e +[Rr
(1−s)s
]2+ 2ReRr
(1−s)s
. (5.6)
The torque is related to power by the synchronous angular velocity ωs as givenin (2.33) which results in a developed shaft torque of
Td =1
ωs
3V 2s
Rr
s
R2e +X2
e +[Rr
(1−s)s
]2+ 2ReRr
(1−s)s
. (5.7)
As seen from (5.7) the maximum torque, also known as the breakdown torque,occurs when 1−s
sis at a minimum. The breakdown slip sb is obtained by
dierentiating (5.7) with respect to s and setting it equal to zero which resultsin
sb =Rr√
R2s +X2
e
. (5.8)
The expression for the breakdown torque Tdm is calculated by substituting(5.8) into (5.7) resulting in
Tdm =3V 2
s
2ωs
[1
Rs +√R2
s +X2e
]. (5.9)
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 5. ASYNCHRONOUS MACHINE MODELLING 71
From (5.9), the measured stator resistance and the values given in Table 5.1,the combined leakage reactance Xe is calculated as 6.60 Ω. The individualleakage reactances are calculated from (5.1) and (5.2) as Xls = 2.64 Ω andXlr = 3.96 Ω respectively.
5.1.3 Rotor Resistance
The rotor resistance Rr (referred from the rotor to the stator) is calculatedfrom the developed power expression given in (5.6) as 1.96 Ω.
5.1.4 Magnetising Reactance
The magnetising reactance Xm is calculated from the no-load test. At no-loadthe slip is almost zero and hence the impedance of the rotor is almost innite.Fig. 5.2 thus reduces to Fig. 5.3.
−Vs
+ IsjXmRc
Figure 5.3: Reduced equivalent circuit of an induction machine under no-load.
The magnetising reactance is given by
Xm =V 2nl
Qnl
, (5.10)
where
Qnl = VnlInlsin(θpf ). (5.11)
The no-load current is given in the datasheet as 2.8 A and the no-load voltageis the same as the rated voltage, 230 V. The power factor at no-load is givenin the datasheet as 0.15. The power factor angle at no-load is thus
θpf = cos−1(0.15) = 81.37 (5.12)
Substituting (5.11) and (5.12) into (5.10) results in
Xm =Vnl
Inlsin(θpf )= 83.08 Ω. (5.13)
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 5. ASYNCHRONOUS MACHINE MODELLING 72
5.1.5 Core Resistance
The core resistance Rc is calculated in a similar manner as the magnetisinginductance. With reference to Fig. 5.3 the core resistance is
Rc =V 2nl
Pnl
, (5.14)
where
Pnl = VnlInlcos(θpf ). (5.15)
The resulting core resistance Rc is calculated as 547.6 Ω. The calculatedmachine parameters are summarised in Table 5.2
Table 5.2: Calculated machine parameters.
Parameter ValueRs 2.1 ΩXls 2.64 ΩXlr 3.96 ΩRr 1.96 ΩXm 83.08 ΩRc 547.6 Ω
5.2 Machine Characteristics
The characteristic curves related to slip are given in the machine datasheet andare used to determine the accuracy of the machine model. Since the machineis rated to operate at 50 Hz all the machine data given in the datasheet is at astator frequency of 50 Hz. The voltage applied to stator phases are consideredto have a constant rms value of 230 V.
5.2.1 Torque-Speed Relationship
The modelled and measured datasheet torque-speed curves are shown in Fig. 5.4.From the datasheet curve, the rotor speed is given as a percentage of the ratedspeed (50 Hz, 3000 rpm), thus the developed torque is given for slip valuesranging from 0 to 1. If the rotor speed is equal to the rated stator speed thenthe slip is 0, and if the rotor speed is zero the slip is 1. The modelled torquecurve was produced by simulating the developed torque as given in (5.7) forslip values also ranging from 0 to 1.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 5. ASYNCHRONOUS MACHINE MODELLING 73
0 10 20 30 40 50 60 70 80 90 1000
0.5
1
1.5
2
2.5
3
Speed relative to rated speed (%)
Torquerelative
toratedtorque
ModelDatasheet
Figure 5.4: Torque vs rated speed characteristic curve.
To achieve the modelled curve of Fig. 5.4 the rotor resistance was changedfrom the calculated value of 1.96 Ω to 1.86 Ω, and the magnetising inductancewas slightly reduced to 80 Ω. As seen from Fig. 5.4 the modelled curve onlyresembles the curve given in the datasheet from 0 % slip to 25 % slip. Theequivalent model is thus not able to predict an accurate torque response forslip values greater than 25 %, at which point the breakdown torque is reached.Since the induction machine will never operate continuously at slip valuesgreater than a few percent, the model need only be accurate for slip valuesranging from 0 % to approximately 6 %. The rated machine slip is 4.33 % andthus the machine will not operate at more than one or two percent slip abovethe rated slip continuously.
The torque-speed curve shown is for motor operation where power is de-livered to the shaft. Since the induction machine will be used as a generator,the machine will be modelled using negative slip values.
5.2.2 Current-Speed Relationship
Similarly the stator current related to the rated speed is shown in Fig. 5.5for both the modelled and measured (from the datasheet) case. The modelledstator current was calculated from Fig. 5.1 as
Is =
∣∣∣∣ Vs
Zeq
∣∣∣∣ , (5.16)
where
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 5. ASYNCHRONOUS MACHINE MODELLING 74
Zeq = Rs + jXls +Rc||jXm||(jXlr +
Rr
s
)= Rs + jXls +
RcjXm
(jXlr +
Rr
s
)RcjXm + (Rc + jXm)
(jXlr +
Rr
s
) . (5.17)
0 10 20 30 40 50 60 70 80 90 1000
1
2
3
4
5
6
7
Speed relative to rated speed (%)
Current
relative
toratedcurrent
ModelDatasheet
Figure 5.5: Stator current vs rated speed characteristic curve.
The modelled current is seen to deviate signicantly more from the measuredcurrent for slip values greater than 6 %. For small slip values the model doeshowever correspond to the current given in the datasheet curve.
Since the measured current and torque curves deviate from the modelledcurves, it is clear that the values of the machine parameters also change atdierent slip values. The locked-rotor test for instance is only performed for afew seconds during which time the machine heats up signicantly. A change intemperature is known to change the eective resistance. The rotor resistancemay vary 50 % up to a 100 % from the nominal value at rated slip due toheating [42]. The magnetising inductance will vary due to ux saturationat high slip values. The leakage reactances change as the stator and rotorcurrents change due to the non-linear behaviour of magnetic materials at highereld-intensity values (which is directly related to current) [43].
The equivalent circuit parameters determined in this chapter are thus ad-equate to predict the behaviour of the chosen induction machine at very lowslip values (less than 6 % slip).
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 5. ASYNCHRONOUS MACHINE MODELLING 75
5.3 Dynamic Machine Model
The dynamic d-q model of an asynchronous (or induction) machine was al-ready discussed in Section 2.4.3. This model, consisting of equations (2.38) to(2.47), is used to model the dynamic behaviour of the 3 kW induction machinedescribed in this chapter at low slip values. The nal machine parameters usedfor modelling are summarised in Table 5.3. The core resistance Rc is not takeninto account for this model.
Table 5.3: Machine parameters used for modelling.
Parameter ValueRs 2.1 ΩXls 2.64 ΩXlr 3.96 ΩRr 1.86 ΩXm 80 Ω
Negative slip values are used to simulate generator operation. The machinemodel was programmed in a MATLAB script where the dierential equationswere solved using the iterative Euler method.
5.3.1 Modelling Results
The simulation was tested at a slip of zero to conrm the behaviour of thedynamic machine model. At a slip of zero the no-load current is given as 2.8 Ain the datasheet and thus a similar value is expected for the model. Threesinusoidal 50 Hz, 230 V rms voltages phase shifted by 120° from each otherwere applied to the dynamic model and the simulation results are shown inFig. 5.6.
The stator voltages were programmed to increase linearly over a time periodof 0.16 s before reaching a nominal value of 230 V rms. Although this didreduce current surges during start-up, to avoid any overcurrent conditionsaltogether, the stator voltages have to increase linearly over a time period ofapproximately 2 s. Once the current settled it reached a steady-state valueof 2.78 A rms which is close to the true no-load current of 2.8 A. Understeady-state no-load conditions the model showed a developed torque of zeroas expected.
Given that the generator will operate at a slip value of around -5 %, anothersimulation was conducted to conrm the generator operation of the dynamicmodel. The resulting stator and rotor currents for one of the three phasesare shown in Fig. 5.7. The rotor shaft was modelled to rotate at a constantangular velocity of 50 Hz.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 5. ASYNCHRONOUS MACHINE MODELLING 76
0 2 4 6 8 10 12 14 16 18 20
·10−2
−400
−200
0
200
400
v s(V
)
0 2 4 6 8 10 12 14 16 18 20
·10−2
−10
−5
0
5
10
Time (s)
i s(A
)
Figure 5.6: Three-phase stator voltages and currents at zero slip.
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
−10
0
10
Time (s)
Current
(A)
isir
Figure 5.7: Stator and rotor currents at -5 % slip.
The stator current is shown to have a frequency of 47.62 Hz while the ro-tor current has a frequency of 2.38 Hz corresponding to a slip of 5 %. Thedeveloped torque of the machine is -15.27 Nm according to the simulation,indicating that 15.27 Nm of torque must be applied to the shaft in order toachieve -5 % slip operation.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 5. ASYNCHRONOUS MACHINE MODELLING 77
5.4 Summary
This chapter derived the asynchronous generator machine parameters fromthe equivalent machine model and datasheet information. The machine modelwas simulated and the results compared to the given datasheet information toconrm the credibility of the model. The model was shown to be reasonablyaccurate for low slip values. The generator operation of the machine was alsoinvestigated. The next chapter contains the design of the inverter controlcircuit which is based on the machine parameters and performance detailed inthis chapter.
Stellenbosch University https://scholar.sun.ac.za
Chapter 6
Inverter Design and
Implementation
The design and implementation of a non-isolated bidirectional current-controlledswitched-mode inverter functioning as a synchronous rectier is described inthis chapter. The inverter connects to the dc bus provided by the dc-dc con-verter presented in Chapters 3 and 4. The asynchronous generator describedin Chapter 5 is used to supply power to the dc bus through the inverter circuit.
The generator is required to supply a maximum of 2 kW of electrical powerto a load connected between the dc bus rails, or if a load is not present, tothe dc-dc converter in order to charge the converter's batteries. The voltagesapplied to the generator stator terminals are controlled as a function of thedc bus current and the dc-dc converter's battery voltage. Two control loopsare used to facilitate the regulation of the battery voltage. The design ofthe control loops are presented along with their closed-loop stability analysis.The design and implementation of various additional safety features are alsodescribed.
6.1 Basic Circuit Operation
The inverter circuit is used to regulate the dc-dc converter battery voltage toensure the batteries are kept at their optimal oating use voltage of 13.6 V each.The positive and negative dc-dc converter bus rails, +Vdc and −Vdc, connectto the inverter terminals. The complete circuit diagram of the inverter circuitconnected to the already designed dc-dc converter circuit is shown in Fig. 6.1.
The inverter circuit is based on the commonly used three-phase half-bridgetopology, as shown in Fig. 6.1, with freewheeling diodes connected across theswitches. The current is assumed to ow into the dc bus thus indicatingthe three-phase machine is operating as a generator and the inverter as asynchronous rectier.
The dc bus terminals of the inverter and converter are connected together
78
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 79
Rin
Vbat
LiL vm
Q2
Q1
C
+Vdc
Rin
Vbat
LiLvm
Q3
Q4
C
−Vdc
Load
Io
idc
Ci
S1
is1
id1
S2
vu
is2
id2
S3
is3
id3
S4
vviv
ib vb
is4
id4
S5
is5
id5
S6
vwiw
iu
is6
id6
ic
vciava
3-PhaseMachine
Figure 6.1: Bidirectional converter and inverter interface circuit.
by means of multi-strand wires. The wires are about 1 m long. A load can alsobe connected to the dc bus if desired. A 100 µF 400 V electrolytic capacitor isconnected in parallel with a 100 nF 400 V ceramic capacitor and the combinedcapacitance is denoted Ci in Fig 6.1. The capacitors are connected betweenthe dc bus rails on the inverter side to lter any possible high-frequency noisedue to the inductance of the wire connection.
The dc-dc converter regulates the dc bus voltage to 355 V rail-to-rail. Thepositive denoted dc bus current idc is regulated and limited to a maximummean value of 5.7 A by the inverter's control circuit. This permits a powertransfer of up to 2023 W from the generator to the nominal 355 V dc bus,thus meeting the power requirement of 2 kW. Since the speed of the generatorshaft varies, as would be the case with an air-driven turbine generator, themaximum electrical power that can be drawn from the generator is dictatedby the mechanical power applied to the generator shaft at a given shaft speed.The current in the generator windings should never exceed the rated generatorcurrent, regardless of the generator shaft speed.
The magnitude and frequency of the fundamental voltages (vuf , vvf andvwf ) applied to the generator stator terminals are controlled to maintain aconstant Volt/Hertz ratio. Note all subscripts ending with an f indicate thefundamental component of said waveform. A constant Volt/Hertz ratio corre-sponds to a constant rms ux within the machine, since ux is proportional tovoltage integrated with respect to time. If constant ux is not maintained, themagnetic core of the machine will saturate causing large saturation currents
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 80
to ow which can damage the machine.The voltages applied to the generator stator terminals, vu, vv and vw, are
produced by the inverter circuit. An IGBT module containing six IGBTswith freewheeling diodes included, are used for switches S1 through S6. Thegeneration of the PWM signals controlling the six switches are shown in Fig. 6.2in block diagram form.
MPSignal
Processing
ωr
ControllerSignal
vcs
dsPIC DACvref
PWMSignal
vr
20 kHz
Figure 6.2: PWM signal generation block diagram.
The control circuit of the inverter produces three sinusoidal modulator refer-ence voltages, denoted vref in Fig. 6.2, each separated 120 degrees from eachother with equal magnitude. The control circuit includes a dsPIC digital signalcontroller and a digital-to-analog converter (DAC) to generate these referencevoltages as shown in Fig. 6.2.
The speed at which the generator shaft is rotating ωr is measured using amagnetic pick-up. The output signal from the magnetic pick-up, denoted MPin Fig. 6.2, is processed and given as an input to the dsPIC. The generatoris designed to function at a constant negative slip of -5.47 % and thus thefrequencies of the modulator reference voltages are approximately 5 % lowerthan the measured shaft speed.
The amplitudes of the modulator reference voltages are dependent uponthe analog control signal vcs, feeding into the dsPIC as shown in Fig. 6.2, aswell as the speed of the generator shaft ωr. The control signal vcs is generatedby the compensation amplier located inside the inverter's inner control loopand is described in more detail in the next section. The amplitudes of thereference voltages are scaled according to the magnitude of vcs and the speedinput ωr in order to maintain constant rms ux within the generator.
The modulator reference voltages are compared to a triangular carrierwaveform vr with a chosen frequency fs of 20 kHz to generate the PWMswitching signals as shown in Fig. 6.2. The top and bottom switches of each ofthe three half-bridge legs switch complementary to each other. With referenceto Fig. 6.1, when one of the top switches S1, S3 or S5 is switched on, the re-spective pole voltage is pulled high to +Vdc. Similarly, when one of the bottomswitches S2, S4 or S6 is switched on, the respective pole voltage is pulled lowto -Vdc.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 81
0 12π π 3
2π 2π
−5
0
5
v ru(V
),v r
(V)
0 12π π 3
2π 2π
−200
−100
0
100
200
v u(V
),v u
f(V
)
0 12π π 3
2π 2π
−10
0
10
ωs1t (rad)
i uf(A
)
Figure 6.3: Inverter phase voltage and current for one half-bridge circuit.
Figure 6.3 illustrates the operation of the inverter circuit for the half-bridgewith the pole voltage vu. It should be noted that the triangular carrier wave-form vr shown in green in Fig. 6.3 has a frequency 2 kHz instead of the designed20 kHz frequency simply to allow a more detailed viewing of the circuit oper-ation in which the switching action can clearly be seen. The rst plot showsthe corresponding modulator reference voltage vru generated by the controlcircuitry and is dened as
vru = Vc cos(ωs1t), (6.1)
where Vc is the dsPIC scaled amplitude of the modulator reference voltageand ωs1 is the angular frequency of the modulator reference voltage. The timeperiod for one complete angular rotation of 2π is dened as T1. The angularfrequency ws1, and thus also the time period T1, will vary due to the variablespeed of the generator shaft. The reference voltage vru is thus used to modulatethe phase voltage vu and the amplitude modulation index ma is given by [11]
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 82
ma =Vc
Vr
, (6.2)
where Vr is the amplitude of the triangular carrier waveform vr. The carrierwaveform is generated in an identical manner to the carrier waveform used inthe converter control circuit detailed in Section 3.2.1. The only dierence isa 2 MHz crystal (instead of a 4 MHz crystal) is used to produce the 20 kHztriangular switching signal. The peak-to-peak output voltage of the triangularwaveform vr is 12 V, thus Vr equals 6 V. The amplitude Vc of each of the mod-ulator reference voltages is limited to 5 V resulting in a maximum modulationindex of
ma =Vc
Vr
=5
6= 0.833. (6.3)
The second plot in Fig. 6.3 shows the modulated phase voltage vu in blueand the fundamental component vuf of the voltage in red. No lter is addedbetween the inverter pole points and the stator terminals; hence the modulatedphase voltages are applied directly to the generator stator terminals. Theinduction machine does not require ltered voltages at its stator terminalssince the internal machine inductance is very high and thus the machine actsas a lter in itself. This property is described in more detail in the nextsubsection.
In Fig. 6.3 the fundamental phase voltage vuf is shown to be in phasewith the modulator reference signal vru. The amplitude Vp of the fundamentalwaveform vuf is thus related to the modulation index by
Vp = ma
Vdc(pp)
2, (6.4)
where Vdc(pp) is the peak-to-peak dc bus voltage of 355 V. The waveform vufis thus given by
vuf = Vp cos(ωs1t). (6.5)
The third plot in Fig. 6.3 shows the fundamental component iuf of the phasecurrent iu to demonstrate the phase shift between the fundamental phase volt-age vuf and current iuf . This phase shift is due primarily to the inductanceof the generator and is calculated from the impedance of the generator. Withthe generator parameters determined in the previous chapter, the impedanceZeq between two stator terminals is calculated using (5.17) as
Zeq = 31.46 6 143.62 Ω, (6.6)
at an angular frequency of 50 Hz. The fundamental phase voltage vuf isused as the reference waveform. With reference to (6.6) and with the phasecurrents dened as positive when owing out of generator and into the dc
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 83
bus, the current is calculated as leading the reference voltage by φ = 36.38°.The amplitude of the fundamental phase current is dened as Ip, hence thefundamental phase current iuf is given by
iuf = Ip cos(ωs1t+ φ). (6.7)
The corresponding currents through switches S1 and S2 are shown in Fig. 6.4with reference to Fig. 6.1 and Fig. 6.3, where current is dened as positive whenowing from the generator into the dc bus. During the positive half-cycle ofiuf shown in Fig. 6.3 and while switch S2 is o, current will ow from thegenerator through the forward biased diode of S1 towards the dc bus. Whenswitch S2 is on current will ow through the IGBT of S2 as shown in Fig 6.4.
0 12π π 3
2π 2π
0
5
10
15
i d1(A
)
0 12π π 3
2π 2π
0
5
10
15
i s2(A
)
0 12π π 3
2π 2π
0
5
10
15
i d2(A
)
0 12π π 3
2π 2π
0
5
10
15
ωs1t (rad)
i s1(A
)
Figure 6.4: Switch and diode currents for S1 and S2 over a single angularrotation.
During the negative half-cycle of iuf current will ow from the dc bus towardsthe generator through the IGBT of S1 as long as switch S1 is on. However,
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 84
when switch S1 is o and iuf is still within the negative half-cycle, current willow through the forward biased diode of S2.
The time-varying duty cycle of switch S1 is related to the modulation indexgiven in (6.2) and the modulator reference voltage vru dened in (6.1) suchthat [31]
d1 =1
2[1 +ma cos(ωs1t)] . (6.8)
The duty cycle d1 is thus out of phase with iuf as seen from Fig. 6.4. Theinstantaneous positive current owing into the dc bus, as shown in Fig. 6.1, isa function of the current through the diodes and IGBTs of either the top orbottom three switches where
idc = (id1 + id3 + id5)− (is1 + is3 + is5)
= −(id2 + id4 + id6) + (is2 + is4 + is6).(6.9)
0 13π 2
3π π
0
5
10
15
ωs1t (rad)
i dc(A
)
Figure 6.5: Instantaneous and average dc bus current.
Figure. 6.5 shows the instantaneous dc bus current in blue and the average dcbus current Idc in red. The dc bus voltage is regulated to a constant 355 V andthe dc bus capacitors labelled C in Fig. 6.1 remain charged, hence the currentowing into the capacitors (at equilibrium) has a mean value of 0 A (onlythe ripple component of the instantaneous dc bus current will ow into thecapacitors). At the maximum average dc bus current of 5.7 A, the rms valueof the ripple component was calculated in MATLAB as 5.07 A. The maximumcombined rms current the capacitance C is rated to is 9.7 A as stated in the dcbus capacitance design of Section 3.1.2. The maximum combined rms ripplecomponent of the dc bus current from both the converter and inverter circuit,at equilibrium, equals 9.61 A and is thus within the allowed operating rangeof the capacitors.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 85
6.1.1 Generator Voltage and Current
The inverter phase voltages vu, vv and vw, shown in Fig. 6.1, are applied tothe respective generator stator terminals. The induction generator is howeverconnected in delta conguration, thus the line-to-line voltages applied acrosseach of the three generator phases, as shown in Fig. 6.1, are given by
va = vu − vv, (6.10)
vb = vv − vw, (6.11)
and
vc = vw − vu. (6.12)
0 12π π 3
2π 2π
−200
−100
0
100
200
v u(V
)
0 12π π 3
2π 2π
−200
−100
0
100
200
v v(V
)
0 12π π 3
2π 2π
−400
−200
0
200
400
ωs1t (rad)
v a(V
)
Figure 6.6: Instantaneous and fundamental line and phase voltages.
The two inverter phase voltages vu and vv are shown in Fig. 6.6 in blue, alongwith the resulting line voltage va. The fundamental component of each wave-form is also indicated in green. The applied line voltages cause induced linecurrents to ow and are calculated using
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 86
ILL =VLL
Zeq
, (6.13)
where ILL and VLL denote the respective phasor line currents and voltagesof the generator as shown in Fig. 6.1. The equivalent generator impedanceZeq is given in (6.6). The amplitudes of the line currents, and thus also theamplitudes of the inverter phase currents, will remain fairly constant with achange in angular frequency since both voltage and impedance change withrespect to frequency. However, at low angular frequencies the relationshipdoes not hold and current is seen to decrease with frequency and voltage.
The generator inductance as seen at the stator terminals is approximatedas the sum of the two leakage inductances, Lls and Llr. The approximation isvalid since the magnetising inductance Lm, connected in parallel with the rotorleakage inductance Llr, is 20 times larger than the rotor leakage inductanceand will thus have a negligible eect on the total generator inductance as seenfrom the stator terminals. The approximate generator inductance Lg is thuscalculated from the machine parameters, derived in Chapter 5, as
Lg =Xls +Xls
ωs0
=6.6
2π50= 21 mH, (6.14)
where ωs0 is the rated angular generator velocity. The inductance is seen tobe very large. Since current is equal to the time integral of the applied volt-age divided by the inductance, the generator will produce a sinusoidal phasecurrent with a small ripple component as shown in Fig 6.7. From simulation,the maximum current ripple is observed to occur when the duty cycle is 50 %,and has a peak-to-peak value of 0.155 A at a switching frequency of 20 kHz.
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
·10−2
−10
−5
0
5
10
Time (s)
i a(A
)
Detail at t=7ms
Figure 6.7: Instantaneous line current.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 87
From the line currents the equivalent inverter phase currents shown in Fig. 6.1are calculated as
iu = ic − ia, (6.15)
iv = ia − ib, (6.16)
and
iw = ib − ic. (6.17)
The amplitudes of the fundamental phase currents are√3 times larger than
that of the fundamental line currents and lags the line currents by 30 degrees.The inverter phase current ripple will thus also increase by
√3 to a peak-to-
peak value of 0.22 A. The currents through the switches and diodes do in facthave a small ripple component, but this is negligible for the purpose of losscalculations.
6.1.2 IGBT Power Loss and Heat Sink Design
The IGBT module used is the FS30R06W1E3 module from Inneon. Table 6.1shows the characteristic values of the chosen IGBT module.
Table 6.1: FS30R06W1E3 IGBT module parameter values.
Parameter Description ValueVCES Maximum collector-emitter voltage 600 VICnom Maximum continuous dc collector current 30 AICRM Maximum repetitive peak collector current 60 AVGES Maximum gate-emitter peak voltage ±20 VVCE(on) Collector-emitter saturation voltage 1.25 Vton Turn-on (rise) time 28 nstoff Turn-o (fall) time 165 nsRθjc Thermal resistance, junction to case 0.9°C/WRθcs Thermal resistance, case to sink 0.85°C/WVF Forward diode voltage 1.23 V
The inverter phase currents iu, iv and iw are of sinusoidal form with a ripplecomponent superimposed on them as described previously. For loss calcula-tions the approximation is made that these currents are pure sinusoids. Thephase currents are out of phase with the fundamental frequency of the phasevoltages vu, vv and vw as explained. The time period for one complete currentor voltage sine wave is denoted T1.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 88
Similarly as with the MOSFET power losses, the IGBTs also have conduction,switching and diode losses. A sinusoidal current is positive and negative forthe same duration of time; hence the switching losses in all six switches areidentical. The losses are calculated for switch S1. The currents through theIGBT and diode of S1 are seen in Fig. 6.4. As described before, during thepositive half-cycle of the fundamental phase current iuf , most of the currentwill ow from the generator towards the dc bus through the forward biaseddiode D1. Alternatively, current will ow through the IGBT of S2 when itis on. During the negative half-cycle of iuf while S1 is on, current will owthrough the IGBT of S1, otherwise current will ow through the diode D2.
First consider the conduction losses. In order for an IGBT to conduct cur-rent it needs an on-voltage across the collector and emitter terminals referredto as Vce(on) or Von. The average power dissipated in the IGBT of switch S1
over a single period T1 is given by [11]
Pcond = VonIs1, (6.18)
where Is1 is the average current through the IGBT of S1. The average currentis calculated from the fundamental phase current through S1 as well as theduty cycle of switch S1. The fundamental phase current iuf is dened in (6.7)and repeated here for convenience:
iuf = Ipcos(ωs1t+ φ) = Ipcos(θ + φ), (6.19)
where Ip is the amplitude of the fundamental phase current and ωs1 is theangular frequency of each of the modulator reference voltages. The angle φdenotes the phase angle and was calculated in (6.6) as 36.38°. Both the ampli-tude of the fundamental phase current Ip and the phase angle φ will remainfairly constant with respect to the angular frequency, except at low angularfrequencies, where the phase angle will increase (resulting in a decreased powerfactor) to a maximum of 180°, at which point the generator acts as a motor.Similarly the current amplitude Ip will decrease to 0 A due to the fundamentalphase voltages decreasing with frequency. The loss calculations are performedfor the worst case scenario, hence current Ip is at a maximum (15 A) and thephase angle is at 36.38°.
Figure 6.8 shows a detailed graph of the current through the IGBT of switchS1 over three switching periods. During the on-time of each switching periodTs, the current is1 is assumed constant as shown in red in Fig. 6.8, but theamplitude at each switching period will vary [31]. By using a Riemann sumapproximation the area of the waveform shown in Fig. 6.8 can be calculated.When divided by the time period T1, this results in the average current Is1such that
Is1 ≈1
T1
N∑i=1
d1Ipcos(ωs1ti + φ)Ts, (6.20)
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 89
ti -1 ti ti +1 t
|d1Ts|
i s1
Ip cos (ωs1ti + φ)
Figure 6.8: Average switching current approximation taken from [31].
where the duty cycle d1 of switch S1 is given in (6.8). By expanding (6.20) intointegral form and also substituting ωs1t for the angle θ, the average currentover one angular rotation of 2π can be calculated. The lower and upper limitsof the integral are derived from Fig. 6.4 which shows current only ows throughthe IGBT of S1 when ωs1t is between π
2−φ and 3π
2−φ radians. With positive
current dened as owing into the dc bus, the average current Is1 is calculatedas
Is1 = − 1
2π
∫ 3π2−φ
π2−φ
1
2[1 +macos(θ)] Ipcos(θ + φ)dθ
= − Ip4π
∫ 3π2−φ
π2−φ
[cos(θ + φ) +macos(θ)cos(θ + φ)] dθ
= − Ip4π
∫ 3π2−φ
π2−φ
[cos(θ + φ) +
ma
2cos(2θ + φ) +
ma
2cos(−φ)
]dθ
=Ip2π
− Ipma
8cos(φ)
= 1.13 A.
(6.21)
The modulation index ma is dened in (6.3). The maximum rated machinecurrent of 15 A was used for Ip. From (6.18) and with Von given as 1.25 Vin Table 6.1, the average conduction losses for switch S1 was calculated as1.41 W.
The switching losses are due to the small amount of time during whichthere is current through as well as voltage across the IGBT when it changesstate, either from on to o or from o to on. These time periods are denedas ton and toff respectively. With reference to Fig. 6.4, the switching losses inthe IGBT of S1 during one angular rotation ws1 are approximated by [31]
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 90
Pswitch ≈ 1
T1
N∑i=1
1
2Vdc(pp)Ipcos(ωs1ti + φ)(ton + toff )
=Vdc(pp)Ip2T1Ts
(ton + toff )N∑i=1
cos(ωs1ti + φ)Ts
≈Vdc(pp)Ip4πTs
(ton + toff )
[−∫ 3π
2−φ
π2−φ
cos(θ + φ)dθ
]
=Vdc(pp)Ip2πTs
(ton + toff )
= 3.27 W,
(6.22)
where the peak-to-peak dc bus voltage Vdc(pp) is 355 V and the switching fre-quency fs is 20 kHz.
When switch S1 is o and the fundamental phase current iuf is negative,current will ow through the diode of S2. With reference to Fig. 6.4 and from(6.19) and (6.21), the average current Id2 is calculated during the negativehalf-cycle of iuf as
Id2 = −iuf − Is1
= − 1
2π
[∫ 3π2−φ
π2−φ
Iscos(θ + φ)dθ
]− Is1
=Isπ
− Is1
= 3.64 A.
(6.23)
The average conduction loss through diode D2 is given by
PD2 = VF Id2 = 4.48 W, (6.24)
where VF is the diode forward voltage given in Table 6.1 as 1.23 V. The totalpower loss through one diode and one IGBT during T1 is thus
Ptot = Pcond + Pswitch + PD2 = 9.16 W. (6.25)
The total power dissipated in each of the other diodes and IGBTs of theinverter is equal to the power dissipated in the IGBT of S1 and the diodeof S2, respectively. The total power loss inside the complete IGBT module isthus given by
Pmod = 6Ptot = 54.96 W. (6.26)
The heat generated by the IGBT module due to the power loss Pmod is dissi-pated with the help of a heat sink. The rise in junction temperature due tothe losses is given in (3.32) and repeated here for convenience:
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 91
∆Tj = PT (Rθjc +Rθcs +Rθsa) . (6.27)
The junction temperature is permitted to rise to a maximum of 150°C. Ab-solute worst case ambient temperature is assumed to be 40°C, therefore ∆Tj
= 110°C. From the IGBT module's datasheet the junction-to-case Rθjc andcase-to-sink Rθcs thermal resistances are given as 0.9°C/W and 0.85°C/W re-spectively. From (6.26) and (6.27) the sink-to-ambient thermal resistance Rθsa
is required to be 0.251°C/W in order to keep ∆Tj below 110°C. A 150 mmby 210 mm MeccAL aluminium heat sink was available for this project. Fromthe datasheet the sink-to-ambient thermal resistance is given as 0.15°C/W atan airow rate of approximately 5 m/s. Airow is provided by a small 12 Vbrushless dc fan mounted on the side of the heat sink. Thus
∆Tj = 54.96(0.9 + 0.85 + 0.15) = 104, (6.28)
and the maximum allowed ambient temperature is thus 46°C. In the case ofthe chosen IGBT module the junction-to-ambient temperature is rather high.This is due to the junction-to-case and case-to-sink thermal resistances that arethemselves high, due to the small size (and thus surface area) of the module.
6.1.3 Over-Temperature Protection
The IGBT module is equipped with a negative temperature coecient (NTC)thermistor which can be used to detect over-temperature conditions. The tem-perature sensor is located on the case of the IGBT module. The circuit shownin Fig. 6.9 is used to detect the temperature of the IGBT module. A voltagedivider circuit is used to obtain a voltage corresponding to the resistance ofthe thermistor, which is a function of the IGBT module's temperature.
Vcc
Rθ
dsPIC
A/DRd D1
Ve
FaultVe
R1
Figure 6.9: Temperature sensing circuit for the IGBT module.
At a case temperature of 100°C (equating to a junction temperature of 150°C),the thermistor Rθ has a value of 493 Ω with a deviation of ±5 % as indicated in
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 92
the IGBT module's datasheet. The recommended maximum power dissipationinside the NTC thermistor is Pmax = 9.6 mW. The maximum current is thus
Imax =
√Pmax
R100
= 3.74 mA. (6.29)
The supply voltage Vcc for the circuit is chosen as 5 V and henceR1 is calculatedas
R1 =Vcc
Imax
−R100 = 843 Ω. (6.30)
A 1 kΩ resistor was used for R1. The voltage at the output of the voltagedivider circuit is 3.35 V for an IGBT module temperature of 100°C. The outputof the voltage divider is fed into one of the dsPIC's 12-bit analog-to-digitalconverter (ADC) inputs. The dsPIC used as part of the inverter controller isthus also used to perform functions related to the protection of the invertercircuit and general housekeeping tasks.
The dsPIC is programmed to produce an ADC interrupt every 0.8 ms.Inside the interrupt subroutine the ADC's value is compared with the decimalvalue of 2744 which corresponds to 3.35 V. If the ADC's value is smaller than2744, the interrupt is cleared and the dsPIC continues to function normally.
If however the value is bigger than 2744, a 5 V error signal Ve is producedat one of the dsPIC's digital outputs and a red LED is turned on indicating anover-temperature condition occurred. Also, one of the dsPIC's digital outputlines referred to as the Fault line, is set low (0 V). The resistor Rd shown inFig. 6.9 has a value of 220 Ω and is used to limit the current through diodeD1 to less than 15 mA.
The six PWM signals (one for each IGBT) generated by the inverter controlcircuitry each pass through a high-speed CMOS logic 2-input CD74HC08 ANDgate before feeding into an optocoupler located inside a gate driver IC. Theother input to the respective AND gates is an error signal generated by thedsPIC Fault line. A block diagram of this circuit for one PWM channel isshown in Fig. 6.10.
ControlCircuit
PWMSignal
dsPIC
FaultSignal
CD74HC08
GateDriverCircuit
Switch
Figure 6.10: PWM signal and error signal integration.
Take for example switch S1. The controller generates a desired PWM switchingsignal for S1. This PWM signal, along with the error signal, is passed through
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 93
the two input AND gate. The output of the AND gate feeds the gate-drivecircuit used to turn switch S1 on and o. If the error signal is low, meaninga fault condition exists, then the output of the AND gate will remain low.Hence, the signal that is sent to the gate driver is also low, leaving switchS1 in an o state. Consequently if the error signal is high, meaning there isno fault condition, the PWM signal is passed through the AND gate to thegate-drive circuit, enabling the desired controller switching action of S1.
The same error signal is fed into all six AND gates, thus if a fault conditionexists, all the switches are simultaneously switched o, setting the pole pointsof the inverter into a high impedance state.
Since a high junction temperature is caused by the switching action, withthe switching action halted, the junction as well as case temperature will dropbelow the fault threshold temperature. The LED powered by the dsPIC willhowever not turn o if the case temperature drops below 100°C, nor will theinverter start switching again. The dsPIC Fault line will remain low. ThedsPIC and controller circuity have to be restarted in order to clear the errorcondition and resume normal switching action at the IGBT gates. This isdone purely for safety purposes and to ensure the cause of the fault conditionis dealt with.
To validate this safety feature a heatgun was used to heat the bottomof the IGBT module casing (which is usually connected to the heat sink) toabove 100°C. The red LED was observed to turn on and the inverter switchingaction terminated. Only once the circuit was restarted did the LED turn oand the inverter resume its switching action. The temperature at which thissafety feature responded was measured by an infrared thermometer as roughly100°C.
6.1.4 Isolated Gate-Drive Circuitry
The isolated gate-drive circuit used to communicate the switching signals tothe gate of each IGBT is similar to the gate-drive circuit used for the MOS-FET switches described in Section 3.1.5. To provide isolation between thecontrol circuit and switches an IGBT gate-drive optocoupler IC is used. Thegate driver is powered by a small push-pull oscillator circuit with a toroidaltransformer and two fast rectier diodes, identical to the circuit used in Sec-tion 3.1.5. The supply voltage to the oscillator circuit is however changed to15 V and the transformer ratio to 10:11. The gate driver IC supply voltage isthus 16.5 V.
The gate driver IC chosen is the ACPL-333J gate driver from Avago Tech-nologies. It is specically designed as an IGBT gate-drive optocoupler equippedwith integrated (Vce) desaturation detection and feedback, undervoltage lock-out and active Miller clamping [35]. The characteristic values of the gate driverare given in Table 6.2.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 94
Table 6.2: ACPL-333J gate driver characteristic values.
Parameter Description ValueVCC2 Positive supply voltage range 15-30 VIO Maximum peak output current 2.5 AVIORM Maximum working insulation voltage 1414 VVUV LO Undervoltage lockout threshold 11.6 VVDESAT Desaturation detection threshold 7 VIM Maximum active Miller clamp current 1.7 Atp Maximum propagation delay time 250 nsCM Common mode transient immunity 50 kV/µs
The maximum voltage across one of the IGBT switches is the rail-to-rail busvoltage of 355 V. The turn-on time of the switches is given as 28 ns and theturn-o time as 165 ns in Table 6.1. The turn-on time is signicantly fasterthan the turn-o time; hence the maximum rate of change of IGBT voltageoccurs when the switch is turned on where
dV
dt=
Vmax
ton= 12.7 kV/µs. (6.31)
The dV/dt rating of the gate driver is thus sucient to support the switchingaction of the IGBTs, since the gate driver is rated to withstand 50 kV/µs.
ACPL-333J
VCC2
VEE
DESAT
Vclamp
Vout
VEE
Rg
RPD
Rb D1 Z1
Cb
ANODE
CATHODE
VCC1
Fault
Vs
Rp
PWMSignal
RF
CF
dsPICInput
+5VOscillatorPowerCircuit
Figure 6.11: Block diagram of the IGBT gate-drive circuit.
A block diagram of the isolated gate-drive circuit is shown in Fig. 6.11. Dur-ing normal operation the PWM signal from the control circuit is fed through
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 95
resistor Rp into the anode of the light emitting diode located inside the gatedriver's optocoupler which then controls the output voltage Vout (as either highor low). The resistor Rp is chosen as 470 Ω. At the same time the desatura-tion detection circuitry is continuously monitored and the Fault line outputvoltage is high.
Once the desaturation detection picks up a Vce value of more than 7 V ittriggers a fault condition, during which time the output voltage Vout is reducedto zero. This allows the switch to be turned o softly and prevent large di/dtinduced voltage spikes. The internal fault feedback is activated which changesthe Fault line voltage from high to low in order to notify the control circuit.The PWM signal from the control circuit is ignored during the fault period.After a xed period of approximately 26 µs the fault pin is reset and normaloperation will continue unless Vce is still larger than 7 V.
The fault feedback is given as a digital input to the dsPIC digital sig-nal processor. Similarly as with an over-temperature fault, a yellow LED isswitched on indicating an overcurrent fault occurred at one or more of theIGBT switches. The dsPIC output Fault line is set low and the same se-quence of events will follow as described in the previous subsection. Essentiallyall switching action is stopped and the inverter is shut down. Again the circuitis required to restart before the dsPIC Fault line will change to a high state,allowing the switching action to resume and for the yellow LED to turn o.
The output of the gate driver is controlled by the PWM control signal,the protection circuitry, the desaturation detection circuitry as well as by theundervoltage lockout protection. If the voltage supplied by the oscillator circuitto the gate driver is less than 11.6 V, Vout is not allowed to go high due tothe undervoltage lockout feature of the gate driver. Once the supply voltageis again greater than the threshold voltage, the output is allowed to go high.This feature protects the gate driver and the switch if the oscillator circuitwere to fail.
The active Miller clamp capability allows the use of a positive supply volt-age only for the gate driver, while still controlling the Miller current duringhigh dV/dt transients. When the IGBT is switched o the gate voltage ismonitored. The clamp circuitry is only activated when the gate voltage dropsbelow 2 V, providing a low impedance path to sink the Miller current quickly.The clamp is released as soon as the next optocoupler LED trigger from thecontrol circuit is received.
To ensure the maximum output current the gate driver can supply is notexceeded, the gate resistor Rg shown in Fig.6.11 is required to have a minimumvalue of
Rg(min) =Vcc2
IO=
16.5
2.5= 6.6 Ω. (6.32)
The nal value for Rg was chosen as 10 Ω, thus limiting the output current to1.65 A. To clamp the output voltage Vout at Vcc2 (16.5 V) a pull-down resistor
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 96
is required and is recommended in the datasheet to sink 650 µA of currentwhile the output is high. The pull-down resistor is calculated as RPD = 16.5
0.65
= 25.4 kΩ. The nal value for the pull-down resistor was chosen as 27 kΩ.The Fault output is connected internally within the gate driver to the
collector of a BJT. A passive pull-up resistor RF is thus required and therecommended value is given as 2.1 kΩ. A 2.2 kΩ resistor was used for RF .Along with the pull-up resistor, a 1 nF ltering capacitor CF is recommendedbetween Fault and Vs thus CF was chosen as 1 nF.
As seen in Fig. 6.11 the desaturation detection circuit requires componentsRb, Cb, D1 and Z1 to be connected externally from the gate driver to functioncorrectly. The desaturation detection is internally set to trigger when Vce
equals 7 V, but for the chosen IGBT module this value is too high. At a Vce
value of approximately 1.7 V the collector current will already reach a value ofaround 25 A. Since the peak current through the inverter switches will neverexceed 15 A, the desaturation protection is chosen to trigger at Vce = 1.7 V.To achieve this a 3.9 V zener diode denoted by Z1 in Fig. 6.11 was placed inseries with D1, a BYV26EGP ultrafast diode with a forward voltage of 1.4 V.
The values for Rb and Cb are specied in the datasheet as 100 Ω and100 pF respectively. Resistor Rb is simply used as a protection resistor to limitthe current from the gate driver such that it will not damage the IC. Thedesaturation fault detection circuitry is designed to stop any fault feedbackonce the IGBT switches on for a short period of time, to allow the Vce value tofall below the threshold value. This time period can be adjusted by changingCb. With Cb = 100 pF the blanking time is 2.7 µs.
6.1.5 Overvoltage Protection
The dc bus voltage is regulated by the dc-dc converter circuit. There areonly two probable situations which will cause the dc bus voltage to exceed thenominal voltage of 355 V rail-to-rail. For both cases it is assumed there is noload connected to drain the bus voltage.
The bus voltage will rise if the current owing from the generator intothe dc bus is greater than the maximum current that each dc-dc converter ispermitted to conduct. This occurs at a mean dc bus current of 6.857 A. Theadditional current will ow into the capacitors thus charging them to a valuehigher than the nominal voltage. A value of 6.857 A corresponds to an averageinductor current of 10 A which is the current limit for charging the batteriesas given in the battery datasheet. Since the mean bus current is regulated bythe inverter circuit, it should never exceed an average current of 5.7 A unlessthe control circuit fails.
Secondly, the dc bus voltage will rise if the converter circuit is switched owhile the generator is still supplying power to the dc bus. If the converter orbatteries are disconnected there is nothing to sink the dc bus current exceptfor the bus capacitors. They will continue to charge until they reach voltages
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 97
higher than what they are rated to operate at. In order to protect the capac-itors as well as the other high voltage components such as the MOSFETs andIGBTs, overvoltage protection was added to the inverter control circuit.
The dc bus voltage is monitored and a scaled version of the bus voltage isfed into a comparator. If the bus voltage exceeds 365 V, which is 5 V higherthan the allowed operating range, the comparator output changes from a high(5 V) to a low (0 V) state. The output of the comparator connects to one ofthe digital input lines on the dsPIC as well as directly to the AND gates usedfor sending the switching signals to the optocoupler inside the gate driver IC.The switching action is thus stopped immediately after the overvoltage faultoccurred. The dsPIC ensures the fault is detected and keeps the digital outputfault line low until the control circuit is reset.
6.2 PWM Controller Design
A block diagram of the proposed PWM control circuit is shown in Fig. 6.12.A double control loop approach is followed with an inner and outer controlloop. The inner loop functions to control the amount of current owing fromthe inverter into the dc bus, while the outer control loop regulates the batteryvoltage of the converter. The inner current control loop measures and comparesthe mean dc bus current with the set-point voltage vseti generated by the outervoltage control loop, where one volt at vseti corresponds to 1.14 A of currentsupplied by the generator to the dc bus. The set-point vseti is dependant uponthe state of charge of the converter's batteries as determined by the voltagecontrol loop.
Vbat∗ Σ
+
−vfbv
VoltageController
vrefi
Limiter
vseti
Σ+
−vfbi
CurrentController
Scaling
MachineModel
InvPlant
idc ConvPlant
vbat
CurrentSensor
VoltageSensor
ModulatorCircuitInner Loop
vε vcs
Figure 6.12: Block diagram of the inverter control circuit.
The battery voltage set-point Vbat∗ is a constant corresponding to 272 V. The
resulting individual battery potential is 13.6 V for each of the twenty lead-acidbatteries connected to the converter circuit.
As seen from Fig. 6.12, the current set-point vseti, generated by the voltagecontrol loop, is limited to between 0 V and 5 V such that the average dc bus
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 98
current is limited to a minimum of 0 A (to ensure the induction machine neveroperates continuously as a motor) and a maximum of 5.7 A.
The magnitude of the control signal vcs generated by the current controller,as seen in Fig. 6.12, is also limited to 5 V since it feeds into an analog inputpin on the dsPIC DSP. The generator shaft speed is measured and also givenas an input to the dsPIC. Inside the dsPIC the magnitude of the control signalis scaled according to the angular speed of the generator shaft to ensure aconstant Volt/Hz ratio is maintained.
The dsPIC is responsible for producing the sinusoidal reference waveformsthat are compared with the triangular carrier waveform to generate the re-quired PWM switching signals for the six IGBTs. In Fig. 6.12 the PWMgeneration circuit is denoted as the modulator circuit and will be described inmore detail in the subsections to follow.
Both the voltage and current controllers are designed using frequency do-main analysis. The necessary time domain equations are derived to obtain thefrequency domain equations by using the Laplace transform. The time domainsimulations are only presented in the next chapter alongside the measured testresults for comparison purposes.
6.2.1 Inverter Plant
The plant portion of the inverter circuit is shown in Fig. 6.13. It is used toidentify the relationship between the generator line voltages and the currentowing into the dc bus. The inverter phase voltages, and hence also the gener-ator line voltages, are controlled by the switching action of the IGBTs dictatedby the current control loop.
The approximation is made that the instantaneous power owing into thedc bus equals the instantaneous power owing out of the generator, hence theswitching and conduction losses are not taken into account. The dc bus currentis related to either the inverter phase voltages and currents or the generatorline voltages and currents, since the total instantaneous power from the in-verter phase voltages and currents must equal the total instantaneous powerfrom the generator line voltages and currents. With reference to Fig. 6.13 theinstantaneous power owing into the dc bus is given by [11]
Vdc(pp)idc = vaia + vbib + vcic, (6.33)
where Vdc(pp) is the converter regulated dc bus voltage of 355 V and idc is thetime-varying dc bus current. The line currents are sinusoidal with a small rip-ple component superimposed on them and can thus be viewed as pure sinusoidswith only the fundamental components present.
The triangular carrier waveform's switching frequency is much larger thanthe fundamental frequency of the line voltages, resulting in a large frequencymodulation ratio. The amplitudes of the subharmonics will thus be very small
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 99
idc
Ci
+Vdc
−Vdc
S1
is1
id1
S2
vu
is2
id2
S3
is3
id3
S4
vviv
ib vb
is4
id4
S5
is5
id5
S6
vwiw
iu
is6
id6
icvcia
va
3-PhaseMachine
Figure 6.13: Bidirectional inverter circuit.
and it can be approximated that only the fundamental frequency component ofthe line voltages will contribute towards the instantaneous power generated bythe induction machine [11]. With the line voltages and currents approximatedas fundamental line voltages and currents, (6.33) reduces to
idc =2Vs(rms)Is(rms)
Vdc(pp)
[cos(ωs1t)cos(ωs1t+ φ)
+ cos(ωs1t− 120)cos(ωs1t− 120 + φ)
+ cos(ωs1t+ 120)cos(ωs1t+ 120 + φ)]
=3Vs(rms)Is(rms)
Vdc(pp)
cosφ,
(6.34)
where Vs(rms) and Is(rms) are the rms values of the fundamental line voltagesand currents, respectively. The angle φ denotes the power factor angle asdescribed earlier. The rms values of the generator line voltages and currentsare related by the equivalent generator impedance Zeq such that
Is(rms) =Vs(rms)
|Zeq|, (6.35)
where |Zeq| is the magnitude of the equivalent generator impedance given in(5.17). Recall the amplitudes, and therefore also the rms values of the fun-damental line voltages, are varied in proportion to the angular frequency of
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 100
the fundamental waveforms in order to maintain constant rms ux within thegenerator core. The impedance Zeq is also frequency dependant and thus from(6.35) the rms value of the line currents will remain fairly constant. At lowfrequencies (below 20 Hz) the real component of the impedance dominates theimaginary component that varies with frequency and hence Is(rms) will changewith Vs(rms).
The control circuit is designed for the worst case scenario and thus Is(rms)
is assumed constant and equal to the maximum delta connected rms currentgiven in the machine datasheet as 6.07 A for the frequency domain analysis.
The phase angle φ with respect to the generator frequency is shown inFig. 6.14 and was calculated using the equivalent generator model derived inChapter 5. The phase angle is seen to remain fairly constant at frequenciesabove 25 Hz. The generator will only be allowed to operate at frequenciesabove 7 Hz since it draws instead of delivers current at lower frequencies. Theangle φ is assumed constant and equal to 36.38° as calculated from (6.6) forthe frequency domain analysis.
10 20 30 40 50 60 70 800
45
90
135
180
Frequency (Hz)
φ(deg)
Figure 6.14: Phase angle over generator operating frequency range.
The Laplace transform of the (6.34) yields
Idc(s) =3Is(rms)
Vdc(pp)
cosφVs(s), (6.36)
where Vs(s) is the rms line voltage of the generator that varies with frequency.The transfer function of the inverter plant Gp(s), describing the average dcbus current as a function of the rms line voltage, is thus given by
Gp(s) =Idc(s)
Vs(s)=
3Is(rms)
Vdc(pp)
cosφ. (6.37)
The plant portion of the inverter is thus combined with the machine modelshown in Fig. 6.12.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 101
6.2.2 Digital Signal Processor Interfacing
The main function of the dsPIC is to assist with the control of the inverterswitches. It is also used alongside protection circuitry as mentioned throughoutthis chapter. The 30F4013 dsPIC from Microchip was chosen for its DSPcapabilities and functionality. The parameters of the dsPIC is summarised inTable. 6.3.
Table 6.3: Parameter values of the dsPIC30F4013 DSP.
Parameter ValueArchitecture 16-bitCPU speed 30 MIPSProgram memory 48 KBRAM 2048 BEEPROM 1024 BOperating voltage 2.5 V to 5.5 VOperating temperature -40°C to 125°CPin count 40Input/output pins 30Communication peripherals 2 UART, 1 SPI, 1 I2CAnalog peripherals 1-ADC, 13 x 12-bit at 200 kspsTimers 5 x 16-bit, 2 x 32-bit
The dsPIC is operated at 5 V and 10 MHz. An external 10 MHz crystaloscillator was used to generate the internal 10 MHz clock. The dsPIC wasprogrammed using a dsPICDEM 2 development board and an MPLAB ICD2in-circuit debugger.
The generator is driven by an identical induction motor controlled by acommercial o-the-shelf variable-speed drive system to simulate airow througha turbine that would otherwise apply torque to the generator shaft. Due to thelarge moment of inertia, the speed at which the generator shaft is turning willnot change rapidly; hence vector control is not required. The stator voltagesare controlled as a function of the dc bus current, but Volt/Hertz control isalso included inside the inner control loop using the dsPIC to ensure the ma-chine's magnetic rms ux stays constant. The detailed operation of the dsPICis described in this section.
The current controller produces a control signal vcs which feeds into one ofthe 12-bit ADC channels on the dsPIC. The dsPIC is operated at 5 V, hencethe maximum allowed voltage at any of the input pins is also 5 V. The controlsignal vcs is limited to between 0 V and 5 V by means of an active clampcircuit identical to the circuit used for limiting current and voltage inside theconverter as described in Section 3.2.4.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 102
To determine the shaft speed of the generator a magnetic pick-up sensor wasused. The functioning of the magnetic pick-up circuit is demonstrated inFig. 6.15. The magnetic pick-up consists of a small permanent magnet and acoil wound around the magnet. The permanent magnet establishes a magneticeld that couples to the coil. When a ferromagnetic material passes in closeproximity to the magnet, the magnetic eld is displaced about the coil. Thus,by Faraday's law, a voltage is induced in the winding. The output voltage isproportional to the strength and rate of change of the magnetic eld coupledto the winding.
Magnetic Pickup
N S
V1
V2
Shaft Coupling
ωr
Figure 6.15: Shaft coupling with magnetic pick-up for speed feedback.
The shaft coupling, connecting the generator and motor shafts, was used as theferromagnetic object, as shown in Fig 6.15. Six holes with a diameter of 14 mmeach, separated exactly 60° from each other, were drilled 10 mm deep into thecircular shaft coupling. The magnetic pick-up was mounted approximately2 mm away from the shaft. One complete rotation thus produces 6 voltagepulses between the two output terminals V1 and V2.
The output signal of the magnetic pick-up is processed using the circuitshown in Fig. 6.16. The coil wound around the permanent magnet can berepresented as an inductor L in series with a resistor R. The output voltagepulses are rst passed through an operational amplier with a gain of approxi-mately 3 to amplify the speed feedback signal. Once the output of the op-ampexceeds 0.45 V, the LM311 comparator changes from a high (5 V) to a low(0 V) state. Even if the shaft is turning at very low speeds, the output voltagewill exceed 0.45 V and thus generate a state-change at the comparator output.Positive feedback is added to provide hysteresis which ensures that no falsestate changes occur.
The output of the comparator is fed into one of the dedicated externalinterrupt pins on the dsPIC. The interrupt will trigger an interrupt subroutine
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 103
L
R V1
V2
−
+
Ri
Rf
Ri
Rf
Rpu
+5V
R1
+5V
R2 Rfb
vmp
Figure 6.16: Magnetic pick-up interfacing with the dsPIC.
on every falling edge of the comparator output voltage vmp. The values of thecomponents used in Fig. 6.16 are given in Table 6.4
Table 6.4: Component values for the magnetic pick-up and dsPIC interfacingcircuit.
Component ValueL 350 mHR 580 ΩRi 3.3 kΩRf 10 kΩR1 10 kΩR2 1 kΩRfb 22 kΩRpu 1.8 kΩ
One of the dsPIC's two 32-bit timers is used to count the time interval betweeninterrupts generated by the magnetic pick-up circuit. As soon as the rstmagnetic pick-up is received the timer is reset and starts counting from zero.One count is added every 100 ns since the timer is set to run at the internal10 MHz clock speed of the processor. Once the second interrupt is received,the timer value is stored in memory and thereafter the timer is reset to onceagain start counting from zero. The value that is stored in memory is used tocalculate the frequency of the generator shaft since
f =Fc
6CT
, (6.38)
where f is the shaft frequency measured in Hertz, CT is the timer value storedin memory and Fc is the clock frequency of 10 MHz. The factor 1
6is present
due to the magnetic pick-up producing six interrupts during one completeshaft rotation. The measured shaft frequency f is further divided by a basefrequency fb of 65 Hz to generate a scaling factor Kc. A base frequency of
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 104
65 Hz was chosen to avoid saturation currents when the generator is operatedat a slip of -5.47 %. The scaling factor scales the control signal vcs to ensure thegenerator is always operated at constant ux, thus the maximum amplitude Vc
of the dsPIC and DAC output modulator reference waveforms are calculatedas
Vc =f
fbvcs = Kcvcs. (6.39)
If the measured frequency of the generator is less than 7 Hz, the control signalvcs is gradually reduced to zero inside the dsPIC since the machine simulationpredicted the generator is only able to supply power when operating at afrequency above 7 Hz. Below 7 Hz the generator will actually draw powerfrom the dc bus to keep it running when operated at a slip of -5.47 %.
Similarly, the control signal vcs is reduced to zero inside the dsPIC whenthe measured shaft frequency exceeds 75 Hz (4500 rpm), which is the speciedmaximum operating frequency of generator system. When the generator isrunning at either under 7 Hz or above 75 Hz an orange LED is switched on,visually indicating the generator shaft speed is out of range. When the controlsignal vcs reaches zero the switches are all switching at a constant duty cycleof 50 % and the amplitudes of the fundamental phase voltages applied to thegenerator stator terminals are also zero. No power will thus be transferredbetween the generator and the dc bus.
The orange LED switch o once the shaft speed is again within the speciedrange. At the same time the control voltage is gradually increased to thedesired value.
In order to generate the required PWM signals, three 120° separated si-nusoids have to be compared with the 20 kHz triangular carrier waveform asexplained in the previous section. The dsPIC is used to generate these three si-nusoidal modulator reference waveforms using a look-up table (LUT) approachsimilar to the approach followed in [44] and [45] where a part of the sinusoidis pre-computed and stored in the memory of the processor used, thus forminga binary LUT.
The maximum resolution of the sine wave generated within the dsPIC ischosen as 12 bits, where the 12th bit is used to indicate the polarity of thesinusoid as either positive or negative. The sampling speed of the sine wave waschosen as 128 samples per 60°. The sine wave was modelled using MATLABwhere the amplitude of the sine wave peaked at 2047 corresponding to theavailable 11-bit magnitude inside the dsPIC.
Since the induction generator is designed to function at a slip of -5.47 %,the number of samples during 60 degrees of rotation is 128
0.948= 135 samples. To
simplify the look-up process inside the dsPIC, the entire positive half of thesine wave generated in MATLAB was stored in the EEPROM. The EEPROMthus contain 135× 3 = 405 samples where each sample is represented using 2bytes.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 105
The three sinusoidal modulator reference waveforms vru, vrv and vrw generatedby the dsPIC, are shown in Fig. 6.17. Each have their own LUT pointer. Theamplitudes of the modulator reference waveforms are equal and denoted Vc,as dened in (6.1) for vru. At a sample speed of 128 samples per every 60°,the counter value CT is divided by 128 to produce a timer limit LT for a16-bit timer. Each time the 16-bit timer reaches the limit LT an interruptis generated and the three pointers are all incremented to point to the nextcorresponding value inside the LUT.
vcs
vmp
dsPIC30F4013
vrv
vru
vrw
Figure 6.17: Three-phase reference waveform generation with the dsPIC.
The LUT value is then manipulated such that the 12th bit is set high if thesinusoid is negative. Also the scaling factor Kc is used to scale the magnitudeof the LUT value (and thus also scale the magnitude of the sinusoid) accordingto the value read in by the ADC for vcs and the speed of the generator shaft.
The dsPIC was programmed in assembly language using the MPLAB ed-itor. Assembly was chosen instead of the C programming language to haveprecise control over the number of instruction cycles that occur after the mag-netic pick-up produces an interrupt. This ensures the correct timer value isstored in memory and allows for a more accurate shaft speed calculation.
The three manipulated LUT values are communicated to a 12-bit AD7398digital-to-analog converter (DAC) using the serial peripheral interface (SPI)protocol. The DAC outputs three modulator reference sinusoids, each with amaximum amplitude of 2.5 V and a dc oset of 2.5 V. The dc oset is removedand a gain of two is added using a dierential LF353 operational amplier. Thethree resulting modulator reference waveforms, with a maximum amplitudeVc of 5 V, are used for comparison with the triangular carrier waveform togenerate the PWM switching signals using high-speed AD790JN comparatorsfrom Analog Devices.
The complete circuit diagram containing the dsPIC, DAC and magneticpick-up interface circuit is given in Appendix A.2.2. A small PCB containingthese circuit components is mounted on top of the analog control PCB. BothPCBs are given in Appendix B.2.2.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 106
6.2.3 Current Control Loop
The current control loop, shown in Fig. 6.18, functions to control the meandc bus current by controlling the switching action of the IGBTs using bothanalog and digital circuitry. Essentially the controlling of the switches allowsthe amount of power generated by the induction machine to be controlled,and since the dc bus voltage is assumed constant, the dc bus current is thuscontrolled. The digital part of the controller was already described in theprevious subsection along with the PWM signal generation.
Idc∗
(Vseti)Σ
+
−Vfbi
Gc(s) KcVc
P (s) KmVs
Gp(s)
H(s)
IdcVcs
Figure 6.18: DC bus current control loop block diagram.
In Fig. 6.18 Gp(s) denotes the transfer function of the inverter plant portionderived in (6.37) and the transfer function H(s) denotes the current sensingfeedback circuit. The current control loop's compensation amplier is de-scribed by the transfer function Gc(s).
The transfer function P (s) shown in Fig. 6.18 is a third order Padé approx-imation [46] used to approximate the various time delays within the currentcontrol loop as a time-continuous function for frequency domain analysis. Thetime delays, mainly due to the dsPIC, can potentially cause system instabil-ity and it is therefore important to include the time delays in the frequencydomain analysis. Note, a time delay introduces phase shift only and does notaect the magnitude of the control loop's response. The total control looptime delay was overestimated as 1 ms. The Padé function within MATLABwas used to model this time delay.
The scaling factor Kc used for the Volt/Hertz scaling within the dsPIC(as detailed in the previous subsection) is also included in the current controlloop as shown in Fig. 6.18. The scaling factor has a maximum value of 1and a minimum value of 0.11 (at a generator shaft frequency of 7 Hz). Thegain denoted Km in Fig. 6.18 is used to represent the modulator circuit and isdescribed in detail next.
A linear model of the PWM circuit is assumed, thus a linear relationshipexists between the sinusoidal modulator reference signals (vru,vrv and vrw) andthe fundamental inverter phase voltages (vuf ,vvf and vwf ). The rms voltageVs(rms) of the fundamental generator line voltages (vaf ,vbf and vcf ) is relatedto the amplitude of the inverter phase voltages Vp by
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 107
Vs(rms) =
√3√2Vp. (6.40)
Substituting (6.4) and (6.2) into the above equation yields
Vs(rms) =
√3
2√2maVdc(pp)
=
√3
2√2
(Vc
Vr
)Vdc(pp)
≈0.612Vdc(pp)
Vr
Vc,
(6.41)
where Vr is the amplitude of the triangular carrier waveform and Vc is dsPICscaled amplitude of the modulator reference voltages. The Laplace transformof the above equation yields:
Vs(s) ≈0.612Vdc(pp)
Vr
Vc(s). (6.42)
With reference to Fig. 6.18, the frequency dependant rms line voltage Vs(s)and modulator reference waveform amplitude Vc(s) are thus related by Km
where
Km =Vs(s)
Vc(s)=
0.612Vdc(pp)
Vr
. (6.43)
The dc bus current is measured using an ACS756 fully integrated Hall ef-fect sensor from Allegro. The characteristics of the sensor is summarised inTable 6.5.
Table 6.5: Allegro ACS756 Hall eect sensor characteristics.
Characteristic Description ValueVcc Supply voltage 5 VIcc Supply current 10 mAtPROP Propagation time 1 µsIp Primary Sampled Current ±50 ASensTA Sensitivity 40 mV/AETOT Maximum total output error ±5 %VISO Voltage Isolation 3 kVTOP Ambient Operating Temperature -40°C to 125°C
At a sensitivity of 40 mV/A and a maximum dc bus current of 5.7 A, themaximum output voltage is 0.228 V. The desired output voltage is 5 V for themaximum dc bus current, hence a gain of 21.93 was required. The circuit used
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 108
to convert the measured current into the equivalent desired voltage is shownin Fig. 6.19. The component values for the current feedback circuit are givenin Table 6.6.
ACS756
GND
VCC
viout
Ip+
Ip−
5Vidc
−
+
LF353N
Ri
Rv
V5p
Rf
R1
R2 C1
−
+
R1
R2
C1
vfbi
Figure 6.19: DC bus current sensor feedback circuit.
Table 6.6: Component values for the current feedback circuit.
Component ValueV5p 5 VRv 10 kΩRi 4.7 kΩRf 10 kΩR1 3.3 kΩR2 36.3 kΩC1 100 nF
An LF353 operational amplier is used to provide the additional gain required.The rst op-amp circuit is used to remove the 2.5 V dc oset from the Halleect sensor's output viout. The resistor Rv is a variable resistor to trim thedc oset in case of a small voltage oset error at the sensor's output for a 0 Ameasurement. The output voltage for the rst op-amp circuit is given by
vout =Rf
Ri +Rv
(viout − V5p) + viout. (6.44)
The ratio between the feedback resistor Rf and the two input resistors Rv+Ri
should ideally equal one thus (6.45) becomes
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 109
vout = 2viout − V5p. (6.45)
The input V5p is a precision 5 V reference source generated by a REF02 ICfrom Analog Devices. With a dc oset of 2.5 V at viout, the output voltagevout has a gain of 2 with respect to the two input voltages viout and V5p.
The second op-amp circuit is a low pass lter with a transfer function of
L(s) =Zf
Zi
=R2
R1
(1
1 + sC1R2
), (6.46)
and a corner frequency of
f1 =1
2πC1R2
, (6.47)
which equals 43.84 Hz with the component values given in Table 6.6. Thelow pass lter is used to remove the ripple component from the measured dcbus current as well as provide a dc gain of 11. The transfer function for thecomplete current feedback circuit shown in Fig. 6.19 is denoted H(s) and isgiven by
H(s) =Vfbi(s)
Idc(s)=
2RcsR2
R1 (1 + sC1R2), (6.48)
where Rcs equals 40 mΩ and is used to include the 40 mV/A sensitivity of theHall eect sensor. With reference to Fig. 6.18 the open-loop transfer functionof the current control loop Gol(s) is
Gol(s) = Gc(s)KcP (s)KmGp(s)H(s), (6.49)
and closed-loop transfer function Gcl(s) is
Gcl(s) =Gc(s)KcP (s)KmGp(s)
1 +Gc(s)KcP (s)KmGp(s)H(s), (6.50)
where Gc(s) is the transfer function of the compensation amplier and Gp(s) isthe transfer function of the inverter plant derived in (6.37). The frequency do-main analysis of Gol(s) with the dsPIC scaling factor Kc at its maximum valueof 1 is shown in Fig. 6.20, without the transfer function of the compensationamplier present.
The magnitude plot shown in Fig. 6.20 has a corner frequency equal to f1and rolls o at -20 db/decade thereafter. The bandwidth of the current controlloop without the compensation amplier present is determined from Fig. 6.20to be 37.3 Hz. The phase plot indicates a large phase shift mainly due to thethird order Padé approximation of the time delay within the current controlloop.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 110
−40
−20
0
Magnitude
(dB)
100 101 102 103 104−720
−540
−360
−180
0
Frequency (Hz)
Phase
(deg)
Figure 6.20: Bode plot of the open-loop response of the current control loopwithout the compensator present.
The speed of the current control loop should ideally be lower than the lowestoperating speed of the generator (7 Hz) since the generator will inherentlyreact slowly to a change in the applied stator voltages.
−
+vcs
Rvfbi
C
Rvseti
C
Figure 6.21: Active integrator compensation circuit.
In order to provide a theoretical innite gain at dc, an integrator is requiredin the compensation amplier. The integrator will also lower the bandwidthof the current control loop and is thus in itself an adequate compensationamplier for the current control loop. The active integrator circuit is shownin Fig. 6.21. The transfer function for the compensation amplier is given by
Gc(s) =Vcs
Vseti − Vfbi
=1
sCR. (6.51)
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 111
The resistor R was chosen as 120 kΩ and the capacitor C as 1 µF. The op-ampused is the LF353 JFET operational amplier from STMicroelectronics.
The frequency domain analysis of both the open- and closed-loop transferfunctions Gol(s) and Gcl(s) are shown in Fig. 6.22. The open-loop transferfunction is seen to have a unity gain crossover frequency of 1.74 Hz whichis within the desired range. The corresponding phase margin is measured as87.1° and the open-loop response shows a -20 db/decade roll-o when it crossesthe zero dB line. The gain margin is determined from Fig. 6.22 as 39.6 dB.With both a positive gain and phase margin the current control loop is shownto be stable.
−150
−100
−50
0
50
Magnitude
(dB)
10−1 100 101 102 103 104−720
−540
−360
−180
0
Frequency (Hz)
Phase
(deg)
Gol(s)Gcl(s)
Figure 6.22: Bode plot of the open-loop response of the complete currentcontrol loop.
The closed-loop response shown in Fig. 6.22 has a dc gain of 1.135 dB corre-sponding to a loop sensitivity of 1.14 A of dc bus current per volt of excitationapplied to the current set-point Vseti. At a maximum set-point voltage of 5 Vthe corresponding dc bus current will rise to and settle at 5.7 A as designedfor.
Due to the large 1 µF capacitor used in the compensation amplier, integralwindup occurs. Recall the output of the compensator Vcs is limited to 5 V.The integral windup will cause a large initial current overshoot in the dc busand will also lead to a slow settling time [47]. Since the current owing into thebatteries is current limited, a large current overshoot is undesirable, especiallyif it will take a prolonged period of time to settle. To combat the integralwindup, a tracking back anti-windup scheme [41] is used. Figure 6.23 shows
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 112
a block diagram of the tracking back scheme when using an integrator for thecompensation circuit.
Vseti Σ+
−Vfbi
Σ+
−Vaw
1
sKi
Limiter
Vcs
Σ
+−
Ka
H(s) Idc
Figure 6.23: Tracking back anti-windup scheme with an integral controller.
The gain block labelled Ki in Fig. 6.23 is the integrating factor 1RC
from (6.51),while Ka is the anti-windup factor. The dierence between the limited (satu-rated) output and the unsaturated output of the integrator is fed back to theinput of the integrator after adding gain Ka to the signal, thus producing Vaw.Although it is desirable to make Ka very large to reduce the integrator windupquickly, it will also increase the settling time. Ideally the constant Ka shouldbe close to unity when using an integral compensation method [41]. By testingthe prototype circuit the optimal value for Ka was determined to be 1.27.
The anti-windup circuitry only aects the controller when the integratoroutput is higher or lower than the clamped output Vcs. When the integratoroutput is within the desired operating range, the saturated and unsaturatedoutputs are equal to each other and thus the dierence between them is zero.The voltage at Vaw is thus also zero and will have no eect on the compensationcircuit.
6.2.4 Voltage Control Loop
The combined current and voltage control loops are shown in Fig. 6.24, wherethe inner current control loop is enclosed by the outer voltage control loop.The functioning of the current control loop was presented in the previoussubsection. The outer voltage control loop as shown in Fig. 6.24 functions toregulate the mean battery potential of the 20 lead-acid batteries used in thedc-dc converter circuit.
The transfer function Fp(s) shown in Fig. 6.24 denotes the plant portionof the dc-dc converter in order to obtain a relationship between the dc buscurrent and the state of charge of the batteries. The battery voltage feedbackcircuit has a transfer function J(s) and the transfer function Fc(s) denotes thecompensation circuit used for the outer voltage control loop.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 113
Vbat∗ Σ
+
−Vfbv
Fc(s)
Vseti
Σ+
−Vfbi
Gc(s) Kc P (s) Km Gp(s)Idc
Fp(s) Vbat(s)
H(s)
J(s)
Vε Vcs
Figure 6.24: Voltage and current control loop block diagram.
Either the converter's batteries, the load connected between the dc bus rails,or both, are responsible for sinking the current owing into the dc bus fromthe inverter. Since the load is not permanently connected, current will owinto the batteries, thus charging them. The batteries have a maximum oatinguse potential of 13.6 V to 13.8 V per battery which should not be exceeded.For the safety of the batteries the voltage control loop will strive to keep thebatteries charged to a maximum mean potential of 13.6 V per battery, thus272 V in total.
With reference to Fig. 6.12 and Fig. 6.24 the plant portion of the convertercircuit is derived next. The 40 Ah batteries will take hours to fully chargeand thus simulating their true response with respect to the current they sinkrequires an extremely long simulation. For time domain analysis the batterywas approximated as a constant voltage source in series with a resistor andhence the total battery voltage vbat is given by
vbat = VB + iLRin, (6.52)
where VB is the nominal battery voltage, iL is the mean battery and inductorcurrent, as dened in Fig. 6.1, and Rin is the internal battery resistance. Thenominal voltage VB will be adjusted in the time domain simulation over aperiod of a few seconds to ensure the voltage control loop responds correctlyto a change in battery voltage.
For frequency domain analysis the battery was modelled as a large capacitorin series with a resistor. From rst principles it is known that electrical chargeis equal to voltage multiplied by capacitance (Q = CV ) and one ampere-hourequals 3600 coulombs of charge. The value of the capacitor Cb representingthe battery is thus approximated by
CB =Q
VB
≈ 3600× 40
12× 20= 600 F. (6.53)
The internal resistance of the twenty batteries Rin is approximated as 5 timesthe resistance of the batteries when fully charged, thus 0.95 Ω.
The mean dc bus current Idc is related to the mean battery and inductorcurrent IL by
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 114
D1 =IdcIL
= 0.6857, (6.54)
as given in (3.3) and repeated here for convenience. The transfer function forthe battery voltage with respect to the dc bus current is given by
Fp(s) =Vbat
Idc=
1
D1
(1
sCB
+Rin
)=
1 + sCBRin
sD1CB
. (6.55)
The battery voltage is measured by means of a fully dierential amplier asshown in Fig. 6.25 to provide the feedback voltage vfbv. Two capacitors wereadded to the dierential amplier to form a low-pass lter that is used toreduce the noise on the measured signal.
R1R2 C1
−
+
R1−vbat
R2
C1
vfbv+vbat
Figure 6.25: Battery voltage feedback circuit.
At a battery reference voltage Vbat∗ of 5 V the measured battery voltage should
rise gradually to 272 V. The amplier thus acts as an attenuator with a gainfactor of 5
272. The values chosen for R1 and R2 are 446 kΩ and 8.2 kΩ re-
spectively. The capacitor C1 is chosen as a 100 nF ceramic capacitor. The lowpass lter's corner frequency is calculated using (6.47) as 194 Hz.
The transfer function for a low pass lter was already derived in (6.46).With reference to Fig. 6.24, the transfer function of the voltage feedback circuitis denoted J(s) and thus J(s) is given by
J(s) =R2
R1
(1
1 + sC1R2
). (6.56)
The voltage control loop compensation amplier is denoted Fc(s). The blockdiagram shown in Fig. 6.24 can be reduced to the block diagram shown inFig. 6.26 where the inner current control loop is replaced by the closed-looptransfer function Gcl(s) derived in (6.50).
With reference to Fig. 6.26, the open-loop transfer function of the voltagecontrol loop Fol(s) is
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 115
Vbat∗ Σ
+
−Vfbv
Fc(s) Idccl(s) Fp(s) Vbat
J(s)
Vε
Figure 6.26: Reduced voltage and current control loop block diagram.
Fol(s) =Vε(s)
Vfbv(s)= Fc(s)Gcl(s)Fp(s)J(s), (6.57)
and the closed-loop transfer function Fcl(s) is
Fcl(s) =Fc(s)Gcl(s)Fp(s)
1 + Fc(s)Gcl(s)Fp(s)J(s). (6.58)
The bode plot of the open-loop response Fcl(s) without the transfer functionof the compensation amplier present is shown in Fig. 6.27. The frequencyresponse shows a unity crossover frequency of approximately 10−5 Hz whichis very low. The desired speed response of the voltage control loop is in factvery slow since the battery voltage will change over a period of minutes tohours. A slight gain can however be added to speed up the voltage controlloop frequency response to approximately 0.01 Hz.
−40
−20
0
20
Magnitude
(dB)
10−6 10−5 10−4 10−3 10−2 10−1 100 101−135
−90
−45
0
Frequency (Hz)
Phase
(deg)
Figure 6.27: Bode plot of the open-loop frequency response of the voltagecontrol loop without compensation.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 116
The ability of the batteries to store charge allows it to act as a large capacitorwhich already provides an integrator as shown in Fig. 6.27, thus the voltagecontrol loop has theoretical innite gain at dc. A low-pass lter with gainand a very low corner frequency is selected as the compensator to ensure thecurrent set-point Vseti is never changed abruptly. The circuit looks identical toFig. 6.25 with R1, R2 and C1 equal to 22 kΩ , 470 kΩ and 100 µF, respectively.The open- and closed-loop response of the voltage control loop are shown inFig. 6.28.
−100
−50
0
50
100
Magnitude
(dB)
10−6 10−5 10−4 10−3 10−2 10−1 100 101−225
−180
−135
−90
−45
0
Frequency (Hz)
Phase
(deg)
Fol(s)Fcl(s)
Figure 6.28: Bode plot of the open- and closed-loop frequency response of thevoltage control loop.
From Fig. 6.28 the gain and phase margins are given by 75.7 dB and 140° re-spectively. The unity crossover frequency increased to 0.0025 Hz. The closed-loop response has a dc gain of 34.71 dB corresponding to a loop sensitivityof 54.33 V of battery voltage per volt of excitation applied to the voltage set-point Vsetv. At a maximum set-point voltage of 5 V the corresponding batteryvoltage will rise to 272 V, as designed for. It should be noted, however, thatregardless of the battery voltage measured by the circuit of Fig. 6.25, the bat-tery charging current will never exceed a mean value of 10 A due to the actionof the dc-dc converter.
The frequency domain analysis thus indicates that both the current andvoltage control loops are stable. The complete circuit schematic for the invertercontrol board is given in Appendix A.2.2.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 6. INVERTER DESIGN AND IMPLEMENTATION 117
6.3 Summary
The detailed operation of the bidirectional inverter circuit and asynchronousgenerator were presented in this chapter. The switching and conduction lossesfor the chosen IGBT module were calculated numerically. The design and im-plementation of the inverter circuit's safety features were presented, amongstwhich are over-temperature, overcurrent and overvoltage protection. A de-tailed design of the inverter control circuit was also presented where two con-trol loops, an inner and outer control loop, were used to regulate the dc buscurrent and battery voltage, respectively. Frequency domain analysis showedboth control loops to be stable. The time domain equations describing theinverter operation were derived and the simulations are presented in the nextchapter, along with the measured results from the manufactured prototypecircuit.
Stellenbosch University https://scholar.sun.ac.za
Chapter 7
Inverter Simulation and System
Integration Test Results
This chapter contains the simulated and measured test results of the integrateddc-dc converter and inverter circuit. The inverter circuit, designed in the pre-vious chapter, is modelled and simulated in the time domain along with thegenerator detailed in Chapter 5. The simulated circuit is compared to the man-ufactured prototype circuit and the results are discussed. The implementedinverter safety features were tested and observations are presented.
The test set-up for the integration of the inverter and converter circuitsis also described in this chapter. The measured battery voltage is comparedto the measured dc bus current to validate the system integration and todemonstrate the practical operation of the completed circuit.
All time domain simulations were implemented in MATLAB where therelevant dierential equations were once again solved using the iterative Eulermethod.
7.1 System Integration
The operation of the dc-dc converter prototype circuit (with and without aload connected to the dc bus) was already presented and discussed in Chapter4. The inverter circuit was designed to operate from the regulated positiveand negative dc bus rails provided by the dc-dc converter. The inverter con-trol circuitry modulates the voltages applied to the asynchronous generatorstator terminals and in return the generator supplies current to the dc busthrough the inverter circuit at a constant negative slip of -5.47 %. The dc-dcconverter's batteries are responsible for sinking the dc bus current owing intothe converter's dc bus terminals.
118
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 7. INVERTER SIMULATION AND SYSTEM INTEGRATION
TEST RESULTS 119
7.1.1 Test Set-Up
The test set-up for the complete system is shown in Fig. 7.1. The prototypeinverter circuit is located on the left-hand side of the table shown in Fig. 7.1,while the prototype dc-dc converter circuit is placed on the right-hand side ofthe table. The 60 Ω , 2.1 kW load is placed between the two prototype circuits.
Figure 7.1: Test set-up for generator and battery interface circuit.
The inverter circuit is constructed in a similar manner to the dc-dc convertercircuit. The control board is mounted on top of the power board connected tothe heat sink. A small 12 V fan is also mounted to the side of the heat sinkand control board. The detailed inverter PCBs are given in Appendix B.2.Both the inverter control and power boards are powered from a 12 V benchpower supply through a DKE10A-15 converter.
The two induction machines (one to drive the other) are mounted on topof the trolley as shown in Fig. 7.1. The VSD controlling the induction motoroperation is located to the right of the two machines.
The measurements for the system operation were obtained using an Agi-lent Technologies oscilloscope, LEM current probe, multimeter and Avometer,depending on the type of measurement required.
7.1.2 Start-Up Procedure
The inverter control circuit is switched on rst (before the dc-dc convertercircuit). The dsPIC ensures the inverter circuit will remain idle (not switching)
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 7. INVERTER SIMULATION AND SYSTEM INTEGRATION
TEST RESULTS 120
until it receives at least two magnetic pick-up interrupts to prevent the inverterfrom loading the dc-bus before it is fully charged.
Next, the dc-dc converter is switched on and the dc bus is charged to355 V as presented in Chapter 4. The VSD system driving the inductionmachine used to simulated airow through a turbine is switched on last. Thegenerator shaft starts to turn, generating magnetic pick-up interrupts at oneof the input channels of the dsPIC. The inverter control circuitry responds bygenerating PWM signals to turn on the IGBTs of the inverter, thus poweringthe generator.
7.2 PWM Generation and Gate Signals
The generation of the six PWM signals to drive each of the IGBTs were de-scribed and simulated in the previous chapter. The measured PWM generationresults are presented in this section as well as the measured IGBT gate signals.
7.2.1 Reference Waveforms
The modulator reference waveforms are generated by the dsPIC DSP and thesinusoidal output waveforms are provided by a DAC, connected to the dsPICas detailed in Section 6.2.2. As soon as the inverter control circuit is switchedon, the battery voltage is measured and the dc bus current set-point is adjustedas a function of the battery voltage.
The nominal battery voltage is 12 V and when fully charged it can reach anominal value of 12.6 V, hence the initial measured battery voltage will alwaysremain below 260 V (13 V per battery). The control circuit will produce amaximum current set-point (5 V), thus demanding 5.7 A of dc bus current, ifthe measured voltage is below 260 V due to the gain provided by the voltagecontrol loop's compensation amplier.
Once the magnetic pick-up senses the generator shaft frequency has ex-ceeded 7 Hz, the orange warning LED switches o and the amplitudes of thesinusoidal modulator reference waveforms slowly increase as shown in Fig. 7.2.The maximum amplitude of a waveform is dependent on the speed of thegenerator shaft and the magnitude of the control signal vcs, generated by thecurrent control loop's compensation amplier.
At a shaft frequency of 25 Hz and a maximum vcs control signal magnitudeof 5 V, the maximum allowed amplitude Vc of the sinusoids are calculatedusing (6.39) as 1.92 V (3.84 V peak-to-peak), due to the scaling factor Kc
dened in Section 6.2.2. Recall, by scaling the amplitudes of the modulatorreference waveforms with respect to the generator shaft speed, the amplitudesof the fundamental stator voltages are also scaled and thus the generator isoperated below or at rated ux. Saturation currents within the generator arethus avoided. The measured modulator sinusoids are shown in Fig. 7.3.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 7. INVERTER SIMULATION AND SYSTEM INTEGRATION
TEST RESULTS 121
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5−1
−0.5
0
0.5
1
Time (s)
v ref
(V)
vruvrvvrv
Figure 7.2: Measured modulator reference waveforms at an increasing genera-tor shaft speed.
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1−2
−1
0
1
2
Time (s)
v ref
(V)
vruvrvvrw
Figure 7.3: Measured modulator reference waveforms at a shaft frequency of25 Hz.
The measured peak-to-peak voltages of the three waveforms shown in Fig. 7.3are 3.83 V, 3.78 V and 3.80 V, respectively, closely matching the designedvoltage of 3.84 V. The amplitudes of the three sinusoids are close to, but notidentical to each other. This indicates a small margin of error introduced bythe op-amp and resistor circuit that is used to remove the dc oset and add again of 2 to each of the DAC outputs. The gain of each sinusoid is thus slightlygreater or lower than two, resulting in the small voltage error. The resultingphase currents will thus also vary ever so slightly in error. The small error incurrent and voltage will not stop the generator from operating as required.
The measured frequency of each of the waveforms shown in Fig. 7.3 is23.7 Hz corresponding to a slip of -5.47 %, thus matching the designed slip of-5.47 %.
Similarly at a shaft frequency of 50 Hz, the maximum allowed sinusoidalvoltage is 7.69 V peak-to-peak. The measured sinusoidal modulator reference
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 7. INVERTER SIMULATION AND SYSTEM INTEGRATION
TEST RESULTS 122
waveforms at a shaft frequency of 50 Hz are shown in Fig. 7.4. The peak-to-peak voltages are measured as 7.65 V, 7.58 V and 7.6 V, respectively. Thevoltages are close to the designed maximum voltage of 7.69 V. The referencewaveform frequencies are measured as 47.4 Hz corresponding to the designedslip of -5.47 %.
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
·10−2
−4
−2
0
2
4
Time (s)
v ref
(V)
vruvrvvrw
Figure 7.4: Measured modulator reference waveforms at a shaft frequency of50 Hz.
The dsPIC and DAC are seen to function as designed and deliver the correctmodulator reference voltage waveforms at various measured generator shaftspeeds. From Fig. 7.3 and Fig. 7.4 it can be seen that the digitally generatedsinusoids look very much like analog signals, indicating the resolution chosenfor the digital sinusoids is adequate.
Lastly, when the generator shaft frequency exceeds the maximum frequencyof 75 Hz, the orange warning LED switches on and the reference waveformsgradually reduce until reaching peak-to-peak values of zero.
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2
−0.4
−0.2
0
0.2
0.4
Time (s)
v ref
(V)
vruvrvvrw
Figure 7.5: Measured modulator reference waveforms after the operating shaftspeed range was exceeded.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 7. INVERTER SIMULATION AND SYSTEM INTEGRATION
TEST RESULTS 123
Similarly when the shaft frequency was slightly reduced to a value below 75 Hz,the reference waveforms were seen to gradually return to an amplitude of 5 Vas shown in Fig. 7.5. The warning LED switched o and normal switchingoperation continued.
7.2.2 Triangular Carrier Waveform
The triangular carrier waveform used for comparison with the sinusoidal refer-ence waveforms, was designed to have a frequency of 20 kHz and a peak-to-peakvoltage of 12 V. The measured carrier waveform is shown in Fig. 7.6.
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
·10−4
−6
−3
0
3
6
Time (s)
v r(V
)
Figure 7.6: Measured triangular carrier waveform.
The measured frequency is 20 kHz and the peak-to-peak voltage is 11.9 V. Theamplitude of the triangular wave is thus 5.95 V. The measured result comparesextremely well to the designed carrier waveform detailed in Section 6.1.
7.2.3 Switching Signals
The measured PWM switching signals generated by the inverter control cir-cuitry for the IGBTs in one half-bridge are shown in Fig. 7.7.
The switching signal of S1 is complementary to S2, as designed. At theinstance of the measurement the duty cycle for S1 was 68.4 % and 26.6 % forS2. The implementation of the dead-time can clearly be seen in Fig. 7.7 dueto the small time period during which both switches are o. The dead-timewas set to 1.2 µs and was chosen as such to avoid any IGBT shoot-throughcurrents.
Similar PWM signals were generated for the other two half-bridges contain-ing switches S3 through S6. The duty cycles vary sinusoidally and in phasewith the modulator reference voltage waveforms, as expected.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 7. INVERTER SIMULATION AND SYSTEM INTEGRATION
TEST RESULTS 124
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
·10−4
0
2.5
5
Time (s)
SwitchingSingals(V
)S1S2
Figure 7.7: Measured PWM switching signals for one inverter half-bridge.
At a maximum reference waveform amplitude of 5 V and a 5.95 V ramp waveamplitude, the maximum modulation index is 0.84. From (6.8) the maximumduty cycle is thus limited to 92 %.
7.2.4 IGBT Gate Signal
The push-pull oscillator circuit used to power the gate driver IC that communi-cates the switching signal to the gate of the IGBT, was determined to oscillateat a frequency of 309 kHz. The optocoupler supply voltage was measured as16.4 V, close to the designed voltage of 16.5 V.
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
·10−4
0
2
4
PWM
Signal(V
)
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
·10−4
0
5
10
15
Time (s)
GateSignal(V
)
Figure 7.8: Measured controller and gate PWM switching signals.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 7. INVERTER SIMULATION AND SYSTEM INTEGRATION
TEST RESULTS 125
At a generator shaft frequency below 7 Hz or above 75 Hz, the duty cycleof all the switches change to a constant value of approximately 50 %, sincethe reference waveforms have reduced to a constant zero volt signal. Themeasured PWM switching signal for S1, as well as the signal at the gate ofthe corresponding IGBT, are shown in Fig. 7.8 at a generator shaft frequencybelow 7 Hz.
The measured duty cycles for both the PWM control signal and the gatesignal were 47.6 %. Due to the dead-time the duty cycles are slightly lowerthan the ideal 50 %. The signal at the gate of the IGBT has a maximumvoltage of 16.3 V corresponding to the gate driver IC supply voltage of 16.4 V.
7.3 Current Simulation and Test Results
The PWM gate drive signals produce PWM voltages at the three pole points ofthe inverter. These voltages are also applied to the generator stator terminalswith their fundamental component at a lower frequency than the measuredshaft frequency. Due to the constant -5.47 % slip operation, current is inducedin the rotor of the generator and current will ow out of the generator into thedc bus.
The outer voltage control loop will demand a current of no greater than5.7 A when a 2 kW load is connected between the dc bus rails, thus no powerows into the batteries to allow them to charge. As soon as the load is removedor its resistance increases, current ows into the batteries and the batteryvoltage rises to above 260 V. In return less current is demanded by the voltagecontrol loop and the battery voltage progressively settle. The batteries arefurther charged slowly and the mean dc bus current gradually decreases, underthe control of the two control loops.
The phase currents and dc bus current were simulated in MATLAB at themaximum allowed mean dc bus current as well as at a lower dc bus current,such as while the batteries are charging. The results are presented in thissection and the time domain analysis showing the stability of the system is alsodiscussed. The corresponding measurements of these currents are presentedand compared with the simulated results.
Without a 2 kW load connected between the dc bus rails, the dc bus currentis shown to ow into the converter to charge the batteries.
7.3.1 Inverter Phase Currents
Initially no load is connected between the dc bus rails. The battery voltagebegins to rise as soon as the bus current rises, since the batteries are beingcharged. The inverter phase current measurements were taken once the batteryvoltage started to settle at a voltage of 266.1 V and a mean dc bus current
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 7. INVERTER SIMULATION AND SYSTEM INTEGRATION
TEST RESULTS 126
of 3.3 A. For comparison purposes the simulation was repeated under theseconditions.
The simulated phase currents are shown in Fig. 7.9 at a set-point voltagecorresponding to a mean dc bus current of 3.3 A and a generator shaft fre-quency of 50 Hz. The corresponding measured phase currents are shown inFig. 7.10.
0 1 2 3 4 5 6 7 8
·10−2
−10
0
10
Time (s)
i p(A
) iuiviw
Figure 7.9: Simulated inverter phase currents at a 3.3 A dc bus current set-point.
0 1 2 3 4 5 6 7 8
·10−2
−10
0
10
Time (s)
i p(A
) iuiviw
Figure 7.10: Measured inverter phase currents at a measured dc bus currentof 3.3 A.
The simulated phase currents each have an rms value of 7.5 A and a frequencyof 47.4 Hz. The rms values of the three measured phase currents are 8.99 A,8.89 A and 8.95 A, respectively. As indicated by the measured modulatorreference waveforms, the phase currents will also vary slightly in magnitude.The frequency of the measured phase currents are 47.5 Hz each. From Fig. 7.10it can be seen that the measured phase currents are not pure sinusoids. This
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 7. INVERTER SIMULATION AND SYSTEM INTEGRATION
TEST RESULTS 127
is due to the fact that the inductance of the generator is, in reality, a functionof the currents passed through the generator windings due to the non-linearmagnetic permeability of the rotor and stator core material. As a generalrule the permeability of the ferromagnetic core material is seen to decline asthe generator line currents increase. We thus observe the distorted sinusoidalphase currents in response to the application of sinusoidal fundamental phasevoltages to the generator stator terminals.
Both the simulated and measured phase currents indicate a constant gener-ator slip of approximately -5.47 %, as designed for. The rms value of the simu-lated and measured phase currents does however dier by approximately 1.5 A,indicating the equivalent generator impedance is lower than the impedance ap-proximated by the equivalent induction machine model derived in Chapter 5.A margin of error was expected since the parameters of induction machineschange with a change in temperature and operating frequency (slip) [42], [43].
The simulated and measured phase current ripple are shown in Fig. 7.11and Fig. 7.12, respectively. The simulated current ripple has a peak-to-peakvalue of 0.22 A. The measured current ripple has a peak-to-peak value of0.356 A. The measured current ripple is larger than the simulated currentripple conrming the statement that the true machine inductance is lowerthan the values calculated in Chapter 5 from the approximate machine model.
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
·10−4
8.2
8.4
8.6
8.8
9
Time (s)
i u(A
)
Figure 7.11: Simulated inverter phase current ripple.
The frequency of both the simulated and measured current ripple is 20 kHzcorresponding to the 20 kHz switching frequency of the IGBTs. The measuredripple current is shown to have a signicant amount of high-frequency noisesuperimposed on the signal. This noise is not real and is simply due to thenoise picked up by the LEM current probe while taking the measurement.The two induction machines, VSD, dc-dc converter and inverter all operatingat dierent frequencies and thus a signicant amount of noise is produced,which is easily picked up by measuring equipment. One of the most signicantsources of radiated noise was determined to be the commercial VSD used tocontrol the motor coupled to the shaft of the generator.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 7. INVERTER SIMULATION AND SYSTEM INTEGRATION
TEST RESULTS 128
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
·10−4
7.8
8
8.2
8.4
Time (s)
i u(A
)
Figure 7.12: Measured inverter phase current ripple.
The simulated modulator reference waveform for switches S1 and S2 is shownin Fig. 7.13, along with the corresponding phase current iu at a mean dcbus current of 3.3 A. The modulator reference voltages are in phase with thefundamental component of the inverter phase voltages, if the modulator circuitis considered ideal. The modulator reference waveforms can thus be used todetermine the phase angle between the phase currents and voltages. As seenfrom Fig. 7.13 the current is leading the modulator reference voltage and thephase angle was determined from the simulation as 35.82°.
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
·10−2
−4
−2
0
2
4
Time (s)
v ru(V
)
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
·10−2
−12
−6
0
6
12
i u(A
)
Figure 7.13: Simulated modulator reference voltage and phase current for oneinverter half-bridge.
The measured phase current iu and the modulator reference waveform forswitches S1 and S2 are shown in Fig. 7.14. Since the inverter phase volt-age waveforms are modulated, their fundamental waveforms are not easilymeasured. Although the modulator circuit itself is not ideal, the phase shiftbetween the fundamental phase voltages and currents are approximately equalto the phase shift between the measured modulator reference waveforms and
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 7. INVERTER SIMULATION AND SYSTEM INTEGRATION
TEST RESULTS 129
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
·10−2
−3
−2
−1
0
1
2
3
Time (s)
v ru(V
)
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
·10−2
−15
−10
−5
0
5
10
15
i u(A
)
Figure 7.14: Measured modulator reference voltage and phase current for oneinverter half-bridge.
the phase current waveforms. From Fig. 7.14 the current is shown to lead themodulator voltage by 35.88°.
The amplitude of the modulator reference waveform is seen to dier for thesimulated and measured case shown in Fig. 7.13 and Fig. 7.14, respectively.This is because the dc bus current is regulated to 3.3 A and, as already shown,the measured phase currents are larger than the simulated phase currents.Thus, a lower modulator reference voltage is required to produce 3.3 A ofcurrent owing in the dc bus for the measured case.
Once the 2.1 kW load is connected, the battery voltage reduces to below260 V since the batteries as well as the generator is used to supply power tothe load. The set-point generated by the voltage control loop increases to 5 V(5.7 A). At a shaft frequency of 34 Hz the corresponding measured modulatorvoltage and phase current are shown in Fig. 7.15.
From Fig. 7.15 the modulator reference voltage is shown to peak at 2.5 Vand the measured frequency is 32.1 Hz. The maximum allowed reference volt-age is 2.6 V at a shaft frequency of 34 Hz. The corresponding phase current hasan rms value of 11.99 A. The power angle is measured as 36.9°. The phase shiftis thus seen to remain constant with a change in frequency when compared tothe power angle of 35.88° at a shaft frequency of 50 Hz.
The amplitudes of the phase currents will remain reasonable constant overa wide generator shaft speed range since both the fundamental stator volt-ages and generator impedance change with speed. From measurements it wasseen that at low shaft speeds (below approximately 20 Hz) this voltage andimpedance relationship does not remain constant and the amplitude of thephase currents will change signicantly as a function of the shaft speed.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 7. INVERTER SIMULATION AND SYSTEM INTEGRATION
TEST RESULTS 130
0 1 2 3 4 5 6 7 8 9 10
·10−2
−3
−2
−1
0
1
2
3
Time (s)
v ru(V
)
−18
−12
−6
0
6
12
18
i u(A
)
Figure 7.15: Measured modulator reference voltage and phase current at ashaft frequency of 34 Hz.
7.3.2 Bus Current
The time domain analysis of the dc bus current during start-up is shown inFig. 7.16 for a maximum current loop set-point of 5 V. The simulation wasconducted at a shaft frequency of 50 Hz. The amplitudes of the modulatorreference waveforms were gradually increased as discussed in Section 7.2.1. Thedc bus current is shown to increase with an increase in the applied inverterphase voltages, as expected. After 2.42 s, the dc bus current is shown to reachthe desired mean dc bus current value of 5.7 A. A small current overshoot of0.4 A follows with the dc bus current thus peaking at 6.1 A. After a total of2.6 s the dc bus current is shown to have settled to the desired value of 5.7 A.
0 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 2.7 30
2
4
6
8
Time (s)
I dc(A
)
MeasuredSet-point
Figure 7.16: Simulated dc bus current during generator start-up.
The time domain analysis shows the dc bus current will settle at the desiredcurrent set-point without a large current overshoot (less than 7.1 %). Thesystem is seen to be stable and able to supply the current required by theouter control loop.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 7. INVERTER SIMULATION AND SYSTEM INTEGRATION
TEST RESULTS 131
The VSD controlling the motor (used to turn the generator shaft) was set toaccelerate the frequency of the stator voltages applied to the motor over aperiod of 10 s to avoid saturation currents in the motor. The slow accelerationof the generator shaft combined with the gradual increase of the amplitudes ofthe modulator reference waveforms cause the measured dc bus current to rampup slowly and with a maximum current overshoot of approximately 0.1 A topeak at 5.8 A. The motor is decelerated in a similar manner using the VSD.
In practice the dc bus current set-point will never experience a step responseafter start-up, since the outer voltage control loop is set to react very slowly(due to the batteries charging slowly) and hence the set-point generated byvoltage control loop will change gradually, allowing the dc bus current to follow.This response is detailed in the next section.
The measured dc bus current under steady state conditions is shown inFig. 7.17 at a maximum set-point of 5 V, corresponding to 5.7 A, and a gen-erator shaft frequency of 50 Hz. The mean value is measured as 5.66 A whichclosely match the desired value of 5.7 A. The error is due to a small inaccu-racy in the current sensing circuit. With a regulated 355 V dc bus, the powerrequirement of 2 kW is still met. The inverter and generator circuit is thusable to supply 2 kW of power to the dc bus.
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
·10−4
4
5
6
7
8
Time (s)
i dc(A
)
Figure 7.17: Measured dc bus current at the maximum set-point current.
The dc bus current is shown to have a fundamental frequency of 20 kHz with a40 kHz component also present. The inverter is switching at 20 kHz while thedc-dc converter is switching at 40 kHz accounting for the additional frequencycomponent. The ripple in the dc bus current is seen to have a peak-to-peakvalue of 2.75 A. An increase or decrease in the amplitudes of the three in-verter phase currents will result in a corresponding increase or decrease in thepeak-to-peak dc bus current ripple.
Due to the diculty in measuring the fundamental component of the in-verter phase voltages, the inverter eciency was not determined. A poweranalyser of some sort is required to accurately measure the eciency of such
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 7. INVERTER SIMULATION AND SYSTEM INTEGRATION
TEST RESULTS 132
a circuit. The lab facilities available for this project did unfortunately notcontain any power analysers.
The dc bus current was also measured whilst charging the batteries. Themeasurement was taken along with the already presented phase current mea-surements at a mean dc bus current of 3.3 A as shown in Fig 7.18. The ripplecurrent has a peak-to-peak magnitude of 1.94 A, which is lower than that ofthe previous dc bus current measurement, as expected.
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
·10−4
2
3
4
5
Time (s)
i dc(A
)
Figure 7.18: Measured dc bus current whilst charging the batteries.
The dc bus current control loop is shown to be stable and able to regulate thecurrent owing into the dc bus eectively. The power transfer requirement of2 kW is also shown to be met by the prototype circuit.
7.3.3 Battery Current
The measured current owing into the batteries at a mean dc bus current ofapproximately 3.3 A is shown in Fig. 7.19. The battery current is shown tohave a mean value of 4.8 A and a peak-to-peak current ripple of 2.04 A at afrequency of 40 kHz.
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
·10−4
3
4
5
6
Time (s)
i L(A
)
Figure 7.19: Measured battery current whilst charging the batteries.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 7. INVERTER SIMULATION AND SYSTEM INTEGRATION
TEST RESULTS 133
The inductor current ripple is seen to be lower than the designed value of2.25 A even if the inductor is experiencing greater magnetisation which wouldincrease the current ripple. This is simply due to the battery voltage thatincreased to 266.1 V from the nominal value of 240 V (12 V per battery). Thevoltage across the inductor is thus less than before when switches Q1 and Q4
are closed, resulting in a smaller current ripple. The duty cycle of switchesQ1 and Q4 are also seen to have changed slightly to 70 % as opposed to theprevious value of 68.57 %.
The dc-dc converter is thus shown to be bidirectional since it can accom-modate current ow to as well as from the dc bus. The results showing powerow from the batteries to the dc bus were already presented in Chapter 4.
7.4 Bus Current and Battery Voltage
Relationship
The results of the outer voltage control loop used to regulate the batteryvoltage are presented in this section. As the battery voltage increases, the dcbus current should decrease until such point where the current supplied bythe generator is just enough to keep the batteries at their oating use voltageof 272 V to 276 V. Both the simulated and measured response of the dc buscurrent with respect to the state of charge of the batteries are presented anddescribed.
7.4.1 Simulated Response
The time domain simulation only deals with the electrical properties of thebatteries, thus the batteries were modelled as a constant voltage source with aninternal resistance, as detailed in Chapter 6. In order to observe the change indc bus current as a function of the battery voltage in the simulation, the batteryvoltage (modelled as the constant voltage source) was increased linearly overa time period of 7 s.
To provide a better approximation of the battery's response to sinkingcurrent, the battery voltage was simulated to increase from 240 V (nominally12 V per battery) to 265 V over a period of 3 s and then increase slowly to272 V over the next 4 s. The results are shown in Fig. 7.20.
Initially the set-point current corresponds to 5.7 A as indicated in Fig. 7.20.The dc bus current is ramped up as explained in Section 7.3.2. At the sametime the battery voltage is seen to rise gradually. After 3 s the battery voltageexceeds 260 V and the current set-point is shown to decline. The simulated dcbus current is seen to respond to the change in the current set-point. Whenthe battery voltage reaches 265 V the batteries start charging at a lower rate.The current set-point is seen to also decline at this lower rate, with the dc buscurrent following its response. Once the battery voltage reaches the desired
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 7. INVERTER SIMULATION AND SYSTEM INTEGRATION
TEST RESULTS 134
0 1 2 3 4 5 6 7−202468
I dc(A
)
MeasuredSet-point
0 1 2 3 4 5 6 7240
250
260
270
280
Time (s)
Vba
t(V
)
Figure 7.20: Simulated mean dc bus current and battery voltage.
value of 272 V, the current set-point reaches zero, thus demanding zero current.The dc bus current is also seen to reduce to the desired value of zero.
The time domain simulation predicts that both the voltage and currentcontrol loops of the inverter are stable.
7.4.2 Measured Response
Due to the time duration required for the batteries to charge, the measurementcould not be obtained as a single measurement using an oscilloscope. The meanbattery voltage was measured at regular intervals with a multimeter and themean dc bus current was measured with an Avometer as well as as with a LEMcurrent probe, to conrm the measurement. As soon as the generator startedto deliver power to the dc bus (a load was not connected between the dc busrails), the battery voltage was seen to increase rapidly to above 260 V. Theresults were taken from a battery voltage of 260 V and are shown in Fig. 7.21.The measurements were taken with the batteries drained beforehand to a valueof approximately 11.92 V per battery.
The results given in Fig. 7.21 show the mean dc bus current does in factdecrease as the mean voltage of the batteries increases, as desired. Initiallythe dc bus current is seen to peak at 5.8 A while the dc bus voltage increasesrapidly. The gain used in the voltage compensation amplier is sucient todemand 5.7 A of current up to a measured battery voltage of 260 V. Thereafterthe demanded current is decreased as shown in Fig. 7.21.
The current is seen to drop at approximately the designed rate until thebattery voltage reaches 265 V. Thereafter the battery voltage starts to settle
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 7. INVERTER SIMULATION AND SYSTEM INTEGRATION
TEST RESULTS 135
260 261 262 263 264 265 266 267 268 269 270 271 2721
2
3
4
5
6
Vbat (V)
I dc(A
)
Figure 7.21: Measured mean dc bus current vs battery voltage.
and requires more current than predicted to continue charging the batteries.From Fig. 7.21 the voltage is seen to never reach exactly 272 V, it remainssomewhere between 271 V and 272 V. Similarly the dc bus current is seen toreach a minimum value of approximately 2.1 A. The battery voltage and dcbus current thus reach equilibrium at these values. The equilibrium values canchange with a change in temperature, since the storage capacity of batteriesare temperature dependent.
The simulation predicted that the dc bus current will drop to zero at abattery voltage of 272 V. In practice the battery voltage will never actuallyreach 272 V but will remain very close to 272 V as long as the necessary currentis supplied to keep it there, thus trickle charging the batteries. The internalresistance of the batteries as well as the eciency of the converter causes thedc bus current to remain above zero in order to keep the batteries at theiroptimal oating use voltage of 272 V.
The last few measurements were taken approximately 30 minutes from eachother. In total the batteries where charged for approximately 8 hours to reacha voltage of 271.5 V. If the batteries are charged even longer, the internalbattery resistance will decrease even further and the required dc bus currentwill decrease slightly.
Overall the inverter control circuit is shown to be stable and provides thedesired controller responses. The dc-dc converter prototype circuit was suc-cessfully integrated with the inverter prototype circuit. The mean dc buscurrent provided by the generator is successfully controlled as a function ofthe measured battery voltage, with the batteries connected to the dc-dc con-verter. The overall system is also shown to be stable under various operatingconditions (with or without a load connected).
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 7. INVERTER SIMULATION AND SYSTEM INTEGRATION
TEST RESULTS 136
7.5 Safety Features Testing
The safety features detailed in Chapter 6 were implemented in the inverterprototype circuit. The soft turn-on and turn-o feature of the amplitudes ofthe modulator reference waveforms when the generator shaft speed exceeds theoperating speed range (above 75 Hz or below 7 Hz), were already presentedand discussed in Section 7.2.1. The correct LED also turned on to visuallyindicate why the generator cannot supply current to the dc bus.
A green LED was also mounted on the PCB containing the dsPIC. TheLED was programmed to turn on and o at 0.5 s intervals. If the LED stopsashing it indicates the dsPIC is either not receiving power or has fallen intoa continuous loop where the low priority interrupt to turn the LED on and ocan never be reached. This provides a visual method to indicate if the controlcircuit is switched on and to see if the dsPIC itself is experiencing any majorproblems. The other safety features were also tested and is discussed in thissection.
7.5.1 Over-Temperature Protection
The over-temperature detection was tested using an external input to simulatethe temperature output from the thermistor circuit. At a voltage of 3.35 V(the designed threshold voltage) the over-temperature sensor tripped and thered LED was switched on by the dsPIC. At the same time all switching actionwas terminated as desired. The mean dc bus current was seen to drop andafter a few seconds it reached zero. The switching action only resumed afterthe dsPIC was reset. Similarly the red LED only turned o once the dsPICcircuit was restarted.
7.5.2 Overcurrent Protection
The overcurrent protection is supplied by the gate driver IC (ACPL-332J)which detects the Vce voltage of the IGBTs. The larger the currents throughthe collector of the IGBT, the higher the measured Vce voltage. The IGBTmodule was tested before using it in the fully functional inverter circuit to seethat the fault output from the gate driver IC tripped when large currents wereowing through the IGBT collector. The fault output was observed to changestate thus indicating it functioned correctly.
Initially the dead-time for the IGBT switches was set to 800 ns. Theinverter circuit was able to function as desired but after a while, when theIGBT module started to heat up, shoot-through currents were detected dueto the increased turn-on and turn-o times of the switches. The overcurrentprotection provided by the gate driver IC sent a fault signal to the dsPIC whileactivating the soft switch turn-o feature built into the chip. The orangeLED was turned on by the dsPIC. Similarly as with the over-temperature
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 7. INVERTER SIMULATION AND SYSTEM INTEGRATION
TEST RESULTS 137
fault condition, the switching action was terminated. The control circuit wasrestarted to allow the switching action of the IGBTs to resume and to allowthe LED to turn o. The dead-time was subsequently increased to 1.2 µs.
7.5.3 Overvoltage Protection
The overvoltage protection was tested using an external input to simulatethe dc bus voltage feedback. As soon as the simulated scaled dc bus voltageexceeded 5 V, the output of the comparator was shown to change state, thusthe overvoltage protection circuitry was triggered. The IGBT switching actionwas once again terminated as with the other protection features. The reversebiased diodes connected across the collector and emitter of each IGBT stillprovide a path for the generator current to ow, but as soon as the switchingaction of the IGBTs are terminated the voltage applied to the generator statorterminals are also terminated. With no voltage applied to the generator statorterminals the induced current immediately reduces and after a few seconds(less than 5 s) the mean dc bus current reaches zero.
The dc bus voltage will thus not rise to a high enough voltage to damageany of the components used in the inverter and converter prototype circuits.All of the high voltage components are rated to operate continuously at avoltage of at least 400 V. The IGBT switching action is resumed after thedsPIC is restarted.
7.6 Summary
The performance of both the inverter and dc-dc converter prototype circuitswere presented in this chapter with the main focus on the integration betweenthe two circuits. The test set-up and start-up procedure to test the completesystem was described in detail. The generation of the gate-drive signals used tocontrol the switching action of the IGBTs was presented using measurements.The inverter phase currents and dc bus current were presented in order todemonstrate the correct operation of the designed inverter current loop whencompared to the simulated response. The measured battery current used tocharge the batteries was given indicating that the dc-dc converter is in factbidirectional. The test results showing the relationship between the mean dcbus current and measured battery voltage were presented and the integrationof the dc-dc converter and inverter circuits were proven to be successful. Theinverter protection circuitry was shown to function properly and provide thenecessary protection for which it was designed.
Stellenbosch University https://scholar.sun.ac.za
Chapter 8
Conclusions
A summary of the work presented in this thesis along with the results are givenin this chapter. Recommendations with regard to the research conducted arediscussed and possible future work is considered.
8.1 Thesis Conculsions
A circuit that demonstrates the interface between an asynchronous generator,a battery of electrochemical cells, and an electrical load was proposed. Thisthesis presented the successful design and implementation of such a circuit.
The circuit was designed such that the battery connects to a dc bus througha bidirectional dc-dc converter, and the asynchronous generator connects tothe same dc bus through a bidirectional inverter. A positive and a negative dcbus rail were used and the load was connected between the two bus rails.
The bidirectional dc-dc converter circuit was successfully designed to reg-ulate the dc bus to 355 V rail-to-rail (within the design specication) by em-ploying two control loops, an inner current control loop and an outer voltagecontrol loop. The inner current control loop was used to control the amountof current owing into or out of the converter's batteries as a function of theset-point generated by the outer control loop that regulates the dc bus voltage.Both frequency and time domain analysis conrmed the closed-loop stabilityof the system.
The converter was practically implemented by manufacturing PCBs to con-struct the prototype circuit. The converter circuit used a half-bridge topologyconsisting of MOSFETs as switches, which were controlled using pulse-widthmodulation. A two-stage soft-start circuit was successfully implemented toavoid in-rush currents during start-up through the reversed biased diodes ofthe MOSFETs.
The transfer of up to 2.1 kW of electrical power between the batteries anddc bus was demonstrated by powering a 2.1 kW load connected between the dcbus rails. The step response of the converter circuit was tested and excellent
138
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 8. CONCLUSIONS 139
transient response was demonstrated. Overall the simulated and test resultscompared extremely well. The converter also showed a high eciency of 94 %.
The asynchronous generator, controlled by the inverter circuit, was success-fully modelled in a computer simulation when compared to the given datasheetinformation.
The detailed design of the bidirectional inverter circuit used to interfacethe asynchronous generator and the dc bus was presented and implemented.The switching action of the IGBTs used in the inverter circuit were controlledusing two control loops. The inner control loop was used to successfully controlthe magnitude of the current owing into the dc bus while the outer controlloop was used to regulate the converter's battery voltage. Both frequency andtime domain analysis showed the closed-loop inverter system to be stable.
A dsPIC digital signal processor was used along with a digital-to-analogconverter to produce the required sinusoidal modulator reference waveformsfor the inverter switches. A magnetic pick-up was used to determine the speedof the generator shaft and the result was processed inside the dsPIC. Thegenerator was shown to function as desired when operated at a constant slipof -5.47 % using current as well as Volt/Hertz control. The generator was ableto supply 2 kW of electrical power to the dc bus when a load was connectedbetween the bus rails.
The integration of the dc-dc converter circuit and inverter circuit was suc-cessfully implemented. The generator was shown to supply power to the dc-dcconverter circuit in order to charge the batteries to their optimal oating usevoltage of 272 V. The bidirectional capability of the dc-dc converter circuitwas thus proven.
The test results showed that as the battery voltage increased, the demandeddc bus current decreased, indicating the inverter control loops functioned asdesired. The dc bus current reached a minimum demanded value of 2.1 A,which was the necessary current required to keep the battery voltage at ap-proximately 272 V, thus trickle charging the batteries.
The prototype circuits showed that the induction generator can be con-trolled as a function of the dc-dc converter's battery voltage by making useof the presented dc-dc converter and inverter circuits. The overall systemperformance was shown to meet all the design specications.
8.2 Recommendations and Future Work
The prototype circuit was tested using another asynchronous machine to drivethe shaft of the generator. It is recommended that a turbine driven by airowis used to drive the generator shaft in future as this is the intended application.
If the system is desired to be used as a stand-alone system, it will requirea separate battery to power all the control and gate-drive circuitry.
Stellenbosch University https://scholar.sun.ac.za
CHAPTER 8. CONCLUSIONS 140
Although the prototype circuit was able to charge the batteries, a batterymanagement system should be considered to allow the state of charge of thevarious converter batteries to remain the same. Dierent charging algorithmsshould be investigated in order to nd the optimal battery charging solution.As an alternative to using lead-acid batteries, the use of lithium-ion batteriesas well as supercapacitors could be investigated.
A dierent IGBT module should be considered for the inverter circuit.The chosen module has a very high junction-to-case and case-to-sink thermalresistance. The resulting IGBT operating junction-to-ambient temperature isvery high and the inverter circuit will not be able to function in areas with anambient temperature above 40°C to 45°C.
The prototype circuit was designed as a concept demonstrator for a circuitused on board an aircraft to supply power to an electrical load. The circuitcan however also be used to supply power to small loads in remote areas byreplacing the RAM air turbine with a wind turbine. This could serve as arenewable energy solution in areas without access to the local power grid.
Stellenbosch University https://scholar.sun.ac.za
List of References
[1] K. Rafal, B. Morin, X. Roboam, E. Bru, C. Turpin, and H. Piquet, Hy-bridization of an aircraft emergency electrical network: Experimentationand benets validation, in 2010 IEEE Vehicle Power and PropulsionConference, Sept 2010, pp. 16.
[2] P. Bolognesi, F. Papini, and L. Taponecco, Hybrid-excitation dc machinesas highly reliable generators for ram air turbines, in Industrial Electron-ics, 2009. IECON '09. 35th Annual Conference of IEEE, Nov 2009, pp.25692574.
[3] X. Roboam, O. Langlois, H. Piquet, B. Morin, and C. Turpin, Hybridpower generation system for aircraft electrical emergency network, IETElectrical Systems in Transportation, vol. 1, no. 4, pp. 148155, December2011.
[4] T. A. and P. R. Surya, Autonomous wind-hydro hybrid system usingcage generators and battery storage, in Power and Energy Systems Con-ference: Towards Sustainable Energy, 2014, March 2014, pp. 16.
[5] P. K. Goel, B. Singh, S. S. Murthy, and N. Kishore, Isolated wind-hydrohybrid system using cage generators and battery storage, IEEE Transac-tions on Industrial Electronics, vol. 58, no. 4, pp. 11411153, April 2011.
[6] J. C. Ferreira, I. R. Machado, E. H. Watanabe, and L. G. B. Rolim, Windpower system based on squirrel cage induction generator, in XI BrazilianPower Electronics Conference, Sept 2011, pp. 943948.
[7] G. C. Konstantopoulos and A. T. Alexandridis, Full-scale modeling, con-trol, and analysis of grid-connected wind turbine induction generatorswith back-to-back ac/dc/ac converters, IEEE Journal of Emerging andSelected Topics in Power Electronics, vol. 2, no. 4, pp. 739748, Dec 2014.
[8] P. H. Mellor, S. G. Burrow, T. Sawata, and M. Holme, A wide-speed-range hybrid variable-reluctance/permanent-magnet generator for futureembedded aircraft generation systems, IEEE Transactions on IndustryApplications, vol. 41, no. 2, pp. 551556, March 2005.
141
Stellenbosch University https://scholar.sun.ac.za
LIST OF REFERENCES 142
[9] F. Lacressonniere, E. Bru, G. Fontes, and X. Roboam, Experimental val-idation of a hybrid emergency network with low and medium voltage liion batteries for more electrical aircraft, in Power Electronics and Appli-cations (EPE), 2013 15th European Conference on, Sept 2013, pp. 19.
[10] D. Holmes and T. Lipo, Pulse Width Modulation for Power Converters:Principles and Practice, ser. IEEE Press Series on Power Engineering.John Wiley & Sons, 2003.
[11] N. Mohan, T. Undeland, and W. Robbins, Power Electronics: Converters,Applications, and Design, 3rd ed., ser. Power Electronics: Converters,Applications, and Design. John Wiley & Sons, 2003.
[12] A. Pressman, K. Billings, and T. Morey, Switching Power Supply Design,3rd Ed. McGraw-Hill Education, 2007.
[13] Y. X. Wang, F. F. Qin, and Y. B. Kim, Bidirectional dc-dc con-verter design and implementation for lithium-ion battery application, in2014 IEEE PES Asia-Pacic Power and Energy Engineering Conference(APPEEC), Dec 2014, pp. 15.
[14] A. K. Verma, B. Singh, and D. T. Shahani, Grid to vehicle and vehicleto grid energy transfer using single-phase bidirectional ac-dc converterand bidirectional dc-dc converter, in Energy, Automation, and Signal(ICEAS), 2011 International Conference on, Dec 2011, pp. 15.
[15] S. Buso and P. Mattavelli, Digital Control in Power Electronics, ser. Lec-tures on power electronics. Morgan & Claypool Publishers, 2006.
[16] B. Bose, Modern Power Electronics and AC Drives, ser. Eastern EconomyEdition. Prentice Hall PTR, 2002, ch. 2 and 8, pp. 3074, 334413.
[17] T. Wildi, Electric Machines, Drives, and Power Systems, 6th ed. Pear-son, 2005, ch. 23, pp. 627656.
[18] L. Trilla, O. Gomis-Bellmunt, A. Junyent-Ferré, A. E. Álvarez, andA. Sudrià-Andreu, Control of a squirrel cage induction generator windfarm connected to a single power converter, in Universities Power En-gineering Conference (UPEC), 2010 45th International, Aug 2010, pp.16.
[19] P. Krause, O. Wasynczuk, S. Sudho, and I. P. E. Society, Analysis ofelectric machinery and drive systems, ser. IEEE Press series on powerengineering. IEEE Press, 2002, ch. 4.
[20] M. H. Granza, H. Voltolini, J. Ivanqui, and P. L. K. Miranda, Windpower generation control system with squirrel cage induction generator,
Stellenbosch University https://scholar.sun.ac.za
LIST OF REFERENCES 143
in Industry Applications (INDUSCON), 2014 11th IEEE/IAS Interna-tional Conference on, Dec 2014, pp. 16.
[21] R. Leidhold, G. Garcia, and M. I. Valla, Field-oriented controlled induc-tion generator with loss minimization, IEEE Transactions on IndustrialElectronics, vol. 49, no. 1, pp. 147156, Feb 2002.
[22] M. M. Bech, J. K. Pedersen, and F. Blaabjerg, Field-oriented control ofan induction motor using random pulsewidth modulation, IEEE Trans-actions on Industry Applications, vol. 37, no. 6, pp. 17771785, Nov 2001.
[23] S. Pati, S. Samantray, and N. C. Patel, A novel adaptive fuzzy controllerfor performance improvement of direct torque controlled induction gener-ator employed for wind power applications, in Emerging Research Areasand 2013 International Conference on Microelectronics, Communicationsand Renewable Energy (AICERA/ICMiCR), 2013 Annual InternationalConference on, June 2013, pp. 16.
[24] Z. Boulghasoul, A. Elbacha, E. Elwarraki, and D. Yous, Combined vec-tor control and direct torque control an experimental review and evalu-ation, in Multimedia Computing and Systems (ICMCS), 2011 Interna-tional Conference on, April 2011, pp. 16.
[25] B. Guru and H. Hiziroglu, Electromagnetic Field Theory Fundamentals,2nd ed. Cambridge University Press, 2004, pp. 282283.
[26] S. Umans, A. Fitzgerald, and C. Kingsley, Electric Machinery, 7th ed.,ser. International Edition. McGraw-Hill Higher Education, 2014, ch. 6.
[27] B. Guru and H. Hiziroglu, Electric Machinery and Transformers, 3rd ed.,ser. The Oxford Series in Electrical and Computer Engineering Series.OUP USA, 2000, ch. 9.
[28] T. Geyer, Model Predictive Control of High Power Converters and Indus-trial Drives. Wiley-Blackwell, 2016, ch. 4, pp. 3339.
[29] B. Ozpineci and L. M. Tolbert, Simulink implementation of inductionmachine model - a modular approach, in Electric Machines and DrivesConference, 2003. IEMDC'03. IEEE International, vol. 2, June 2003, pp.728734.
[30] P. Kemp, The design of an analogue class-d audio amplier using z-domain methods, M.S. thesis, University of Stellenbosch, March 2012.
[31] T. Mouton, Electronics 414 class notes, Depeartment of Electrical andElectronic Engineering, University of Stellenbosch, pp. 157160, 2014.
Stellenbosch University https://scholar.sun.ac.za
LIST OF REFERENCES 144
[32] Termal resistance calculator - plate n heat sink, calculation tools.MyHeatSinks Pte. Ltd. Livermore, California, USA. [Online]. Available:http://www.myheatsinks.com/calculate/thermal-resistance-plate-n/
[33] W.-R. Liou, P.-H. Chen, and J.-C. Tzeng, A synchronous boost regulatorwith pwm/pfm mode operation, in ASIC, 2009. ASICON '09. IEEE 8thInternational Conference, October 2009, pp. 10661069.
[34] D. Neamen, Microelectronics: Circuit Analysis and Design, 4th ed., ser.International Edition. McGraw-Hill, 2010, ch. 6, pp. 439440.
[35] P. Horowitz and W. Hill, The Art of Electronics, 2nd ed. CambridgeUniversity Press, 1995, ch. 2,4,15.
[36] F. Golnaraghi and B. C. Kuo, Automatic Control Systems, 9th ed. JohnWiley and Sons, Inc, 2009, ch. 4,8.
[37] G. C. Chryssis, High-Frequency Switching Power Supplies: Theory andDesign, 2nd ed. McGraw-Hill, 1989, pp. 237241.
[38] M. Brown, Power Supply Cookbook, 2nd ed., ser. EDN Series for DesignEngineers. Newnes, 2001, pp. 212216.
[39] T. Mouton, A. de Beer, B. Putzeys, and B. McGrath, Modelling anddesign of single-edge oversampled pwm current regulators using z-domainmethods, in ECCE Asia Downunder (ECCE Asia), 2013 IEEE, June2013, pp. 3137.
[40] G. Goodwin, S. Graebe, and M. Salgado, Control System Design. Pren-tice Hall, 2001, ch. 3, pp. 5257.
[41] X. Li, J. G. Park, and H. B. Shin, Comparison and evaluation of anti-windup pi controllers, Journal of Power Electronics, vol. 11, no. 1, pp.4550, 2011.
[42] G. Kenne, R. S. Simo, F. Lamnabhi-Lagarrigue, A. Arzande, and J. C.Vannier, An online simplied rotor resistance estimator for induction mo-tors, IEEE Transactions on Control Systems Technology, vol. 18, no. 5,pp. 11881194, Sept 2010.
[43] H. M. Jabr and N. C. Kar, Starting performance of saturated inductionmotors, in Power Engineering Society General Meeting, 2007. IEEE,June 2007, pp. 17.
[44] K. K. Tan, H. X. Zhou, and T. H. Lee, New interpolation method forquadrature encoder signals, IEEE Transactions on Instrumentation andMeasurement, vol. 51, no. 5, pp. 10731079, Oct 2002.
Stellenbosch University https://scholar.sun.ac.za
LIST OF REFERENCES 145
[45] H. V. Hoang and J. W. Jeon, Signal compensation and extraction ofhigh resolution position for sinusoidal magnetic encoders, in Control,Automation and Systems, 2007. ICCAS '07. International Conference on,Oct 2007, pp. 13681373.
[46] B. Kulakowski, J. Gardner, and J. Shearer, Dynamic Modeling and Con-trol of Engineering Systems. Cambridge University Press, 2007.
[47] G. Franklin, J. Powell, and A. Emami-Naeini, Feedback control of dynamicsystems, 6th ed., ser. International Edition. Pearson, 2010, pp. 633636.
Stellenbosch University https://scholar.sun.ac.za
Appendices
146
Stellenbosch University https://scholar.sun.ac.za
Appendix A
Complete Circuit Schematics
A.1 DC-DC Converter Circuit Schematics
All of the circuit diagrams used for the implementation of the bidirectionalcurrent-controller voltage-regulated dc-dc converter prototype circuit are givenin this section.
A.1.1 Power Stage Schematics
The schematics for the power stage of the dc-dc converter are given in thissection. It includes the half-bridge converter topologies, gate-drive circuitryand pre-charge soft-start mechanisms. The schematic for the top-half of theconverter is shown in Fig. A.1 and the schematic for the bottom half is shownin Fig. A.2.
147
Stellenbosch University https://scholar.sun.ac.za
APPENDIX A. COMPLETE CIRCUIT SCHEMATICS 148
11
22
33
44
55
66
77
88
DD
CC
BB
AA
D1
BA
T85
D2
BA
T85
Q2
BC
556
Q1
BC
556
T1 1:1
C1
100p
F
C2
100p
F
R1
0.22
K
R2
0.22
K
R3
1M
Ano
de2
Cat
hode
3
SGND 5
Vo(
B)
6V
o(A
)7
SVCC8
IC1
TLP2
50
R4
0.22
K
C3
100u
F
M1
IRFP
264
R5
27oh
m
+12V
D3
BA
T85
D4
BA
T85
Q4
BC
556
Q3
BC
556
T2 1:1
C4
100p
F
C5
100p
F
R6
0.22
K
R7
0.22
K
R8
1M
Ano
de2
Cat
hode
3
SGND 5
Vo(
B)
6V
o(A
)7
SVCC8
IC2
TLP2
50
R9
0.22
K
C6
100u
F
M2
IRFP
264
R10
27oh
m
+12V
1 2
CO
N1
PWM
1 1 2
CO
N2
PWM
2
PWM
1
PWM
2
L1 406u
H
1 2 3
CO
N3
BA
TTER
Y
R11
R12
0E68
R13
R14
R15
R16
R17
R18
R19
R20
R21
PWM
1
PWM
2
K1
Rel
ay
C11
1mF
Vs+
1 2
CO
N4
LOA
D
1
3
BA
L/ST
B6
BA
L5
V-
42
7
V+
8IC
3
LM31
1P
12
CO
N5
CS
Sign
al
+12V
Vs+ R
2222
0K
R24
10K
+12V
R26
1K
D5
5V6
D6
1N40
07
C8
100n
F
C9
100n
F
C10
100n
F
C7
100n
F
C12
1mF
R27
R28
R29100ohm
R30
4K7
R31
4K7
C13
100n
F
C14
100n
F
C15
10uF
SGN
D1
SGN
D1
SGN
D2
SGN
D2
D7
1N41
48
1 2
CO
N6
VS
Sign
al
+12V
1 2
CO
N7
12V
Sup
ply
0V
0V
0V
0V0V
GN
D
R23
2K
R32
10K
Q5
2N22
19A
D8
BY
V96
E
D9
BY
W96
E
D10
BY
W96
E
0V+12V
PIC101
PIC1
02
COC1
PIC201
PIC202
COC2
PIC301 PIC302COC
3
PIC401
PIC4
02
COC4
PIC501
PIC502
COC5
PIC601 PIC602COC
6
PIC701PIC702COC
7
PIC801PIC802COC
8
PIC901PIC902COC
9
PIC1001PIC1002COC10
PIC1101 PIC1102
COC11
PIC1201 PIC1202
COC12
PIC1301PIC1302COC13
PIC1401PIC1402COC14
PIC1501 PIC1502
COC15
PICON101
PICON102
COCON1
PICON201
PICON202
COCON2
PICO
N301
PICO
N302
PICO
N303
COCON3
PICON401
PICON402CO
CON4
PICON501PICON502
COCON5
PICO
N601
PICO
N602
COCON6
PICON701
PICON702
COCON7
PID101
PID102
COD1
PID201
PID202
COD2
PID301
PID302
COD3
PID401
PID402
COD4
PID501PID502CO
D5
PID601PID602CO
D6
PID701PID702CO
D7
PID801PID802CO
D8
PID901PID902COD
9
PID1001PID1002COD10
PIIC102
PIIC103
PIIC105
PIIC106
PIIC107
PIIC108COIC1
PIIC202
PIIC203
PIIC205
PIIC206
PIIC207
PIIC208COIC2
PIIC301
PIIC302
PIIC303
PIIC
304
PIIC305
PIIC306
PIIC307
PIIC308
COIC3
PIK101 PIK102
PIK103 PIK104
COK1
PIL101
PIL1
02
COL1
PIM101
PIM102 PIM103CO
M1
PIM201
PIM202 PIM203CO
M2
PIQ101PIQ102
PIQ103CO
Q1
PIQ201PIQ202
PIQ203 COQ2
PIQ301PIQ302
PIQ303CO
Q3
PIQ401PIQ402
PIQ403 COQ4
PIQ501
PIQ502 PIQ503COQ5
PIR101
PIR102
COR1
PIR201
PIR202
COR2
PIR301PIR302 COR3
PIR401PIR402 COR4
PIR501
PIR502
COR5
PIR601
PIR602
COR6
PIR701
PIR702
COR7
PIR801PIR802 COR8
PIR901PIR902 COR9
PIR1001
PIR1002
COR10
PIR1101PIR1102 COR11
PIR1
201
PIR1202
COR12
PIR1
301
PIR1302
COR13
PIR1
401
PIR1402
COR14
PIR1
501
PIR1502
COR15
PIR1
601
PIR160
2COR16
PIR1
701
PIR1702
COR17
PIR1
801
PIR1802
COR18
PIR1
901
PIR1902
COR19
PIR2
001
PIR2002
COR20
PIR2
101
PIR2102
COR21
PIR2201PIR2202 COR22
PIR2301PIR
2302PIR2303 CO
R23
PIR2401PIR2402 COR24
PIR2601PIR2602 COR26
PIR2701PIR2702 COR27PIR2801PIR2802
COR28
PIR2901PIR2902 COR29
PIR3001PIR3002 COR30
PIR3101PIR3102 COR31
PIR3201PIR3202 COR32
PIT101PIT102
PIT103PIT104
PIT105PIT106
PIT107PIT108 COT
1
PIT201PIT202
PIT203PIT204
PIT205PIT206
PIT207PIT208 COT
2
PIC702
PIC802 PIC902
PIC1002
PICO
N302
PICON701
PID602PID702
PID802
PIIC305
PIIC306
PIIC308
PIK103
PIQ103PIQ203
PIQ303PIQ403
PIR2602
PIC701
PIC801 PIC901
PIC1001
PIC1502PI
CON3
03
PICON501
PICON702
PID501
PID801
PID901
PIIC
304
PIQ503
PIR301
PIR801
PIR1
201
PIR1
301
PIR1
401
PIR1
501
PIR1
601
PIR1
701
PIR1
801
PIR1
901
PIR2
001
PIR2
101
PIR2401PIR3201
PIT106PIT107
PIT206PIT207
PIC602
PIC1102PIC1202
PIC1401
PICON402
PICON502
PICO
N602
PIIC205
PIM203PIR1202
PIR1302
PIR1402
PIR1502
PIR160
2
PIR1702
PIR1802
PIR1902
PIR2002
PIR2102
PIR3101
PIT202PIT203
PIC101
PIQ101
PIT105
PIC1
02
PIR201
PIR302PIC201
PIR102
PIC202
PIQ201
PIT108
PIC301
PIC1302
PID102
PID202
PIIC108
PIC401
PIQ301
PIT205
PIC4
02
PIR701 PI
C501
PIR602
PIR802PIC502
PIQ401
PIT208
PIC601
PIC1402
PID302
PID402
PIIC208
PIC1501PID701PIIC303
PIR2201
PIR230
2PIR2303
PICON101
PIR402
PICON102
PIIC103
PICON201
PIR902PICON202
PIIC203
PICO
N301
PIK102PIR1101
PIR2701PIR2801
PIR2901
PID101
PIT104
PID201
PIT101
PID301
PIT204
PID401
PIT201
PID502PIIC302
PIR2601PID601
PIIC307
PIK104PIQ502
PID902PID1001
PIK101PIL101
PIR1102PIR2702
PIR2802PIR2902
PIIC102
PIR401
PIIC106
PIIC107
PIR501
PIIC202
PIR901
PIIC206
PIIC207
PIR1001
PIIC301
PIQ501
PIR3202
PIQ102PI
R101
PIQ202
PIR202
PIQ302PI
R601
PIQ402
PIR702
PIR2301 PIR2402
PIM101
PIR502
PIR3002
NLPWM1
PIM201
PIR1002
PIR3102
NLPWM2
PIC302
PIC1301
PIIC105
PIL1
02
PIM103 PIM202
PIR3001
PIT102PIT103
PIC1101PIC1201
PICON401
PICO
N601
PID1002PIM102
PIR2202
Figure A.1: Schematic for the top-half of the dc-dc converter circuit.
Stellenbosch University https://scholar.sun.ac.za
APPENDIX A. COMPLETE CIRCUIT SCHEMATICS 149
11
22
33
44
55
66
77
88
DD
CC
BB
AA
D1
BAT8
5
D2
BAT8
5
Q2
BC55
6Q
1BC
556
T1 1:1
C1 100p
F
C2 100p
F
R1 0.22
K
R2 0.22
K
R3 1M
Ano
de2
Cath
ode
3
SGND 5
Vo(
B)6
Vo(
A)
7
SVCC8
IC1
TLP2
50
R4 0.22
K
C3 100u
F
R5 27oh
m
0V
D3
BAT8
5
D4
BAT8
5
Q4
BC55
6Q
3BC
556
T2 1:1
C4 100p
F
C5 100p
F
R6 0.22
K
R7 0.22
K
R8 1M
Ano
de2
Cath
ode
3
SGND 5
Vo(
B)6
Vo(
A)
7
SVCC8
IC2
TLP2
50
R9 0.22
K
C6 100u
F
M2
IRFP
264
R10
27oh
m
0V
1 2
CON
1
PWM
1 1 2
CON
2
PWM
2
PWM
1
PWM
2
L1 406u
H
1 2 3
CON
3
BATT
ERY
R11
R12
0E68
R13
R14
R15
R16
R17
R18
R19
R20
R21
PWM
2
K1
Rela
y
C11
1mF
1 2
CON
4
LOA
D
1
3
BAL/
STB
6
BAL
5
V-
42
7
V+
8IC
3
LM31
1P
12
CON
5
CS S
igna
l
-12V
-VsR2
222
0K
R24
10K
0V
R26
1K
D5
5V6
D6
1N40
07
C8 100n
F
C9 100n
F
C10
100n
F
C7 100n
F
C12
1mF
R27
R28
R29100ohm
R30
4K7
R31
4K7
C13
100n
F
C14
100n
F
C15
10uF
SGN
D2
SGN
D2
D7
1N41
48
1 2
CON
6
VS
Sign
al
0V1 2
CON
7
12V
Sup
ply
M1
IRFP
264
PWM
1
SGN
D1
SGN
D1
0V
-12V
-12V
-12V
-12V
-Vs
R23
2K
R32
10K
GN
DD
8BY
V96
E
D9
BYW
96E
D10
BYW
96E
Q5
2N22
19A
PIC101
PIC1
02
COC1
PIC201
PIC202
COC2
PIC301 PIC302COC
3
PIC401
PIC4
02
COC4
PIC501
PIC502
COC5
PIC601 PIC602COC
6
PIC701PIC702COC
7
PIC801PIC802COC
8
PIC901PIC902COC
9
PIC1001PIC1002COC10
PIC1101 PIC1102
COC11
PIC1201 PIC1202
COC12
PIC1301PIC1302COC13
PIC1401PIC1402COC14
PIC1501 PIC1502
COC15
PICO
N101
PICO
N102
COCON1
PICO
N201
PICO
N202
COCON2
PICO
N301
PICO
N302
PICO
N303
COCON3
PICO
N401
PICO
N402
COCON4
PICON501PICON502
COCON5
PICON601
PICON602CO
CON6
PICON701
PICON702
COCON7
PID101
PID102
COD1
PID201
PID202
COD2
PID301
PID302
COD3
PID401
PID402
COD4
PID501PID502CO
D5
PID601PID602CO
D6
PID701PID702CO
D7
PID801PID802CO
D8
PID901PID902COD
9
PID1001PID1002CO
D10
PIIC102
PIIC103
PIIC105
PIIC106
PIIC107
PIIC108COIC1
PIIC202
PIIC203
PIIC205
PIIC206
PIIC207
PIIC208COIC2
PIIC
301
PIIC302
PIIC303
PIIC304
PIIC305
PIIC306
PIIC307
PIIC
308
COIC3
PIK101 PIK102
PIK103 PIK104
COK1
PIL101
PIL1
02
COL1
PIM101
PIM102 PIM103CO
M1
PIM201
PIM202 PIM203CO
M2
PIQ101PIQ102
PIQ103CO
Q1
PIQ201PIQ202
PIQ203 COQ2
PIQ301PIQ302
PIQ303CO
Q3
PIQ401PIQ402
PIQ403 COQ4
PIQ501
PIQ502 PIQ503COQ5
PIR101
PIR102
COR1
PIR201
PIR202
COR2
PIR301PIR302 COR3
PIR401PIR402 COR4
PIR5
01PI
R502
COR5
PIR601
PIR602
COR6
PIR701
PIR702
COR7
PIR801PIR802 COR8
PIR901PIR902 COR9
PIR1
001
PIR1
002
COR10
PIR1101PIR1102 COR11
PIR1
201
PIR1
202
COR12
PIR1
301
PIR1
302
COR13
PIR1
401
PIR1
402
COR14
PIR1
501
PIR1
502
COR15
PIR1
601
PIR1
602
COR16
PIR1
701
PIR1
702
COR17
PIR1
801
PIR1
802
COR18
PIR1
901
PIR1
902
COR19
PIR2
001
PIR2
002
COR20
PIR2
101
PIR2
102
COR21
PIR2201PIR2202 COR22
PIR2301PIR2302PIR2303COR23
PIR2401PIR2402 COR24
PIR2601PIR2602 COR26
PIR2701PIR2702COR27
PIR2801PIR2802 COR28PIR2901PIR2902
COR29
PIR3001PIR3002 COR30
PIR3101PIR3102 COR31
PIR3201PIR3202 COR32
PIT101PIT102
PIT103PIT104
PIT105PIT106
PIT107PIT108 CO
T1
PIT201PIT202
PIT203PIT204
PIT205PIT206
PIT207PIT208 COT
2
PIC701
PIC801 PIC901
PIC1001
PICO
N302
PICON702
PID701
PID801
PIIC304
PIQ503
PIR301
PIR801
PIR2601PIR3201
PIT106PIT107
PIT206PIT207
PIC302
PIC1102PIC1202
PIC1301
PICO
N402
PICON602
PID1001
PIIC105
PIM103PIR2201
PIR3001
PIT102PIT103
PIC702
PIC802 PIC902
PIC1002PIC1501
PICO
N301
PICON502PICON701
PID502PID602
PID802PID902
PIIC305
PIIC306
PIIC
308
PIK103
PIQ103PIQ203
PIQ303PIQ403
PIR1
201
PIR1
301
PIR1
401
PIR1
501
PIR1
601
PIR1
701
PIR1
801
PIR1
901
PIR2
001
PIR2
101
PIR2302PIR2303
PIC1101PIC1201
PICO
N401
PICON501
PICON601
PIM202
PIR1
202
PIR1
302
PIR1
402
PIR1
502
PIR1
602
PIR1
702
PIR1
802
PIR1
902
PIR2
002
PIR2
102
PIC101
PIQ101
PIT105
PIC1
02
PIR201
PIR302PIC201
PIR102
PIC202
PIQ201
PIT108
PIC301
PIC1302
PID102
PID202
PIIC108
PIC401
PIQ301
PIT205
PIC4
02
PIR701 PI
C501
PIR602
PIR802PIC502
PIQ401
PIT208
PIC601
PIC1402
PID302
PID402
PIIC208
PIC1502 PID702PIIC302
PIR2202PIR2401
PICO
N101
PIR402
PICO
N102
PIIC103
PICO
N201
PIR902PI
CON2
02
PIIC203
PICO
N303 PIK101
PIR1102PIR2702
PIR2802PIR2902
PID101
PIT104
PID201
PIT101
PID301
PIT204
PID401
PIT201
PID501PIIC303
PIR2602
PID601PIIC307
PIK104PIQ502
PID901 PID1002
PIK102PIL101
PIR1101PIR2701
PIR2801PIR2901
PIIC102
PIR401
PIIC106
PIIC107
PIR5
01
PIIC202
PIR901
PIIC206
PIIC207
PIR1
001
PIIC
301
PIQ501
PIR3202
PIQ102PI
R101
PIQ202
PIR202
PIQ302PI
R601
PIQ402
PIR702
PIR2301 PIR2402
PIM101
PIR5
02
PIR3002
NLPWM1
PIM201
PIR1
002
PIR3102
NLPWM2
PIC602
PIC1401
PIIC205
PIL1
02
PIM102PIM203
PIR3101
PIT202PIT203
Figure A.2: Schematic for the bottom-half of the dc-dc converter circuit.
Stellenbosch University https://scholar.sun.ac.za
APPENDIX A. COMPLETE CIRCUIT SCHEMATICS 150
A.1.2 Control Stage Schematics
The complete control schematic for the top- and bottom-half of the dc-dcconverter circuit is given in this section in Fig.A.3.
Stellenbosch University https://scholar.sun.ac.za
APPENDIX A. COMPLETE CIRCUIT SCHEMATICS 151
11
22
33
44
55
66
77
88
DD
CC
BB
AA
2 31
A8 4
[H(s
)]
IC2A
LF35
3N
Vin
-1
Vin
+2
Vou
t+3
COM
4
Vou
t-5
IC1
DK
E10A
-15
1 2
CON
1
Batte
ry 1
2V
+12V 0V
+15V
-15V
GN
D
+15V
-15V
GN
D
R1 4K7
R2 4K7
R3 33K
R4 33K
567
B
8 4
[F(s
)]
IC2B
LF35
3N
+15V
-15V
R5 3K3
R6 3K3
R7 4K7
R8 4K7
C2 470n
FC1 10nF
C3 10nF
C4 470n
F
GN
D
IN2
VTE
MP
3
GN
D4
TRIM
5O
UT
6N
C8
NC
7N
C1
IC3
REF0
3GP
+15V
GN
D
C5 100n
F
8 5
2 3
4
6
7
1
IC4
OP0
7CP
8 5
2 3
4
6
7
1
IC5
OP0
7CP
+15V
+15V
-15V
-15V
R9 4K7
R10
4K7
R11
4K7
R12
4K7
GN
D
GN
D
+5V
REF
-5V
REF
2 31
A
8 4
IC6A
LF35
3N
567
B8 4
IC6B
LF35
3N
+15V
+15V
-15V
-15V
R13
0K47
-5V
REF
D1
BAT8
5D
2BA
T85
Ve
Cla
mpe
d1
C6 10uF
+2.5
V R
EF
2 31
A8 4
IC7A
LF35
3N
567
B8 4
IC7B
LF35
3N
+15V
+15V
-15V
-15V
R16
6K8
R17
3K3
R18
3K3
R19
10K
R15 2K G
ND+2.5V REF
+3V
REF
R20
2K
GN
D
+2.5
V R
EF
-6V
REF
+15V
+15V
-15V
-15V
-6V
REF
2 31
A8 4
IC8A
LF35
3N
567
B8 4
IC8B
LF35
3N
D3
BAT8
5D
4BA
T85
R21
0K47 +3
V R
EF
567
B8 4
[J(s
)]
IC10
BLF
353N
2 31
A8 4
IC10
ALF
353N
+15V
+15V
-15V
-15V
R30
1K
R31
100K
R32
6K8
R33
18K
R34
18K
R35
6K8
R36
220K
R37
220K
12
IC12
A
M74
HCT
14B
34
IC12
B
M74
HCT
14B
98
IC12
D
M74
HCT
14B
RST
1
CLK
9
D0
3Q
02
D1
4Q
15
D2
6Q
27
D3
11Q
310
D4
13Q
412
D5
14Q
515
VCC
16
GN
D8
IC13
M74
HC1
74B
1R
RST
1
CLK
9
D0
3Q
02
D1
4Q
15
D2
6Q
27
D3
11Q
310
D4
13Q
412
D5
14Q
515
VCC
16
GN
D8
IC14
M74
HC1
74B
1R
2 31
A8 4
IC15
ALF
353N
567
B8 4
IC15
BLF
353N
12
OSC
1
4MH
zR4
5
10M
R46
1KR4
71K
R48
22K
R49
22K
R50
1MR51
1M
R52
1K5
R53
4K7
C9 22pF
C10
22pF
C11
100n
F
C12
100n
F
C13
100u
FC1
410
0uF
GN
D
GN
D
+5V
56
IC12
C
M74
HCT
14B
1110
IC12
E
M74
HCT
14B
1312
IC12
F
M74
HCT
14B
+5V
GN
D
+5V
+5V
+15V
+15V
-15V
-15V
C15
100n
F
C16
1nF
C17
1nF
GN
D
GN
D
GN
D
Vtri
567
B8 4
IC18
BLF
353N
2 31
A8 4
IC18
ALF
353N
+15V
+15V
-15V
-15V
R64
4K7
R65
4K7
R66
4K7
R67
4K7
R68
4K7
R69
4K7
R70
4K7R7
1
4K7
GN
DV
eS2
VeS
1
Ve
Cla
mpe
d1
Ve
Cla
mpe
d1
1
23
4
5
6
7
8 IC19
AD
790J
N
1
23
4
5
6
7
8 IC20
AD
790J
N
GN
D
GN
D
+15V
+15V
-15V
-15V+5
V
+5V
C20
100n
F
C21
100n
F
C22
100n
F
C23
100n
F
C24
100n
F
C25
100n
F
GN
D
GN
D
GN
D
GN
D
GN
DV
r1
Vr1
1 2
CON
5
PWM
2
1 2
CON
4
PWM
1
GN
D
GN
D
IN1
2
OU
T3
GN
D
VS1
MC7
805C
T
GN
D
+15V
+5V
8
1
4
32
1IC
17A
LM35
8N
8 4
756
2IC
17B
LM35
8N
+15v
+15V
GN
D
GN
D
D6
1N41
48
GN
D
R55
100K
R56
1K8
R57
5K6
R58
1K8
R59
5K6
R54
100K
D5
1N41
48
C18
100n
F
C19
100n
F
C37
470u
F
C26
100n
F
+5V
GN
D
C27
100n
F
C28
100n
F
GN
D
GN
D
C29
100n
FG
ND
C30
100n
F
+5V
REF
C31
100n
F
C32
100n
FC3
310
0nF
GN
D
C34
100n
F
GN
D
GN
D
R25
1K5
R26
1K5
12
CON
2
CS S
igna
l 1C7 47
nF
C8 100p
F
C35
100p
FC3
647
nF
R29
100K
R28
1K
+5V
REF
GN
D
+5V
REF
GN
D
Vof
fset
Vof
fset
C38
100u
FC3
9
100u
F
(+5V
)(-
5V)
123
CON
3
VS
Sign
al 1
GN
D
SetI1
Vre
f1
C40
100n
F
GN
D
C41
100n
FC4
2
100n
F
C43
100n
F
(-5V
)(+
5V)
GN
D
GN
D
567
B8 4
IC23
BLF
353N
2 31
A8 4
IC23
ALF
353N
+15V
+15V
-15V
-15V
R80
4K7
R81
4K7
R79
4K7
R82
4K7
R86
4K7
R84
4K7
R85
4K7R8
3
4K7
GN
DV
eS3
VeS
4
Ve
Cla
mpe
d2
Ve
Cla
mpe
d2
1
23
4
5
6
7
8 IC24
AD
790J
N
1
23
4
5
6
7
8 IC25
AD
790J
N
GN
D
GN
D
+15V
+15V
-15V
-15V+5
V
+5V
C58
100n
F
C57
100n
F
C56
100n
F
C61
100n
F
C60
100n
F
C59
100n
F
GN
D
GN
D
GN
D
GN
D
GN
DV
r2
Vr2
1 2
CON
9
PWM
3
1 2
CON
8
PWM
4
GN
D
GN
D
Vof
fset
2 31
A8 4
[H(s
)]
IC21
ALF
353N
+15V
-15V
GN
D
R72
4K7
R73
4K7
R76
33K
R74
33K
567
B
8 4
[F(s
)]
IC21
BLF
353N
+15V
-15V
R22
3K3
R23
3K3
R27
4K7
R24
4K7
C52
470n
FC53
10nF
C50
10nF
C51
470n
F
GN
D
2 31
A
8 4
IC22
ALF
353N
567
B8 4
IC22
BLF
353N
+15V
+15V
-15V
-15V
R14
0K47
-5V
REF
D9
BAT8
5D
10BA
T85
Ve
Cla
mpe
d2
+15V
+15V
-15V
-15V
-6V
REF
2 31
A8 4
IC11
ALF
353N
567
B8 4
IC11
BLF
353N
D7
BAT8
5D
8BA
T85
R44
0K47 +3
V R
EF
567
B8 4
[J(s
)]
IC16
BLF
353N
2 31
A8 4
IC16
ALF
353N
+15V
+15V
-15V
-15V
R60
1K
R63
100K
R40
6K8
R39
18K
R42
18K
R43
6K8
R38
220K
R41
220K
+5V
REF
GN
D
R75
1K5
R78
1K5
12
CON
7
CS S
igna
l 2C4
7
47nF
C46
100p
F
C44
100p
FC4
547
nF
R62
100K
R61
1K
+5V
REF
GN
D
123
CON
6
VS
Sign
al 2
GN
D
SetI2
Vre
f2
C48
100n
F
GN
D
C49
100n
FC5
4
100n
F
C55
100n
F
(-5V
)(+
5V)
GN
D
GN
D
2 31
A8 4
IC26
ALF
353N
567
B8 4
IC26
BLF
353N
+15V
+15V
-15V
-15V
R87
1K
R88
1K
R89
1K R90
1K
GN
D
Vtri
Vr1
R91
1K
R92
1KG
ND
GN
D
Vtri
Vr2
PIC1
01PIC102
COC1
PIC201
PIC2
02
COC2
PIC301PIC302CO
C3
PIC401PIC402COC
4
PIC501PIC502COC
5PIC601 PIC602
COC6
PIC701
PIC7
02
COC7PI
C801
PIC802
COC8
PIC901PIC902COC
9PIC1001PIC1002
COC10
PIC1101PIC1102COC11
PIC1201PIC1202COC12
PIC1301 PIC1302COC13
PIC1401 PIC1402
COC14
PIC1
501
PIC1502
COC15
PIC1601
PIC1602
COC16
PIC1701PIC1702CO
C17
PIC1801
PIC1
802
COC18
PIC1901PIC1902CO
C19
PIC2
001
PIC2002COC20
PIC2101
PIC2
102
COC21
PIC2201
PIC2
202
COC22
PIC2
301
PIC2302COC23
PIC2401
PIC2
402
COC24
PIC250
1PI
C250
2
COC25
PIC2601PIC2602COC26
PIC2701
PIC2702CO
C27
PIC2801
PIC2
802COC28
PIC2901
PIC2902CO
C29
PIC3001
PIC3
002
COC30
PIC310
1PI
C310
2
COC31
PIC320
1PI
C320
2
COC32
PIC3301PIC3302COC33
PIC3401PIC3402COC34
PIC3501PIC3502COC35
PIC3601PIC3602COC36
PIC3701 PIC3702COC37
PIC3
801
PIC3802
COC38 PIC3
901
PIC3902COC39
PIC4001
PIC4
002COC40
PIC410
1PI
C410
2
COC41
PIC4
201
PIC4202
COC42
PIC4301
PIC4
302
COC43
PIC4401PIC4402COC44
PIC4501PIC4502COC45
PIC4
601
PIC4602
COC46
PIC4701
PIC4
702
COC47
PIC4801
PIC4
802COC48
PIC4901
PIC4
902
COC49
PIC5001PIC5002COC50
PIC5101PIC5102COC51
PIC520
1PI
C520
2
COC52
PIC5
301
PIC5302
COC53
PIC5
401
PIC540
2
COC54
PIC550
1PI
C550
2
COC55
PIC5601
PIC5
602
COC56
PIC5701
PIC5
702
COC57
PIC5
801
PIC5802COC58
PIC590
1PI
C590
2
COC59
PIC6001
PIC6
002
COC60
PIC6
101
PIC6102COC61
PICON101
PICON102
COCON1
PICON201
PICON202
COCON2
PICON301
PICON302
PICON303
COCON3
PICO
N401
PICO
N402
COCON4
PICO
N501
PICO
N502
COCON5
PICON601
PICON602
PICON603
COCON6
PICON701
PICON702
COCON7
PICO
N801
PICO
N802
COCON8
PICO
N901
PICO
N902
COCON9
PID101 PID102CO
D1
PID201PID202CO
D2
PID301 PID302COD
3
PID401PID402CO
D4
PID501PID502COD
5
PID601PID602CO
D6
PID701 PID702CO
D7
PID801PID802CO
D8
PID901 PID902CO
D9
PID1001PID1002COD10
PIIC101
PIIC102
PIIC103
PIIC104
PIIC105
COIC1
PIIC201
PIIC202
PIIC203
PIIC204PIIC208
COIC2A
PIIC204PIIC205
PIIC206
PIIC
207
PIIC208COIC2B
PIIC301
PIIC302
PIIC303
PIIC304
PIIC305
PIIC306
PIIC307
PIIC308
COIC3
PIIC401
PIIC402
PIIC403
PIIC404PIIC405
PIIC406
PIIC407PIIC408
COIC4
PIIC501
PIIC502
PIIC503
PIIC504PIIC505
PIIC506
PIIC507PI
IC50
8
COIC5
PIIC601
PIIC602
PIIC603
PIIC604PIIC608
COIC6A
PIIC604PIIC605
PIIC
606
PIIC607
PIIC608
COIC6B
PIIC701
PIIC702
PIIC703
PIIC704PIIC708 COIC
7A
PIIC704PIIC705
PIIC706
PIIC
707
PIIC708
COIC
7B
PIIC801
PIIC802
PIIC
803
PIIC804PIIC808
COIC8A
PIIC804PI
IC80
5
PIIC806
PIIC807
PIIC808
COIC8B
PIIC
1001
PIIC1002
PIIC1003
PIIC1004PIIC1008
COIC
10A
PIIC1004PIIC1005
PIIC1006
PIIC
1007
PIIC1008
COIC
10B
PIIC1101
PIIC1102
PIIC1103
PIIC1104PIIC1108
COIC
11A
PIIC1104PIIC1105
PIIC1106
PIIC
1107
PIIC1108
COIC
11B
PIIC1201
PIIC1202
COIC
12A
PIIC1203
PIIC1204
COIC
12B
PIIC1205
PIIC
1206
COIC
12C
PIIC1208
PIIC1209
COIC
12D
PIIC
1201
0PI
IC12
011
COIC
12E
PIIC
1201
2PI
IC12
013
COIC
12F
PIIC
1301
PIIC1302
PIIC
1303
PIIC
1304
PIIC1305
PIIC
1306
PIIC1307
PIIC
1308
PIIC
1309
PIIC
1301
0PI
IC13
011
PIIC
1301
2PI
IC13
013
PIIC
1301
4PI
IC13
015
PIIC
1301
6
COIC
13
PIIC
1401
PIIC1402
PIIC
1403
PIIC
1404
PIIC1405
PIIC
1406
PIIC1407
PIIC
1408
PIIC
1409
PIIC
1401
0PI
IC14
011
PIIC
1401
2PI
IC14
013
PIIC
1401
4PI
IC14
015
PIIC
1401
6
COIC
14
PIIC
1501
PIIC1502
PIIC1503
PIIC1504PIIC1508 COIC
15A
PIIC1504PIIC1505
PIIC1506
PIIC1507
PIIC1508CO
IC15
B
PIIC
1601
PIIC1602
PIIC1603
PIIC1604PIIC1608
COIC
16A
PIIC1604PIIC1605
PIIC1606
PIIC
1607
PIIC1608
COIC
16B
PIIC1701
PIIC1702
PIIC1703
PIIC1704PIIC1708 COIC
17A
PIIC1704PIIC1705
PIIC1706
PIIC1707
PIIC1708CO
IC17
B
PIIC1801
PIIC
1802
PIIC
1803
PIIC1804PIIC1808 COIC
18A
PIIC1804PI
IC18
05
PIIC
1806
PIIC1807
PIIC1808COIC
18B
PIIC1901
PIIC1902
PIIC1903
PIIC1904
PIIC1905
PIIC1906PIIC1907
PIIC1908 COIC19
PIIC2001
PIIC2002
PIIC2003
PIIC2004
PIIC2005
PIIC2006PIIC2007
PIIC2008 COIC
20
PIIC2101
PIIC
2102
PIIC
2103
PIIC2104PIIC2108
COIC
21A
PIIC2104PI
IC21
05
PIIC
2106
PIIC2107
PIIC2108CO
IC21
B
PIIC2201
PIIC
2202
PIIC
2203
PIIC2204PIIC2208
COIC
22A
PIIC2204PIIC2205
PIIC2206
PIIC
2207
PIIC2208
COIC
22B
PIIC2301
PIIC
2302
PIIC
2303
PIIC2304PIIC2308 COIC
23A
PIIC2304PI
IC23
05
PIIC
2306
PIIC2307
PIIC2308COIC
23B
PIIC2401
PIIC2402
PIIC2403
PIIC2404
PIIC2405
PIIC2406PIIC2407
PIIC2408 COIC
24
PIIC2501
PIIC2502
PIIC2503
PIIC2504
PIIC2505
PIIC2506PIIC2507
PIIC2508 COIC25
PIIC2601
PIIC2602
PIIC2603
PIIC2604PIIC2608 COIC
26A
PIIC2604PIIC2605
PIIC2606
PIIC2607
PIIC2608COIC
26B
PIOSC101
PIOSC102
COOSC1
PIR101
PIR102
COR1
PIR201
PIR202
COR2
PIR301
PIR302
COR3
PIR401PIR402 COR4
PIR501
PIR5
02CO
R5 PIR6
01PI
R602COR
6
PIR7
01PIR702
COR7
PIR801PIR802 COR8
PIR9
01PI
R902CO
R9PIR1001
PIR1
002
COR10
PIR1101
PIR1102
COR11
PIR1201
PIR1202
COR12
PIR130
1PIR
1302
COR13
PIR1401
PIR1402
COR14
PIR1501 PIR1502
PIR1503COR15
PIR1601
PIR1
602
COR16
PIR1701PIR1702 COR17
PIR1801
PIR1802
COR18
PIR1
901
PIR1902
COR19
PIR2001 PIR2002
PIR2003
COR20
PIR2101
PIR2102
COR21
PIR2201
PIR2
202
COR22
PIR2
301
PIR2
302COR23
PIR2401PIR2402 COR24
PIR2
501
PIR2
502COR25
PIR2601
PIR2602
COR26
PIR2
701
PIR270
2COR27
PIR2
801
PIR280
2COR28
PIR2901PIR2902 COR29
PIR3
001
PIR3
002
COR30
PIR3101
PIR3102
COR31
PIR3
201
PIR3202
COR32
PIR3
301
PIR3302
COR33
PIR3
401
PIR340
2COR34
PIR3501PIR3502 COR35
PIR3601
PIR3602CO
R36 PIR
3701
PIR370
2COR37 PI
R3801
PIR3802CO
R38
PIR3
901
PIR3902
COR39
PIR4
001
PIR400
2COR40
PIR4101
PIR4102CO
R41
PIR4
201
PIR4202
COR42
PIR4301PIR4302 COR43
PIR4401
PIR4402
COR44
PIR4501
PIR4502
COR45
PIR4601PIR4602 COR46
PIR4701PIR4702 COR47
PIR4801
PIR4802
COR48
PIR490
1PIR
4902
COR49
PIR5001PIR5002 COR50
PIR5101
PIR5102
COR51
PIR5201
PIR5202
COR52
PIR5
301
PIR5
302
COR53
PIR5401 PIR5402
PIR5403
COR54
PIR5501PIR5502 COR55
PIR5601PIR5602 COR56
PIR5701PIR5702 COR57
PIR5801 PIR5802COR58
PIR5
901
PIR5902COR
59
PIR6
001
PIR6
002
COR60 PIR6
101
PIR6102COR61
PIR6201PIR6202 COR62
PIR6301
PIR6302
COR63
PIR6401
PIR6402
COR64 PIR
6501
PIR650
2COR65
PIR6
601
PIR6
602
COR66
PIR6701
PIR6
702
COR67
PIR6801
PIR6
802
COR68
PIR6901
PIR6902
COR69
PIR7001
PIR7002
COR70
PIR7101
PIR7102
COR71
PIR7201
PIR7202
COR72
PIR7301
PIR7302
COR73
PIR7401PIR7402 COR74
PIR7
501
PIR7
502COR75
PIR7601
PIR7602
COR76
PIR7801
PIR7802
COR78
PIR7
901
PIR7
902
COR79
PIR8001
PIR8002
COR80 PI
R8101
PIR8102CO
R81
PIR820
1PI
R820
2COR82
PIR8301
PIR8302
COR83
PIR8401
PIR8402
COR84
PIR8501
PIR8502
COR85
PIR8601
PIR8
602
COR86
PIR8
701
PIR8702
COR87
PIR8
801
PIR8
802
COR88
PIR8
901
PIR8902
COR89
PIR9001
PIR9
002
COR90
PIR9101PIR9102 COR91
PIR9
201
PIR9
202
COR92
PIVS101
PIVS102
PIVS103
COVS1
PIIC306
PIIC403
PIR1501
PIR2001NL0205V REF
PIC3302PIIC701
PIIC
803
PIIC1103
PIR1
602
NL03V REF
PIC1102
PIC1202
PIC2101
PIC2401
PIC2602
PIC3
202
PIC5701
PIC6001
PIIC
1201
4
PIIC
1301
6PI
IC14
016
PIIC1905
PIIC1908
PIIC2005
PIIC2008
PIIC2405
PIIC2408
PIIC2505
PIIC2508
PIR4602PIR4702
PIVS103
PIC3701
PIIC406
PIIC603
PIIC
2203
PIR1
002
PIR1101
PIR3
001
PIR5401
PIR6
001
NL05V REF
PICON101
PIIC102
PIC502PIC601
PIC2
202
PIC2
502
PIC2
802
PIC2902
PIC310
1
PIC3
801
PIC410
1
PIC4301
PIC4901
PIC550
1
PIC5
602
PIC5
902
PIIC103
PIIC208
PIIC302
PIIC407PIIC507
PIIC608
PIIC708
PIIC808
PIIC1008
PIIC1108
PIIC1508
PIIC1608 PIIC1708
PIIC1808PIIC1901 PIIC2001
PIIC2108
PIIC2208PIIC2308
PIIC2401 PIIC2501
PIIC2608
PIVS101
PIIC506
PIIC605
PIIC2205
PIR1202
NL05V REF
PIC3402PI
IC70
7
PIIC
805
PIIC1105
PIR1902
NL06V REF
PIC2
001
PIC2
301
PIC2701
PIC3001
PIC3902
PIC4001
PIC4202
PIC4801
PIC540
2
PIC5
801
PIC6
101
PIIC105
PIIC204
PIIC404PIIC504
PIIC604
PIIC704
PIIC804
PIIC1004
PIIC1104
PIIC1504
PIIC1604
PIIC1804PIIC1904 PIIC2004
PIIC2104
PIIC2204PIIC2304
PIIC2404 PIIC2504
PIIC2604
PICON102
PIIC101
PIC301
PIC401
PIC501PIC602
PIC901PIC1001
PIC1101
PIC1201
PIC1302PIC1402
PIC1701PIC1901
PIC2002
PIC2
102
PIC2201
PIC2302
PIC2
402
PIC250
1
PIC2601
PIC2702
PIC2801
PIC2901
PIC3
002
PIC3
102
PIC320
1
PIC3301PIC3401
PIC3501
PIC3601
PIC3802
PIC3
901
PIC4
002
PIC4
102
PIC4
201
PIC4
302
PIC4401
PIC4501PI
C480
2
PIC4
902
PIC5001
PIC5101
PIC5
401
PIC5
502
PIC5601
PIC5
702
PIC5802
PIC590
1
PIC6
002
PIC6102
PICON303
PICO
N402
PICO
N502
PICON603
PICO
N802
PICO
N902
PID501
PIIC104
PIIC304
PIIC503
PIIC705
PIIC1207
PIIC
1201
2
PIIC
1308
PIIC
1301
5
PIIC
1408
PIIC
1401
5
PIIC1505
PIIC1704PIIC1705
PIIC1906 PIIC2006 PIIC2406 PIIC2506
PIIC2603
PIR9
02
PIR1503
PIR1701
PIR2003
PIR2
502
PIR3501 PIR4301
PIR5001PIR5701
PIR5802
PIR650
2
PIR7101
PIR7
502
PIR8102
PIR8301
PIR8
901
PIR9101
PIVS102
PIC1
01
PIC201
PIIC206
PIR5
02
PIC102
PIIC
207
PIR702
PIR130
1
PIC2
02PI
R701
PIC302
PIIC205
PIR6
01
PIR802 PIC402PIR801
PIC701
PIC8
01
PIIC1006
PIR3
002
PIC7
02PIR3101
PIC902PIIC1201
PIOSC101
PIR4501
PIC1002
PIIC1202PIIC1203
PIOSC102
PIR4502
PIC1301
PIIC
1301
PIR4601
PIC1401
PIIC
1401
PIR4701
PIC1
501
PIC1602
PIIC
1501
PIR5102
PIC1502PIR5201
PIC1601
PIIC1502
PIR4802PIR5101
PIC1702PIIC1503
PIR490
2PIR5002
PIC1801
PIIC1702
PIR5801
PIR5902
PIC1902PIIC1703
PIR5601 PIR5702
PIC3502
PIIC1005
PIR2
801
PIR2902 PIC3602PIR2901
PIC3702 PID502
PIR5501 PIR5602
PIC4402
PIIC1605
PIR6
101
PIR6202 PIC4502PIR6201
PIC4
601
PIC4701
PIIC1606
PIR6
002
PIC4
702
PIR6301
PIC5002
PIIC
2105
PIR2
301
PIR2402 PIC5102PIR2401
PIC520
1PIC5
301
PIIC
2106
PIR2
202
PIC5
202P
IR27
01
PIC5302
PIIC2107PIR1401
PIR270
2
PICON201 PIR201
PICON202PIR101
PICON301
PIR370
2
PICON302
PIR3602
PICO
N401
PIIC1907
PICO
N501
PIIC2007
PICON601
PIR4102
PICON602
PIR3802
PICON701 PIR7301
PICON702PIR7201
PICO
N801
PIIC2407
PICO
N901
PIIC2507
PID102PIIC601
PID201PIIC607
PID302PIIC801
PID401PIIC807
PID601PIIC1707
PID702PIIC1101
PID801PI
IC11
07
PID902PIIC2201
PID1001PI
IC22
07
PIIC201
PIR501
PIR2602
PIIC202
PIR102
PIR301
PIIC203
PIR202
PIR402
PIIC301
PIIC303
PIIC305
PIIC307
PIIC308
PIIC401
PIIC402
PIR9
01PIR1001
PIIC405
PIIC408
PIIC501
PIIC502
PIR1102
PIR1201
PIIC505
PIIC
508
PIIC702
PIR1601
PIR1702PIIC703
PIR1502
PIIC706
PIR1802PIR1
901
PIIC
1001
PIR280
2
PIR3202
PIIC1002
PIR3
201
PIR3302 PIIC1003
PIR340
2 PIR3502
PIIC1204
PIIC
1309
PIIC1205
PIIC
1301
2
PIIC
1206
PIIC
1303
PIIC1208P
IR4901
PIIC1209
PIIC
1401
0
PIIC
1401
3
PIR4801
PIIC
1201
0
PIIC
1403
PIIC
1201
1
PIIC
1401
2
PIIC
1201
3
PIIC1302
PIIC
1304
PIIC1305
PIIC
1306
PIIC1307
PIIC
1301
1PI
IC13
010
PIIC
1301
3
PIIC
1409
PIIC
1301
4
PIIC1402
PIIC
1404
PIIC1405
PIIC
1406
PIIC1407
PIIC
1401
1
PIIC
1401
4
PIIC1506
PIR5202PI
R530
1
PIIC
1601
PIR400
2
PIR6102
PIIC1602
PIR3902PIR4
001
PIIC1603
PIR4202 PIR4302
PIIC
1802
PIR6
602PI
R6701
PIIC
1803
PIR6402
PIR650
1 PIIC
1805
PIR6902
PIR7002
PIIC
1806
PIR6801
PIR7102
PIIC2101
PIR2201
PIR7802
PIIC
2102
PIR7202PI
R7601
PIIC
2103
PIR7302 PIR7402
PIIC
2302
PIR7
902PIR
8201
PIIC
2303
PIR8002
PIR8101 PIIC
2305
PIR8402
PIR8502
PIIC
2306
PIR8302PIR8601
PIIC2602
PIR8702
PIR8
801
PIIC2605
PIR9
002 PIR9102PI
IC2606
PIR8902
PIR9
201
PIR302
PIR2601
PIR401PI
R250
1
PIR1801
PIR2002
PIR3
301
PIR3601 PI
R340
1PIR
3701
PIR3801PIR
3901
PIR4101PIR
4201
PIR5402
PIR5403 PIR5502
PIR7401PI
R750
1
PIR7602
PIR7801
PID301PID402
PIIC802
PIIC806
PIR6
02
PIR2102
NLSetI
1
PID701PID802
PIIC1102
PIIC1106
PIR2
302
PIR4402
NLSetI
2
PID101PID202
PIIC602
PIIC
606
PIR130
2PIR6401
PIR7001
NLVe Clamped1
PID901PID1002
PIIC
2202
PIIC2206
PIR1402
PIR8001
PIR8501
NLVe Clamped2
PIIC1801
PIIC1902
PIR6
702
NLVe
S1
PIIC1807
PIIC2003
PIR6
802
NLVe
S2
PIIC2307
PIIC2503
PIR8
602
NLVe
S3
PIIC2301
PIIC2402
PIR8
202
NLVe
S4
PIC1
802
PID602PIIC1701
PIIC1706
PIR5
901
PIR6
601
PIR6901
PIR7
901
PIR8401
NLVo
ffse
t
PIIC1903
PIIC2002
PIIC2601
PIR8
802
NLVr1
PIIC2403
PIIC2502
PIIC2607
PIR9
202
NLVr2
PIC802
PIIC
1007
PIR2101
PIR3102
NLVref
1
PIC4602
PIIC
1607
PIR4401
PIR6302
NLVr
ef2
PIIC1507
PIR5
302
PIR8
701
PIR9001
NLVt
ri
Figure A.3: Control schematic for the dc-dc converter circuit.
Stellenbosch University https://scholar.sun.ac.za
APPENDIX A. COMPLETE CIRCUIT SCHEMATICS 152
A.2 Inverter Circuit Schematics
All of the circuit diagrams used for the implementation of the bidirectionalcurrent-controller three-phase inverter prototype circuit are given in this sec-tion.
A.2.1 Power Stage Schematics
The schematics for the power stage of the inverter circuit are given in thissection. It includes the three-phase inverter topology, gate-drive circuitry anddc bus current sensor. The schematic is shown in Fig.A.4.
Stellenbosch University https://scholar.sun.ac.za
APPENDIX A. COMPLETE CIRCUIT SCHEMATICS 153
11
22
33
44
55
66
77
88
DD
CC
BB
AA
D1
BAT8
5
D2
BAT8
5
Q2
BC55
6Q
1BC
556
T110
:11
C1 100p
F
C2 100p
F
R1 0.22
K
R2 0.22
K
R3 1M
R4 0.47
K
C4 100u
FR5 10
ohm
+15V
C3 100n
F
R6 27K
C5 100n
F
SGN
D1
GN
D
D3
BAT8
5
D4
BAT8
5
Q4
BC55
6Q
3BC
556
T210
:11
C6 100p
F
C7 100p
F
R7 0.22
K
R8 0.22
K
R9 1M
C9 100u
F
+15V
C8 100n
F
C10
100n
F
GN
D
D5
BAT8
5
D6
BAT8
5
Q6
BC55
6Q
5BC
556
T310
:11
C11
100p
F
C12
100p
F
R13
0.22
K
R14
0.22
K
R15
1M
C14
100u
F
+15V
C13
100n
F
C15
100n
F
GN
D
D7
BAT8
5
D8
BAT8
5
Q8
BC55
6Q
7BC
556
T410
:11
C16
100p
F
C17
100p
F
R19
0.22
K
R20
0.22
K
R21
1M
C19
100u
F
+15V
C18
100n
F
C20
100n
F
GN
D
D9
BAT8
5
D10
BAT8
5
Q10
BC55
6Q
9BC
556
T510
:11
C21
100p
F
C22
100p
F
R25
0.22
K
R26
0.22
K
R27
1M
C24
100u
F
+15V
C23
100n
F
C25
100n
F
GN
D
D11
BAT8
5
D12
BAT8
5
Q12
BC55
6Q
11BC
556
T610
:11
C26
100p
F
C27
100p
F
R31
0.22
K
R32
0.22
K
R33
1M
C29
100u
F
+15V
C28
100n
F
C30
100n
F
GN
D
G1
G2
G3
G4
G5
G6
UV
W
EU
EV
EW
E1U
E1V
E1W
T1 T2PM
1FS
30R
06W
1E3_
B11
C31
100u
F
1 2 3
P5 3Pha
se
1 2
P4 DC
Bus
PWM
1
PWM
2
PWM
3
PWM
4
PWM
5
PWM
6
1 2
P6
Vbu
s Mon
itor
1 2
P7 12V
Sup
ply+1
2V 0V
R39
1K
GN
D
+5V
VCC
1
GN
D2
VIO
UT
3
IP+
4
IP-
5
H4
ACS
756x
CB
VCC
213
Vou
t11
DES
AT
14
VEE
12
VCL
AM
P10
VE
16
VLE
D15
AN
OD
E6
CATH
OD
E5
VCC
12
Vs
1
nFA
ULT
3
IC1
TLP5
214
C33
100n
F
+5V
R37
2K2
R38
0K1
C34
100p
F
PWM
1
D13
BYV
26E
C35
1nFnF
1 GN
D
+Vs
-Vs
+Vs
1 2 3
CON
1
PWM
1
C39
100n
F
1 2 3
CON
2
PWM
2 1 2 3
CON
3
PWM
3 1 2 3
CON
4
PWM
4
1 2 3
CON
5
PWM
51 2 3
CON
6
PWM
6
+5V
nF2 G
ND
SGN
D1
R10
0.47
K
R11
10oh
mR1
227
K
-Vs
VCC
213
Vou
t11
DES
AT
14
VEE
12
VCL
AM
P10
VE
16
VLE
D15
AN
OD
E6
CATH
OD
E5
VCC
12
Vs
1
nFA
ULT
3
IC2
TLP5
214
C40
100n
F
R40
2K2
R41
0K1
C41
100p
F
PWM
2
D14
BYV
26E
C42
1nF +5
V
nF3 G
ND
+Vs
R16
0.47
K
R17
10oh
mR1
827
K
SGN
D3
VCC
213
Vou
t11
DES
AT
14
VEE
12
VCL
AM
P10
VE
16
VLE
D15
AN
OD
E6
CATH
OD
E5
VCC
12
Vs
1
nFA
ULT
3
IC3
TLP5
214
C43
100n
F
R42
2K2
R43
0K1
C44
100p
F
PWM
3
D15
BYV
26E
C45
1nF +5
V
nF4 G
ND
SGN
D3
R22
0.47
K
R23
10oh
mR2
427
K
-Vs
VCC
213
Vou
t11
DES
AT
14
VEE
12
VCL
AM
P10
VE
16
VLE
D15
AN
OD
E6
CATH
OD
E5
VCC
12
Vs
1
nFA
ULT
3
IC4
TLP5
214
C46
100n
F
R44
2K2
R45
0K1
C47
100p
F
PWM
4
D16
BYV
26E
C48
1nF
+5V
nF5 G
ND
+Vs
R28
0.47
K
R29
10oh
mR3
027
K
SGN
D5
VCC
213
Vou
t11
DES
AT
14
VEE
12
VCL
AM
P10
VE
16
VLE
D15
AN
OD
E6
CATH
OD
E5
VCC
12
Vs
1
nFA
ULT
3
IC5
TLP5
214
C49
100n
F
R46
2K2
R47
0K1
C50
100p
F
PWM
5
D17
BYV
26E
C51
1nF+5
V
nF6 G
ND
SGN
D5
R34
0.47
K
R35
10oh
mR3
627
K
-Vs
VCC
213
Vou
t11
DES
AT
14
VEE
12
VCL
AM
P10
VE
16
VLE
D15
AN
OD
E6
CATH
OD
E5
VCC
12
Vs
1
nFA
ULT
3
IC6
TLP5
214
C52
100n
F
R48
2K2
R49
0K1
C53
100p
F
PWM
6
D18
BYV
26E
C54
1nF
GN
D
GN
D
GN
D
GN
D
GN
D
GN
D
Vin
-1
Vin
+2
Vou
t+3
COM
4
Vou
t-5
IC7
DK
E10A
-15
+15V
-15V
GN
D
C55
100u
F
C56
100u
F
C57
100n
F
C58
100n
F
IN1
2
OU
T3
GN
D
VS1
MC7
805C
T
1 2
P8 Tem
p Se
nsor
+15V
+5V
GN
D
C67
100n
FC6
810
0nF
C69
100u
F
SGN
D1
SGN
D3
SGN
D5
C32
100n
F
GN
D
1 2 3
P9 CS D
C B
us
D19
3V9
D20
3V9
D21
3V9
D22
3V9
D23
3V9
D24
3V9
PIC101
PIC1
02
COC1
PIC201
PIC2
02
COC2
PIC301PIC302COC
3
PIC401 PIC402CO
C4
PIC501PIC502 COC
5
PIC601
PIC6
02
COC6
PIC701
PIC7
02
COC7
PIC801PIC802COC
8
PIC901 PIC902CO
C9
PIC1001PIC1002COC10
PIC1101
PIC1
102
COC11
PIC1201
PIC1
202
COC12
PIC1301PIC1302COC13
PIC1401 PIC1402
COC14
PIC1501PIC1502COC15
PIC1601
PIC1
602
COC16
PIC1701
PIC1
702
COC17
PIC1801PIC1802COC18
PIC1901 PIC1902COC19
PIC2001PIC2002COC20
PIC2
101
PIC2102
COC21
PIC2201
PIC2202
COC22
PIC2301PIC2302COC23
PIC2401 PIC2402
COC24
PIC2501PIC2502COC25
PIC2
601
PIC2602
COC26
PIC2701
PIC2702
COC27
PIC2801PIC2802COC28
PIC2901 PIC2902
COC29
PIC3001PIC3002COC30PIC3101 PIC3102
COC31
PIC3201PIC3202COC32
PIC3301PIC3302COC33
PIC3401PIC3402COC34
PIC3501PIC3502COC35
PIC3
901
PIC390
2
COC39
PIC4001PIC4002COC40
PIC4101PIC4102COC41
PIC4201PIC4202COC42
PIC4301PIC4302COC43
PIC4401PIC4402COC44
PIC4501PIC4502COC45
PIC4601PIC4602COC46
PIC4701PIC4702COC47
PIC4801PIC4802COC48
PIC4901PIC4902COC49
PIC5001PIC5002COC50
PIC5101PIC5102COC51
PIC5201PIC5202COC52
PIC5301PIC5302COC53
PIC5401PIC5402COC54
PIC5501
PIC5
502
COC55
PIC560
1PI
C560
2COC5
6
PIC5701
PIC5
702
COC57
PIC580
1PI
C580
2COC58
PIC6701PIC6702COC67
PIC6801PIC6802COC68
PIC6901 PIC6902
COC69
PICO
N101
PICO
N102
PICO
N103
COCON1
PICO
N201
PICO
N202
PICO
N203
COCON2
PICO
N301
PICO
N302
PICO
N303
COCON3
PICO
N401
PICO
N402
PICO
N403
COCON4
PICON501
PICON502
PICON503
COCON5
PICON601
PICON602
PICON603
COCON6
PID101
PID102
COD1
PID201
PID202
COD2
PID301
PID302
COD3
PID401
PID402
COD4
PID501
PID502
COD5
PID601
PID602
COD6
PID701
PID702
COD7
PID801
PID802
COD8
PID901
PID902
COD9
PID1001
PID1002
COD10
PID1101
PID1102
COD11
PID1201
PID1202
COD12
PID1301
PID1302
COD13
PID1401
PID1402
COD14
PID1501
PID1502
COD15
PID1
601
PID1
602
COD16
PID1701
PID1702
COD17
PID1
801
PID1
802
COD18
PID1901
PID1902COD19 PID2001
PID2002COD20 PID2101
PID2102COD21 PI
D220
1PI
D220
2COD22
PID2301
PID2302COD23PI
D2401
PID2402COD24
PIH4
01
PIH402
PIH403
PIH4
04
PIH405CO
H4
PIIC101
PIIC
102
PIIC103
PIIC105
PIIC106
PIIC1010
PIIC1011
PIIC1012
PIIC1013
PIIC1014
PIIC1015
PIIC1016
COIC1
PIIC
201
PIIC
202
PIIC203
PIIC205
PIIC206
PIIC2010
PIIC2011
PIIC2012
PIIC2013
PIIC2014
PIIC2015
PIIC2016
COIC2
PIIC301
PIIC302
PIIC303
PIIC305
PIIC306
PIIC3010
PIIC3011
PIIC3012
PIIC3013
PIIC3014
PIIC3015
PIIC3016
COIC3
PIIC
401
PIIC402
PIIC403
PIIC
405
PIIC406
PIIC4010
PIIC4011
PIIC4012
PIIC4013
PIIC4014
PIIC4015
PIIC4016
COIC4
PIIC501
PIIC502
PIIC503
PIIC505
PIIC
506
PIIC
5010
PIIC
5011
PIIC
5012
PIIC
5013
PIIC
5014
PIIC
5015
PIIC
5016
COIC5
PIIC601
PIIC602
PIIC603
PIIC
605
PIIC606
PIIC
6010
PIIC
6011
PIIC
6012
PIIC
6013
PIIC
6014
PIIC
6015
PIIC
6016
COIC6
PIIC701
PIIC702
PIIC703
PIIC704
PIIC705
COIC7
PIM10E1U
PIM1
0E1V
PIM10E1W
PIM10EUPIM10EV
PIM10EW
PIM10G1
PIM10G2
PIM10G3
PIM10G4
PIM1
0G5
PIM1
0G6
PIM1
0P
PIM1
0T1
PIM1
0T2
PIM10U
PIM10V
PIM10W
COM1
PIP401
PIP402COP
4
PIP501
PIP502
PIP503COP
5
PIP601
PIP602
COP6
PIP701
PIP702
COP7
PIP801
PIP802
COP8
PIP901
PIP902
PIP903COP
9
PIQ101PIQ102
PIQ103COQ
1
PIQ201PIQ202
PIQ203 COQ2
PIQ301PIQ302
PIQ303CO
Q3
PIQ401PIQ402
PIQ403 COQ4
PIQ501PIQ502
PIQ503CO
Q5
PIQ601PIQ602
PIQ603 COQ6
PIQ701PIQ702
PIQ703CO
Q7
PIQ801PIQ802
PIQ803 COQ8
PIQ901PIQ902
PIQ903COQ
9
PIQ1001PIQ1002
PIQ1003 COQ10
PIQ1101PIQ1102
PIQ1103COQ11
PIQ1201PI
Q120
2
PIQ1203 COQ12
PIR101
PIR102
COR1
PIR201
PIR202
COR2
PIR301PIR302 COR3
PIR401 PIR402
COR4
PIR501
PIR5
02CO
R5
PIR601PIR602 COR6
PIR701
PIR702
COR7
PIR801
PIR802
COR8
PIR901PIR902 COR9
PIR1001 PIR1002
COR10
PIR110
1PI
R110
2COR11
PIR1201PIR1202COR
12
PIR130
1PIR
1302
COR13
PIR140
1PIR
1402
COR14
PIR1501PIR1502 COR15
PIR1601 PIR1602
COR16
PIR1701
PIR1
702
COR17
PIR1801PIR1802COR18
PIR1901
PIR1902
COR19
PIR2001
PIR2002
COR20
PIR2101PIR2102 COR21
PIR2201 PIR2202
COR22
PIR2301
PIR2
302
COR23
PIR2401PIR2402 COR24
PIR2
501
PIR2502
COR25
PIR2601
PIR2602
COR26
PIR2701PIR2702 COR27
PIR2801 PIR2802COR28
PIR2901
PIR2902
COR29
PIR3001PIR3002COR30
PIR3
101
PIR310
2COR31
PIR320
1PIR
3202
COR32
PIR3301PIR3302 COR33
PIR3401 PIR3402COR34
PIR3501
PIR3502
COR35
PIR3601PIR3602 COR36
PIR3701PIR3702COR37
PIR3
801
PIR3
802
COR38
PIR3901PIR3902 COR39
PIR4001PIR4002COR40
PIR4
101
PIR4
102
COR41
PIR4201PIR4202COR42
PIR4
301
PIR4
302
COR43
PIR4401PIR4402COR44
PIR450
1PIR
4502
COR45
PIR4601
PIR4602
COR46
PIR4
701
PIR4
702
COR47
PIR4801
PIR4802
COR48
PIR490
1PIR
4902
COR49
PIT101PIT102
PIT103PIT104
PIT105PIT106
PIT107PIT108 COT
1
PIT201PIT202
PIT203PIT204
PIT205PIT206
PIT207PIT208 COT
2
PIT301PIT302
PIT303PIT304
PIT305PIT306
PIT307PIT308 COT
3
PIT401PIT402
PIT403PIT404
PIT405PIT406
PIT407PIT408 COT
4
PIT501PIT502
PIT503PIT504
PIT505PIT506
PIT507PIT508 CO
T5
PIT601PIT602
PIT603PIT604
PIT605PIT606
PIT607PIT608 CO
T6
PIVS101
PIVS102
PIVS103
COVS1
PIC3302 PIC4002 PIC4302 PIC4602
PIC4902PIC5202
PIC6802PIC6901
PIIC
102
PIIC
202
PIIC302
PIIC402
PIIC502
PIIC602
PIM1
0T1
PIR3702 PIR4002 PIR4202 PIR4402
PIR4602
PIR4802
PIVS103
PIIC702
PIP701
PIC302 PIC802 PIC1302 PIC1802
PIC2302PIC2802
PIC5501
PIC580
1
PIC6702PIIC703
PIQ103PIQ203
PIQ303PIQ403
PIQ503PIQ603
PIQ703PIQ803
PIQ903PIQ1003
PIQ1103PIQ1203
PIVS101
PIC3101PIC3202
PID1901
PID2101
PID2301
PIH4
04
PIM1
0P
PIC5
602
PIC5701
PIIC705
PIC902PIC1001
PIC1902PIC2001
PIC2902PIC3001
PIC3102PIC3201
PIC4101 PIC4701
PIC5301
PIIC2012
PIIC2016
PIIC4012
PIIC4016
PIIC
6012
PIIC
6016
PIM10E1U
PIM1
0E1V
PIM10E1W
PIM10EUPIM10EV
PIM10EWPIP402
PIP602
PIR1201 PIR2401
PIR3601
PIT202PIT203
PIT402PIT403
PIT602PIT603
PIIC701
PIP702
PIC301 PIC801 PIC1301 PIC1801
PIC2301PIC2801
PIC3301PIC3501
PIC4001PIC4201
PIC4301PIC4501
PIC4601PIC4801
PIC4901PIC5101
PIC5201PIC5401
PIC5
502
PIC560
1
PIC5
702
PIC5
802
PIC6701PIC6801
PIC6902
PICO
N102
PICO
N202
PICO
N302
PICO
N402
PICON502
PICON602
PIIC101
PIIC105
PIIC
201
PIIC205
PIIC301
PIIC305
PIIC
401
PIIC
405
PIIC501
PIIC505
PIIC601
PIIC
605
PIIC704
PIP802
PIR301 PIR901 PIR1501 PIR2101
PIR2701PIR3301
PIR3901
PIT106PIT107
PIT206PIT207
PIT306PIT307
PIT406PIT407
PIT506PIT507
PIT606PIT607
PIVS102
PIC101
PIQ101
PIT105
PIC1
02
PIR201
PIR302PIC201
PIR102
PIC2
02
PIQ201
PIT108
PIC401PIC502
PID102
PID202
PIIC1013
PIC601
PIQ301
PIT205
PIC6
02
PIR801
PIR902PIC701
PIR702
PIC7
02
PIQ401
PIT208
PIC901PIC1002
PID302
PID402
PIIC2013
PIC1101
PIQ501
PIT305
PIC1
102
PIR140
1
PIR1502PIC1201
PIR130
2
PIC1
202
PIQ601
PIT308
PIC1401PIC1502
PID502
PID602
PIIC3013
PIC1601
PIQ701
PIT405
PIC1
602
PIR2001
PIR2102PIC1701
PIR1902
PIC1
702
PIQ801
PIT408
PIC1901PIC2002
PID702
PID802
PIIC4013
PIC2
101
PIQ901
PIT505
PIC2102
PIR2601
PIR2702PIC2201
PIR2502
PIC2202
PIQ1001
PIT508
PIC2401PIC2502
PID902
PID1002
PIIC
5013
PIC2
601
PIQ1101
PIT605
PIC2602
PIR320
1
PIR3302PIC2701
PIR310
2
PIC2702
PIQ1201
PIT608
PIC2901PIC3002
PID1102
PID1202
PIIC
6013
PIC3402PIIC1014
PIR3
801
PIC3
901
PIH4
01PIP901
PIC390
2
PIH402
PIP902
PIC4102PIIC2014
PIR4
101
PIC4402PIIC3014
PIR4
301
PIC4702PIIC4014
PIR450
1
PIC5002PI
IC50
14PI
R470
1
PIC5302PI
IC60
14PIR
4901
PICO
N101
PIR401
PICO
N201
PIR1001
PICO
N301
PIR1601
PICO
N401
PIR2201
PICON501
PIR2801
PICON601
PIR3401
PID101
PIT104 PID201
PIT101
PID301
PIT204 PID401
PIT201
PID501
PIT304 PID601
PIT301
PID701
PIT404 PID801
PIT401
PID901
PIT504 PID1001
PIT501
PID1101
PIT604 PID1201
PIT601
PID1301
PIR3
802
PID1302PID1902
PID1401
PIR4
102
PID1402PID2002
PID1501
PIR4
302
PID1502PID2102
PID1
601
PIR450
2PI
D160
2PID2
202
PID1701
PIR4
702
PID1702PID2302
PID1
801
PIR490
2PI
D180
2PID2402
PIH403
PIP903
PIH405
PIP401
PIP601
PIIC106
PIR402
PIIC1011
PIR501
PIR602PIIC1015
PIIC206
PIR1002
PIIC2011
PIR110
1
PIR1202PIIC2015
PIIC306
PIR1602
PIIC3011
PIR1701
PIR1802PIIC3015
PIIC406
PIR2202
PIIC4011
PIR2301
PIR2402PIIC4015
PIIC
506
PIR2802
PIIC
5011
PIR2901
PIR3002PI
IC50
15
PIIC606
PIR3402
PIIC
6011
PIR3501
PIR3602PI
IC60
15
PIM1
0T2
PIP801
PIR3902
PIQ102PI
R101
PIQ202
PIR202
PIQ302PI
R701
PIQ402
PIR802
PIQ502PI
R1301
PIQ602
PIR140
2
PIQ702PI
R1901
PIQ802
PIR2002
PIQ902PIR2
501
PIQ1002
PIR2602
PIQ1102PIR3
101
PIQ1
202
PIR320
2
PIC3502
PICO
N103
PIIC103
PIR3701NLnF1 PIC4202
PICO
N203
PIIC203
PIR4001NLnF2 PIC4502
PICO
N303
PIIC303
PIR4201NLnF3 PIC4802
PICO
N403
PIIC403
PIR4401NLnF4
PIC5102
PICON503
PIIC503
PIR4601
NLnF5PIC5402
PICON603
PIIC603
PIR4801
NLnF6
PIIC1010
PIM10G1
PIR5
02
NLPWM1
PIIC2010
PIM10G2
PIR1
102
NLPWM2
PIIC3010
PIM10G3
PIR1
702
NLPWM3
PIIC4010
PIM10G4
PIR2
302
NLPWM4
PIIC
5010
PIM1
0G5
PIR2902
NLPWM5
PIIC
6010
PIM1
0G6
PIR3502
NLPWM6
PIC402PIC501
PIC3401
PID2001
PIIC1012
PIIC1016
PIM10U
PIP501
PIR601
PIT102PIT103
PIC1402PIC1501
PIC4401
PID2
201
PIIC3012
PIIC3016
PIM10V
PIP502
PIR1801
PIT302PIT303
PIC2402PIC2501
PIC5001
PID2401
PIIC
5012
PIIC
5016
PIM10W
PIP503
PIR3001
PIT502PIT503
Figure A.4: Schematic for the power stage of the inverter circuit.
Stellenbosch University https://scholar.sun.ac.za
APPENDIX A. COMPLETE CIRCUIT SCHEMATICS 154
A.2.2 Control Stage Schematics
The complete control schematics for the three-phase inverter circuit are givenin this section. The schematic containing the dsPIC, DAC and magnetic pick-up interface circuit is shown in Fig. A.5. The schematic for the analog part ofthe control circuit is shown in Fig. A.6.
Stellenbosch University https://scholar.sun.ac.za
APPENDIX A. COMPLETE CIRCUIT SCHEMATICS 155
11
22
33
44
DD
CC
BB
AA
1 2 3
P1 Mag
netic
Pic
kup
1
3
BA
L/ST
B6
BA
L5
V-
42
7
V+
8IC
1
LM31
1P
GN
D
R1
3K3
R2
3K3
R3
10K
R4
10K
GN
D
+15V
-15V
C1
100n
F
C2
100n
F
GN
D
+5V
R5
1K8
R6
22K
R7
10K
R8
1KG
ND
MC
LR1
AN
0/V
REF
+/C
N2/
RB
02
AN
1/V
REF
-/CN
3/R
B1
3
AN
2/SS
1/LV
DIN
/CN
4/R
B2
4
AN
3/C
N5/
RB
35
AN
4/IC
7/C
N6/
RB
46
AN
5/IC
8/C
N7/
RB
57
PGC
/EM
UC
/AN
6/O
CFA
/RB
68
PGD
/EM
UD
/AN
7/R
B7
9
AN
8/R
B8
10
VD
D11
VSS
12
OSC
1/C
LKI
13
OSC
2/C
LKO
/RC
1514
EMU
D1/
SOSC
I/T2C
K/U
1ATX
/CN
1/R
C13
15
EMU
C1/
SOSC
O/T
1CK
/U1A
RX
/CN
0/R
C14
16
INT0
/RA
1117
IC2/
INT2
/RD
918
OC
4/R
D3
19
VSS
20
VD
D21
OC
3/R
D2
22
IC1/
INT1
/RD
823
EMU
C3/
SCK
1/R
F624
EMU
D3/
U1T
X/S
DO
1/SC
L/R
F325
U1R
X/S
DI1
/SD
A/R
F226
U2T
X/C
N18
/RF5
27U
2RX
/CN
17/R
F428
C1T
X/R
F129
C1R
X/R
F030
VSS
31
VD
D32
EMU
D2/
OC
2/R
D1
33EM
UC
2/O
C1/
RD
034
AN
12/C
OFS
/RB
1235
AN
11/C
SDO
/RB
1136
AN
10/C
SDI/R
B10
37A
N9/
CSC
K/R
B9
38
AV
SS39
AV
DD
40
IC3
DSP
IC30
F401
3-30
I/P
VO
UTB
1V
OU
TA2
VSS
3
VR
EFA
4
VR
EFB
5
GN
D6
LDA
C7
RS
8
CS
9
CLK
10
SDI
11
VR
EFD
12V
REF
C13
VD
D14
VO
UTD
15V
OU
TC16
IC4
AD
7398
BR
Z
IN2
TEM
P3
GN
D4
TRIM
5
OU
T6
NC
7N
C8
NC
1
IC5
REF
02A
P
+15V
GN
DC3
100n
FC
410
0nF
C6
1uF
+G
ND
C5
100n
F8 5
2 3
4
6
7
1
IC2
OP0
7CP
5VP
C7
100n
F
C8
100n
F
GN
D
5VP
5VP
SDI
SDI
SCK
SS
SS
SCK
GN
D
MP
MP
Tp
1 2
X1
10M
HZ
C9
22pF
C10
22pF
DG
ND
IN1
2
OU
T3
GN
D
VS
MC
7805
CT
+15V
+5V
5VP
GN
D
+5V
DG
ND
GN
D
C11
100n
FC
1210
0nF
Ioz5V
P GN
D
PWM
+5V
R9
0K22
R11
0K22
R12
0K22
R13
0K22
DG
ND
DG
ND
D2
D3
D4
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
P2 Hea
der 1
6
+15V
-15VG
ND
Ia Ib Ic
PWM
IcIbIaTp Ioz
D5
DG
ND
DG
ND
CN
Err
NO
CN
Err
NO
TEM
P
TEM
P
PIC1
01PI
C102
COC1
PIC2
01PI
C202
COC2
PIC301PIC302COC
3PIC401PIC402
COC4
PIC501PIC502COC5
PIC601PIC602COC
6PIC701PIC702
COC7
PIC801PIC802COC
8
PIC901
PIC9
02
COC9
PIC1
001
PIC1002
COC1
0
PIC1101PIC1102CO
C11
PIC1201PIC1202CO
C12
PID2
01PI
D202
COD2
PID3
01PI
D302
COD3
PID4
01PI
D402
COD4
PID5
01PI
D502
COD5
PIIC101
PIIC102
PIIC103
PIIC104
PIIC105
PIIC106
PIIC107
PIIC108
COIC1
PIIC201
PIIC202
PIIC203
PIIC204PIIC205
PIIC206
PIIC207PIIC208
COIC2
PIIC301
PIIC302
PIIC303
PIIC304
PIIC305
PIIC306
PIIC307
PIIC308
PIIC309
PIIC3010
PIIC3011
PIIC3012
PIIC3013
PIIC3014
PIIC3015
PIIC3016
PIIC3017
PIIC3018
PIIC3019
PIIC3020
PIIC3021
PIIC3022
PIIC3023
PIIC3024
PIIC3025
PIIC3026
PIIC3027
PIIC3028
PIIC3029
PIIC3030
PIIC3031
PIIC3032
PIIC3033
PIIC3034
PIIC3035
PIIC3036
PIIC3037
PIIC3038
PIIC3039
PIIC3040CO
IC3
PIIC401
PIIC402
PIIC403
PIIC404
PIIC405
PIIC406
PIIC407
PIIC408
PIIC409
PIIC
4010
PIIC4011
PIIC
4012
PIIC4013
PIIC4014
PIIC4015
PIIC4016
COIC4
PIIC501
PIIC502
PIIC503
PIIC504
PIIC505
PIIC506
PIIC507
PIIC508
COIC5
PIP101
PIP102
PIP103
COP1
PIP201
PIP202
PIP203
PIP204
PIP205
PIP206
PIP207
PIP208
PIP209
PIP2010
PIP2011
PIP2012
PIP2013
PIP2014
PIP2015
PIP2016
COP2
PIR1
01PI
R102
COR1
PIR2
01PI
R202
COR2
PIR3
01PI
R302
COR3
PIR401PIR402 COR4
PIR501PIR502COR
5
PIR601PIR602COR
6
PIR701PIR702COR
7
PIR801PIR802COR
8
PIR9
01PI
R902
COR9
PIR110
1PIR
1102
COR1
1
PIR1
201
PIR1
202
COR1
2
PIR1
301
PIR1
302
COR1
3
PIVS01
PIVS02
PIVS03
COVS
PIX101 PIX102COX1
PIC502
PIC1202
PIIC105
PIIC106
PIIC108
PIIC301
PIIC3011
PIIC3021
PIIC3032
PIR502PIR702
PIVS03
PIC1
01
PIC302
PIC1102
PIIC207
PIIC502
PIP201
PIP202
PIVS01
PIC2
01
PIIC204
PIP205
PIP206
PIC402
PIC602PIC702
PIC802
PIIC302
PIIC3040
PIIC404
PIIC405
PIIC408
PIIC4013
PIIC4014
PIIC506
NL5VP
PIIC307
PIP209
NLCN
PIIC308
PIP2010
NLErr
PIC1
02
PIC2
02
PIC301PIC401
PIC501
PIC601PIC701
PIC801
PIC901
PIC1
001
PIC1101PIC1201
PID2
02
PID3
02
PID4
02
PID5
02
PIIC101
PIIC104
PIIC303
PIIC3012
PIIC3020
PIIC3026
PIIC3031
PIIC3039
PIIC403
PIIC406
PIIC407
PIIC
4012
PIIC504
PIP102
PIP203
PIP204
PIR401
PIR801
PIVS02
PIIC402
PIP2013
NLIaPIIC401
PIP2014
NLIb
PIIC4016
PIP2015
NLIc
PIIC306
PIP208
NLIoz
PIIC107
PIIC3023
PIR501 PIR602
NLMP
PIC9
02PIIC3013
PIX101
PIC1002
PIIC3014
PIX102
PID2
01PI
R902 PID3
01PI
R120
2
PID4
01PIR
1102
PID5
01PI
R130
1
PIIC102
PIR601
PIR701 PIR802
PIIC103
PIIC206
PIR3
02
PIIC201
PIIC202
PIR1
02
PIR3
01 PIIC203
PIR2
02 PIR402PIIC205
PIIC208
PIIC304
PIIC3010
PIR9
01
PIIC3016
PIR1
201
PIIC3017
PIIC3018
PIIC3027
PIIC3028
PIR1
302
PIIC3029
PIIC3030
PIIC3033
PIIC3034
PIIC3035
PIR110
1
PIIC3036
PIIC3037
PIIC3038
PIIC4015
PIIC501
PIIC503
PIIC505
PIIC507
PIIC508
PIP101
PIR1
01
PIP103
PIR2
01
PIIC309
PIP2011
NLNO
PIIC3019
PIP2016
NLPW
M
PIIC3024
PIIC
4010
NLSC
K
PIIC3025
PIIC4011
NLSD
IPIIC3022
PIIC409
NLSS
PIIC3015
PIP2012
NLTE
MP
PIIC305
PIP207
NLTp
Figure A.5: Schematic for the inverter digital control circuit.
Stellenbosch University https://scholar.sun.ac.za
APPENDIX A. COMPLETE CIRCUIT SCHEMATICS 156
11
22
33
44
55
66
77
88
DD
CC
BB
AA
12
Y1
2MH
zR1 10
MC1 22
pFC2 22
pF
RST
1
CLK
9
D0
3Q
02
D1
4Q
15
D2
6Q
27
D3
11Q
310
D4
13Q
412
D5
14Q
515
VCC
16
GN
D8
IC2
M74
HC1
74B
1R
RST
1
CLK
9
D0
3Q
02
D1
4Q
15
D2
6Q
27
D3
11Q
310
D4
13Q
412
D5
14Q
515
VCC
16
GN
D8
IC3
M74
HC1
74B
1R
R2 1KR3 1K
C4 100n
F
C6 100n
F
C3 100u
FC5 10
0uF
+5V
56
IC1C
M74
HCT
14B
1110
IC1E
M74
HCT
14B
+5V
+5V
+5V
12
IC1A
M74
HCT
14B
34
IC1B
M74
HCT
14B
98
IC1D
M74
HCT
14B
2 31
A
8 4
IC4A
LF35
3N56
7B
8 4
IC4B
LF35
3N
R5 12K
R4 12K
R7 330KR6 33
0K
R8 12K
R9 12K
1312
IC1F
M74
HCT
14B
+15V
+15V
-15V
-15V
C9 100n
F
C8 1nF
C7 1nF
GN
DG
ND
GN
D
Vtri
GN
DG
ND
GN
D
1 2
P1 12V
Sup
ply+1
2V 0VV
in-
1
Vin
+2
Vou
t+3
COM
4
Vou
t-5
IC5
DK
E10A
-15
+15V
-15VC1
210
0uF
C13
100u
F
C14
100n
F
C11
100n
FIN
1
2
OU
T3
GN
D
VS1
MC7
805C
T+1
5V+5
V
C15
100n
FC1
610
0nF
C17
100u
F
GN
DG
ND
C10
100n
F
+5V
GN
D
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
P2
Hea
der 1
6
Ia Ib Ic PWM
Ioz
Tp
GN
D
+15V
-15V
8
1
4
32IC
6ALF
353N
8 4
756
IC6B
LF35
3N
8 4
756
IC8B
LF35
3N
8
1
4
32IC
12A
LF35
3N
8 4
756
IC28
BLF
353N
8
1
4
32IC
8ALF
353N
+15V
+15V
+15V
-15V
-15V
-15V
IN2
TEM
P3
GN
D4
TRIM
5
OU
T6
NC
7N
C8
NC
1
VR1
REF0
2AP
C30
100n
FC3
110
0nF
+15V
GN
DG
ND
+5V
P
Vae
Vbe
Vce
GN
D
R27
120K
R29
120K
R30
470K
R31
22K
R32
22K
R33
470K
C33 1u
F
C34
1uF
+15V
-15V
GN
D
+15V
-15V
R34
10K
GN
DR35
33K
+5V
P
Dea
dT
567
B8 4
IC23
BLF
353N
2 31
A8 4
IC23
ALF
353N
+15V
+15V
-15V
-15V
R80
4K7
R81
4K7
R79
4K7
R82
4K7
R86
4K7
R84
4K7
R85
4K7
R83
4K7
GN
DV
eS2
VeS
1
Vae
Vae
1
23
4
5
6
7
8 IC24
AD
790J
N
1
23
4
5
6
7
8 IC25
AD
790J
N
GN
D+1
5V
+15V
-15V
-15V+5
V
+5V
C58
100n
F
C57
100n
F
C56
100n
F
C60
100n
FG
ND
GN
D
GN
D
GN
DV
tri
Vtri
GN
D
GN
D
Dea
dT
1 23
IC14
A
SN74
HC0
8N
4 56
IC14
B
SN74
HC0
8N
89 10
IC14
C
SN74
HC0
8N
12 1311
IC14
D
SN74
HC0
8N
1 23
IC15
A
SN74
HC0
8N
4 56
IC15
B
SN74
HC0
8N
89 10
IC15
C
SN74
HC0
8N
12 1311
IC15
D
SN74
HC0
8N
1 2 4 5
6
IC16
A
SN74
HC2
1N
9 10 12 13
8
IC16
B
SN74
HC2
1N
8
1
4
32IC
17A
LF35
3N
8 4
756
IC17
BLF
353N
R36
1K
+5V
P
D7
BAT8
5D
8BA
T85
+15V
+15V
-15V
-15V
Ioz
GN
D
1 2 3
P7 PWM
1
1 2 3
P8 PWM
2
nF nF
Vsa
t1
567
B8 4
IC18
BLF
353N
2 31
A8 4
IC18
ALF
353N
+15V
+15V
-15V
-15V
R38
4K7
R39
4K7
R37
4K7
R40
4K7
R44
4K7
R42
4K7
R43
4K7
R41
4K7
GN
DV
eS4
VeS
3
Vbe
Vbe
1
23
4
5
6
7
8 IC19
AD
790J
N
1
23
4
5
6
7
8 IC20
AD
790J
N
GN
D+1
5V
+15V
-15V
-15V+5
V
+5V
C37
100n
F
C36
100n
F
C35
100n
F
C39
100n
FG
ND
GN
D
GN
D
GN
DV
tri
Vtri
GN
D
GN
D
Dea
dT
1 2 3
P9 PWM
3
1 2 3
P10
PWM
4
nF nF
567
B8 4
IC26
BLF
353N
2 31
A8 4
IC26
ALF
353N
+15V
+15V
-15V
-15V
R46
4K7
R47
4K7
R45
4K7
R48
4K7
R52
4K7
R50
4K7
R51
4K7
R49
4K7
GN
DV
eS6
VeS
5
Vce
Vce
1
23
4
5
6
7
8 IC21
AD
790J
N
1
23
4
5
6
7
8 IC22
AD
790J
N
GN
D+1
5V
+15V
-15V
-15V+5
V
+5V
C43
100n
F
C42
100n
F
C41
100n
F
C45
100n
FG
ND
GN
D
GN
D
GN
DV
tri
Vtri
GN
D
GN
D
Dea
dT
1 2 3
P11
PWM
5
1 2 3
P12
PWM
6
nF nF
Vsa
t2
Vsa
t3
Vsa
t4
Vsa
t5
Vsa
t6
Vsa
t1
Vsa
t2
Vsa
t3
Vsa
t4
Vsa
t5V
sat6
Vbu
s
PWM
Tem
pnF
Tem
p
81
432
IC27
ALF
353N
8 4
756
IC27
BLF
353N
R53
1K
+5V
P
D9
BAT8
5D
10BA
T85
+15V
+15V
-15V
-15V
+15V
-15V
+5V
P
8 4
756
IC12
BLF
353N
8
1
4
32IC
28A
LF35
3N
+15V
-15V
R54
8K2
R55
56K
R56
390K
R57
390K
R58
56K
R59
8K2
GN
D
12
P13
Vba
t
12
P14
Vbu
s
GN
D
1 2
P15
Vte
mp
+15V
-15V
GN
D
+15V
-15V
+5V
C49
100n
FC4
810
0nF
C47 10
0nF
GN
D
8 4
756
IC30
BLF
353N
+15V
-15V
R62
560K
R63
39K
R64
8K2
R65
560K
R66
39K
R67
8K2
2 31
A8 4
IC30
ALF
353N
GN
D
+15V
-15V
+5V
P
IN2
VTE
MP
3
GN
D4
TRIM
5O
UT
6N
C8
NC
7N
C1
IC31
REF0
3GP
+5V
GN
D
Vof
faR6
810
K
R71
4k7
R72
4K7
R73
4K7
R74
4K7
R75
4K7
R76
4K7
R77
4K7
R78
4K7
R87
4K7
R88
4K7
R89
4K7
R90
4K7
GN
D
GN
D
GN
D
Ia Ib Ic
C50
100n
F
C51
100n
F
C52
100n
F
C53
100n
F
C54
100n
F
C6210
0nF
C63
100n
F
C65
100n
F
GN
D
+5V
C55
100n
F
GN
DGN
D
GN
DGN
D
GN
D
GN
D
C66
100n
F
1
3
BAL/
STB
6
BAL
5
V-
42
7
V+
8IC
29
LM31
1P
R69
1K2
R70
22K
R91
10K
Vbu
s
GN
D
+15V
-15V
GN
DG
ND
-15V
R92
1KR9
3
330K
R94
1KR9
5
330K
R96
1KR9
7
330K
R98
330K
R99
1K R100
1KR1
01
330K
R102
1KR1
03
330K
-15V
-15V
C44
100n
F
GN
D
+5V
Err
Err
C64
1nF
C71 1n
F
C70
1nF
C67 1n
F
C68
1nF
C69 1n
F
R10
1K
GN
D
C18
100n
F
C20
100n
F
C19
100n
F
C21
100n
F
+15V
GN
D
R11
4K7
R12
4K7
R13
4K7
R14
4K7
R15
4K7
R16
4K7
C22
100n
F
C23
100n
FC2
410
0uF
C25
100u
F
R24
4K7
R25
4K7
R26
4K7
R23
4K7
R17
10K
R18
10K
R19
10K
R20
22K
R21
22K
R22
18K
1 2 3
P6 CS F
B
8
1
4
32IC
9ALF
353N
8 4
756
IC9B
LF35
3N
8
1
4
32IC
10A
LF35
3N
8 4
756
IC10
BLF
353N
+15V
+15V
+15V
+15V
-15V
-15V
-15V
-15V
R61
10K
R60
4K7
R104
10K
R105
3K3
R106
3K3
R107
33K
R108
33K
R109
3K3
R110
3K3
C26
100n
F
C27
100n
F
GN
D
+5V
+5V
P
GN
D
GN
D
Idcf
b
Idcf
b
GN
D
GN
D
Idc
C28
100n
F
GN
D
C29
100n
F
GN
D+5V
P
+5V
P
PIC101PIC102CO
C1PIC201PIC202
COC2
PIC301 PIC302CO
C3PIC401PIC402
COC4
PIC501 PIC502
COC5
PIC601PIC602COC
6
PIC701PIC702CO
C7
PIC801
PIC8
02
COC8
PIC9
01PIC902
COC9
PIC1001PIC1002COC10
PIC1101
PIC1
102
COC11
PIC1201
PIC1
202
COC12
PIC1301
PIC1
302
COC13
PIC1401
PIC1
402
COC14
PIC1501PIC1502COC15
PIC1601PIC1602COC16
PIC1701 PIC1702
COC17
PIC1801PIC1802COC18
PIC1901PIC1902COC19
PIC2001
PIC2002
COC20
PIC2101PIC2102COC21
PIC2
201
PIC2202
COC22
PIC2301PIC2302COC23
PIC2401 PIC2402
COC24
PIC2
501
PIC2502COC25
PIC2601
PIC2
602
COC26
PIC2701PIC2702COC27
PIC2801PIC2802COC
28
PIC2901PIC2902COC29
PIC3001PIC3002COC30
PIC3101PIC3102COC
31
PIC3301
PIC3
302
COC33
PIC3401PIC3402COC34
PIC3501
PIC3502
COC35
PIC360
1PIC
3602
COC36
PIC3701 PIC3702COC37
PIC3901
PIC3902
COC39
PIC4101
PIC4102
COC41
PIC4201
PIC4202
COC42
PIC4301 PIC4302COC43
PIC4401PIC4402COC44
PIC4501
PIC4502
COC45
PIC4701
PIC4
702
COC47
PIC4801PIC4802COC48
PIC490
1PIC
4902
COC49
PIC5001PIC5002COC50
PIC5101
PIC5
102
COC51
PIC5201 PIC5202COC52
PIC5301
PIC5
302
COC53
PIC5401
PIC5402
COC54
PIC5501PIC5502COC55
PIC5601
PIC5602
COC56
PIC5701
PIC5702
COC57
PIC5801 PIC5802COC
58
PIC6001
PIC6002
COC60
PIC6201
PIC6202 COC62
PIC6
301
PIC6302
COC63
PIC6401
PIC6
402
COC64
PIC650
1PI
C650
2
COC65
PIC6601 PIC6602COC66
PIC6701PIC6702COC67
PIC6
801
PIC680
2COC68
PIC6901PIC6902COC69
PIC700
1PIC
7002
COC70
PIC7101PIC7102COC71
PID701 PID702COD
7
PID801PID802COD
8
PID901 PID902CO
D9
PID1001PID1002COD10
PIIC101
PIIC102
COIC1A
PIIC103
PIIC104
COIC1B
PIIC105
PIIC106
COIC1C
PIIC108
PIIC109
COIC1D
PIIC
1010
PIIC1011
COIC1E
PIIC1012
PIIC
1013
COIC1F
PIIC201
PIIC
202
PIIC
203
PIIC204
PIIC205
PIIC206
PIIC207
PIIC
208
PIIC209
PIIC2010
PIIC2011
PIIC2012
PIIC2013
PIIC2014
PIIC2015
PIIC2016
COIC2
PIIC301
PIIC
302
PIIC303
PIIC304
PIIC305
PIIC306
PIIC307
PIIC308
PIIC309
PIIC3010
PIIC
3011
PIIC3012
PIIC
3013
PIIC
3014
PIIC3015
PIIC3016
COIC3
PIIC401
PIIC402
PIIC403
PIIC404PIIC408 COIC
4A
PIIC404PIIC405
PIIC406
PIIC407
PIIC408
COIC
4B
PIIC501
PIIC502
PIIC503
PIIC504
PIIC505
COIC5
PIIC601
PIIC602
PIIC603
PIIC604PIIC608 COI
C6A
PIIC604PIIC605
PIIC606
PIIC607
PIIC608
COIC
6B
PIIC
801
PIIC802
PIIC803
PIIC804PIIC808 COI
C8A
PIIC804PIIC805
PIIC806
PIIC
807
PIIC808COIC
8B
PIIC
901
PIIC902
PIIC903
PIIC904PIIC908 COIC9A
PIIC904PI
IC90
5
PIIC906
PIIC907
PIIC908COIC9B
PIIC1001
PIIC
1002
PIIC
1003
PIIC1004PIIC1008CO
IC10
A
PIIC1004PIIC1005
PIIC1006
PIIC
1007
PIIC1008CO
IC10
B
PIIC
1201
PIIC1202
PIIC1203
PIIC1204PIIC1208 COIC1
2A
PIIC1204PIIC1205
PIIC1206
PIIC
1207
PIIC1208
COIC
12B
PIIC
1401
PIIC
1402
PIIC1403
COIC
14A
PIIC
1404
PIIC
1405
PIIC1406
COIC
14B
PIIC1408
PIIC
1409
PIIC
1401
0 COIC
14C
PIIC
1401
1
PIIC
1401
2
PIIC
1401
3 COIC
14D
PIIC1501
PIIC1502
PIIC
1503
COIC
15A
PIIC
1504
PIIC
1505
PIIC1506
COIC
15B
PIIC1508
PIIC
1509
PIIC
1501
0
COIC
15C
PIIC
1501
1
PIIC
1501
2
PIIC
1501
3
COIC
15D
PIIC1601
PIIC1602
PIIC1604
PIIC1605
PIIC1606
COIC
16A
PIIC
1608
PIIC1609
PIIC
1601
0
PIIC
1601
2
PIIC
1601
3
COIC
16B
PIIC1701
PIIC
1702
PIIC
1703
PIIC1704PIIC1708 COIC
17A
PIIC1704PIIC1705
PIIC1706
PIIC
1707
PIIC1708CO
IC17
B
PIIC1801
PIIC1802
PIIC1803
PIIC1804PIIC1808 COIC
18A
PIIC1804PIIC1805
PIIC1806
PIIC1807
PIIC1808COIC
18B
PIIC1901
PIIC1902
PIIC1903
PIIC1904
PIIC1905
PIIC1906PIIC
1907
PIIC1908 COIC
19
PIIC2001
PIIC2002
PIIC2003
PIIC2004
PIIC2005
PIIC2006PIIC
2007
PIIC2008 COIC20
PIIC2101
PIIC2102
PIIC2103
PIIC2104
PIIC2105
PIIC2106PIIC
2107
PIIC2108 COIC21
PIIC2201
PIIC2202
PIIC2203
PIIC2204
PIIC2205
PIIC2206PIIC
2207
PIIC2208 COIC22
PIIC2301
PIIC2302
PIIC2303
PIIC2304PIIC2308 COIC
23A
PIIC2304PIIC2305
PIIC2306
PIIC2307
PIIC2308COIC
23B
PIIC2401
PIIC2402
PIIC2403
PIIC2404
PIIC2405
PIIC2406PIIC
2407
PIIC2408 COIC
24
PIIC2501
PIIC2502
PIIC2503
PIIC2504
PIIC2505
PIIC2506PIIC
2507
PIIC2508 COIC
25
PIIC2601
PIIC2602
PIIC2603
PIIC2604PIIC2608 COIC
26A
PIIC2604PIIC2605
PIIC2606
PIIC2607
PIIC2608COIC
26B
PIIC2701
PIIC
2702
PIIC
2703
PIIC2704PIIC2708 COIC2
7A
PIIC2704PI
IC27
05
PIIC
2706
PIIC
2707
PIIC2708CO
IC27
BPIIC2801
PIIC2802
PIIC2803
PIIC2804PIIC2808 COIC
28A
PIIC2804PI
IC28
05
PIIC
2806
PIIC2807
PIIC2808CO
IC28
B
PIIC2901
PIIC2902
PIIC2903
PIIC2904
PIIC2905
PIIC2906
PIIC2907
PIIC2908
COIC29
PIIC3001
PIIC3002
PIIC3003
PIIC3004PIIC3008 COIC
30A
PIIC3004PI
IC30
05
PIIC
3006
PIIC3007
PIIC3008
COIC
30B
PIIC
3101
PIIC
3102
PIIC3103
PIIC
3104
PIIC3105
PIIC3106
PIIC
3107
PIIC3108
COIC
31
PIP101
PIP102
COP1
PIP201
PIP202
PIP203
PIP204
PIP205
PIP206
PIP207
PIP208
PIP209
PIP2010
PIP2011
PIP2012
PIP2013
PIP2014
PIP2
015
PIP2016
COP2
PIP601
PIP602
PIP603
COP6
PIP701
PIP702
PIP703COP
7
PIP801
PIP802
PIP803COP
8
PIP901
PIP902
PIP903COP
9
PIP1001
PIP1002
PIP1
003COP10
PIP1101
PIP1102
PIP1
103COP11
PIP1201
PIP1202
PIP1203COP12
PIP1301
PIP1302
COP13
PIP1401
PIP1
402
COP14
PIP1501
PIP1
502COP
15
PIR101
PIR102
COR1
PIR201PIR202 COR2
PIR301PIR302 COR3
PIR401
PIR402
COR4
PIR501
PIR502
COR5
PIR601
PIR602
COR6
PIR701PIR702COR
7
PIR801
PIR802
COR8
PIR9
01PIR902
COR9
PIR1
001
PIR1
002
COR10
PIR1
101
PIR1
102
COR11
PIR120
1PIR
1202
COR12
PIR1
301
PIR1
302
COR13
PIR140
1PIR
1402
COR14
PIR1
501
PIR1
502
COR15
PIR1
601
PIR1
602
COR16
PIR1
701
PIR170
2COR17
PIR1801PIR1802COR18
PIR1
901
PIR190
2COR19
PIR2
001
PIR200
2COR20
PIR2
101
PIR2
102
COR21
PIR220
1PI
R220
2COR22
PIR2
301
PIR2
302
COR23
PIR240
1PI
R240
2COR24
PIR2501PIR2502 COR25PI
R260
1PI
R260
2COR26
PIR2
701
PIR2
702
COR27
PIR290
1PI
R290
2COR29
PIR3
001
PIR3
002
COR30
PIR3
101
PIR3
102
COR31
PIR3
201
PIR3
202
COR32
PIR3301PIR3302 COR33
PIR3401 PIR3
402
PIR3403COR34
PIR3501PIR3502 COR35
PIR3
601
PIR3
602
COR36
PIR3
701
PIR3702
COR37
PIR3
801
PIR3802
COR38 PI
R3901
PIR3
902COR39
PIR4
001
PIR4002
COR40
PIR4
101
PIR4102
COR41
PIR4
201
PIR4202
COR42
PIR4
301
PIR430
2COR43
PIR4
401
PIR4402
COR44
PIR4
501
PIR4502
COR45
PIR4
601
PIR4602
COR46 PIR
4701
PIR4
702COR47
PIR4
801
PIR4802
COR48
PIR4
901
PIR4902
COR49
PIR5
001
PIR5002
COR50
PIR5
101
PIR5102
COR51
PIR5
201
PIR5202
COR52
PIR5
301
PIR5
302
COR53
PIR5
401
PIR5
402
COR54
PIR5
501
PIR5
502
COR55
PIR5
601
PIR5
602
COR56
PIR570
1PIR
5702
COR57
PIR580
1PIR
5802
COR58
PIR5901PIR5902 COR59
PIR6001PIR6002 COR60
PIR6101 PIR6
102
PIR6103
COR61
PIR6
201
PIR6
202
COR62
PIR630
1PI
R630
2COR63
PIR6
401
PIR6
402
COR64
PIR650
1PIR
6502
COR65
PIR660
1PIR
6602
COR66
PIR6701PIR6702 COR67
PIR6801PIR
6802PIR6803COR68
PIR6901PIR6902 COR69
PIR7
001
PIR7
002COR70
PIR7
101
PIR7
102
COR71
PIR720
1PIR
7202
COR72
PIR7
301
PIR7
302
COR73
PIR7401PIR7402 COR74
PIR7
501
PIR7
502
COR75
PIR7
601
PIR7
602
COR76
PIR7
701
PIR7
702
COR77
PIR7801PIR7802 COR78
PIR7
901
PIR790
2COR79
PIR8
001
PIR8002
COR80 PIR
8101
PIR8
102COR81
PIR8
201
PIR8202
COR82
PIR8
301
PIR8302
COR83
PIR8
401
PIR8402
COR84
PIR8
501
PIR8502
COR85
PIR8
601
PIR8602
COR86
PIR8
701
PIR8
702
COR87
PIR8
801
PIR8
802
COR88
PIR8
901
PIR8
902
COR89
PIR9001PIR9002 COR90
PIR9101PIR9102 COR91
PIR9
201
PIR9
202
COR9
2
PIR9
301
PIR9
302
COR93
PIR9
401
PIR9
402
COR9
4
PIR950
1PIR
9502
COR95
PIR9
601
PIR9
602
COR96
PIR9
701
PIR9
702
COR97
PIR9
801
PIR9
802
COR98
PIR9
901
PIR9
902COR9
9
PIR100
01PIR
10002
COR1
00
PIR101
01PIR
10102
COR101
PIR102
01PIR
10202
COR1
02
PIR103
01PIR
10302
COR103
PIR104
01PIR
10402
COR1
04
PIR105
01PIR
10502
COR1
05
PIR106
01PIR
10602
COR106
PIR107
01PIR
10702
COR107
PIR10801PIR10802 COR1
08
PIR109
01PIR
10902COR109
PIR11001PIR11002CO
R110
PIVR101
PIVR102
PIVR103
PIVR104
PIVR105
PIVR106
PIVR107
PIVR108
COVR
1PIVS101
PIVS102
PIVS103
COVS1
PIY101
PIY102
COY1
PIC402
PIC602
PIC1002
PIC1602PIC1701
PIC360
1
PIC3901
PIC4201
PIC4402
PIC4501
PIC4701
PIC5002
PIC5701
PIC6001
PIC6601
PIIC1014
PIIC2016
PIIC3016
PIIC
1401
4
PIIC
1501
4PI
IC16
014
PIIC1905
PIIC1908
PIIC2005
PIIC2008
PIIC2105
PIIC2108
PIIC2205
PIIC2208
PIIC2405
PIIC2408
PIIC2505
PIIC2508
PIIC
3102
PIP601
PIR202PIR302
PIR6803
PIR6902
PIVS103
PIC2802
PIC2902
PIC3102
PIC5502
PIIC1705
PIIC
2705
PIIC3003
PIR3
201
PIR3502
PIR6101
PIVR106
NL05VP
PIIC502
PIP101
PIC1101
PIC1201
PIC1502PIC3002
PIC3502
PIC4102
PIC4802
PIC5101
PIC5201
PIC5301
PIC5602
PIC6
301
PIIC408
PIIC503
PIIC608
PIIC808
PIIC908
PIIC1008
PIIC1208PIIC1708
PIIC1808PIIC1901 PIIC2001 PIIC2101 PIIC2201
PIIC2308PIIC2401 PIIC2501
PIIC2608PIIC2708
PIIC2808
PIIC2908
PIIC3008PIP201
PIP202
PIVR102
PIVS101
PIC1
302
PIC1401
PIC2102
PIC3701 PIC4301
PIC490
1
PIC5401
PIC5801
PIC6201
PIC6
502
PIIC404
PIIC505
PIIC604
PIIC804
PIIC904
PIIC1004
PIIC1204PIIC1704
PIIC1804PIIC1904 PIIC2004 PIIC2104 PIIC2204
PIIC2304PIIC2404 PIIC2504
PIIC2604PIIC2704
PIIC2804
PIIC2904
PIIC3004
PIP205
PIP206
PIIC501
PIP102
PIIC601
PIIC602
PIR3
701
PIR4
201
PIR4
501
PIR5
001
PIR7
901
PIR8
401
NLDe
adT
PIIC
1504
PIIC
1608
PIP2010
NLErr
PIC101PIC201
PIC302PIC401
PIC502
PIC601
PIC701PIC1001
PIC1
102
PIC1
202
PIC1301
PIC1
402
PIC1501PIC1601
PIC1702
PIC1801
PIC1901PIC2101
PIC2301
PIC2402
PIC2701PIC2801
PIC2901
PIC3001PIC3101
PIC3401
PIC3501
PIC360
2
PIC3702PIC3902
PIC4101
PIC4202
PIC4302
PIC4401
PIC4502
PIC4
702
PIC4801
PIC490
2
PIC5001
PIC5
102
PIC5202
PIC5
302 PIC5402
PIC5501
PIC5601
PIC5702
PIC5802PIC6002
PIC6202
PIC6302
PIC650
1
PIC6602
PIC6701 PIC6901PIC7101
PIIC107
PIIC1012
PIIC
208
PIIC2015
PIIC308
PIIC3015
PIIC405
PIIC504
PIIC1407
PIIC
1507
PIIC1607
PIIC
1703
PIIC1906 PIIC2006 PIIC2106 PIIC2206PIIC2406 PIIC2506
PIIC
2703
PIIC2901
PIIC
3104
PIP203
PIP204
PIP602
PIP702
PIP802
PIP902
PIP1002
PIP1102
PIP1202
PIP1
502
PIR701
PIR120
2
PIR140
2
PIR1
602
PIR170
2
PIR1801
PIR2501
PIR3301
PIR3403
PIR3
902
PIR4
101
PIR4
702
PIR4
901
PIR5901
PIR6701
PIR6801
PIR8
102
PIR8
301
PIR105
01 PIR11001
PIVR104
PIVS102
PIP2013
PIR7
601
NLIa
PIP2014
PIR720
1NLI
b
PIP2
015
PIR8
801
NLIc
PID901PID1002
PIIC
2702
PIIC
2706
PIR240
1
PIR5
302
NLIdc
PIC2
602 PIIC907
PIR2
301
PIR109
01
NLIdcf
b
PID701PID802
PIIC
1702
PIIC1706
PIP208
PIR200
2
PIR3
602
NLIoz
PIC102
PIIC101 PIR101
PIY101
PIC202
PIIC102PIIC103
PIR102
PIY102
PIC301
PIIC201
PIR201
PIC501
PIIC301
PIR301
PIC702PIIC403
PIR402
PIR702PIC801
PIIC402
PIR502PI
R601
PIC8
02
PIC9
01PIIC401
PIR602
PIC902PI
R801
PIC1902
PIIC
3005
PIR660
2 PIR6702
PIC2001
PIIC
3006
PIR6
302
PIR6
401
PIC2002
PIIC2903
PIIC3007
PIR6
402
PIC2
201
PIIC2802
PIR5
401
PIR5
502
PIC2202
PIIC2801
PIR3
101
PIR5
402
PIC2302
PIIC2803
PIR580
2 PIR5902
PIC2401
PIIC
2805
PIR3
202 PIR3302
PIC2
501
PIIC2807
PIR3
002
PIR5
301
PIC2502
PIIC
2806PIR3
001
PIR3
102
PIC2601
PIIC906
PIR105
02PIR107
01
PIC2702PIIC
905
PIR106
02PIR10802
PIC3301
PIIC1202
PIR2
702
PIC3
302
PIIC
1201
PIR2
102
PIR3
601
PIC3402PIIC1203
PIR2
902
PIC6401
PIIC606
PIR7
502PI
R770
1
PIC6702PIIC605
PIR7
602
PIR7802 PIC6
801
PIIC806
PIR8
702PI
R890
1
PIC6902PIIC805
PIR8
802
PIR9002
PIC700
1
PIIC802
PIR7
102PIR7
301
PIC7102PIIC803
PIR720
2PIR7402
PID702PI
IC17
07PID801
PIIC1701
PID902PI
IC27
07PID1001
PIIC2701
PIIC104
PIIC209
PIIC105
PIIC2012
PIIC106
PIIC
203
PIIC108PIR401
PIIC109
PIIC3010
PIIC
3013
PIR501
PIIC
1010
PIIC303
PIIC1011
PIIC3012
PIIC
1013
PIIC
202
PIIC204
PIIC205
PIIC206
PIIC207
PIIC2011
PIIC2010
PIIC2013
PIIC309
PIIC2014
PIIC
302
PIIC304
PIIC305
PIIC306
PIIC307
PIIC
3011
PIIC
3014
PIIC406
PIR802
PIR9
01
PIIC603
PIR3
402
PIIC
901
PIR104
02
PIR106
01
PIIC902
PIR6001
PIR104
01
PIIC903
PIP603
PIIC1001
PIR220
1
PIR2
701
PIIC
1002P
IR17
01
PIR2
202
PIIC
1003
PIIC
1007
PIR1
901
PIIC1005 PIR1802PIR2
101
PIIC1006
PIR190
2
PIR2
001
PIIC1205
PIR2
402
PIR2502
PIIC1206
PIR2
302
PIR2
601
PIIC
1207
PIR2
602
PIR290
1
PIIC
1401
PIIC
2007
PIR9
802
PIIC1403PIP1001
PIIC
1404
PIIC
2507
PIR950
2
PIIC1406PIP801
PIIC1408PIP901
PIIC
1409
PIIC
1907
PIR9
702
PIIC
1401
1PIP701
PIIC
1401
2PI
IC24
07
PIR9
302
PIIC1502
PIIC1506
PIIC1508PIP1201
PIIC
1509
PIIC
2207
PIR103
02
PIIC
1501
1PIP1101
PIIC
1501
2PI
IC21
07
PIR101
02
PIIC1606
PIIC
1601
0
PIIC1802
PIR3702PIR4
001
PIIC1803
PIR3802
PIR3901 PIIC1805
PIR4202
PIR430
2PIIC1806
PIR4102PIR4
401
PIIC1902
PIR9
602 PIR9
701
PIIC2002
PIR9
801
PIR9
901 PIIC2102
PIR100
02 PIR101
01
PIIC2202
PIR102
02 PIR103
01
PIIC2302
PIR790
2PIR8
201
PIIC2303
PIR8002
PIR810
1 PIIC2305
PIR8402
PIR8502PIIC2306
PIR8302PIR8
601
PIIC2402
PIR9
202 PIR9
301
PIIC2502
PIR9
402 PIR950
1
PIIC2602
PIR4502PIR4
801
PIIC2603
PIR4602
PIR470
1 PIIC2605
PIR5002
PIR5102PIIC2606
PIR4902PIR5
201
PIIC2902
PIR7
002
PIR9101PIIC2905
PIIC2906
PIIC3001
PIIC3002
PIR9102
PIIC
3101
PIIC3103
PIIC3105
PIR680
2PI
IC31
07
PIIC3108
PIP209
PIP2011
PIP1301 PIR
5701
PIP1302PIR5
601
PIP1401 PIR
6501
PIP1
402PIR6
201
PIR902
PIR1
001
PIR1
101
PIR7
702
PIR120
1PIR7801
PIR1
301
PIR7
302
PIR140
1PIR7401
PIR1
501
PIR8
902
PIR1
601
PIR9001
PIR3401PIR3501
PIR5
501
PIR5
602
PIR570
2PIR580
1
PIR6002PIR6
102
PIR6103
PIR6
202PIR630
1
PIR650
2PIR660
1
PIR107
02PIR
10902
PIR10801 PIR11002
PIVR101
PIVR103
PIVR105
PIVR107
PIVR108
PIIC
1402
PIIC
1405
PIIC
1401
0
PIIC
1401
3
PIIC
1503
PIIC
1501
0
PIIC
1501
3NLn
F
PIIC
1505
PIP2016
NLPWM
PIIC1501
PIP2012
NLTemp
PIC1802PIP207
PIP1501
NLTp
PIC6
402
PIIC607
PIR1
102
PIR8
001
PIR8
501
NLVae
PIC700
2 PIIC
801
PIR1
302
PIR3
801
PIR4
301
NLVbe
PIIC1609
PIIC2907
PIR6901
PIR7
001
NLVb
us
PIC680
2
PIIC
807
PIR1
502
PIR4
601
PIR5
101
NLVce
PIIC2301
PIR8202
PIR9
201
NLVeS1
PIIC2307
PIIC2503
PIR8602
NLVe
S2
PIIC1801
PIR4002
PIR9
601
NLVeS3
PIIC1807
PIIC2003
PIR4402
NLVe
S4
PIIC2601
PIR4802
PIR100
01
NLVeS5
PIIC2203
PIIC2607
PIR5202
NLVe
S6
PIIC3106
PIR7
101
PIR7
501
PIR8
701
NLVo
ffa
PIIC1604
PIP703
NLVsat
1
PIIC1602
PIP803
NLVs
at2
PIIC1605
PIP903
NLVs
at3
PIIC1601
PIP1
003
NLVs
at4
PIIC
1601
2
PIP1
103
NLVs
at5
PIIC
1601
3
PIP1203
NLVsat
6
PIIC407
PIIC1903
PIIC2103
PIIC2403
PIR1
002
PIR9
401
PIR9
902
PIR102
01
NLVtri
Figure A.6: Schematic for the inverter analog control circuit.
Stellenbosch University https://scholar.sun.ac.za
Appendix B
Printed Circuit Boards
B.1 DC-DC Converter Printer Circuit Boards
The PCBs used for the dc-dc converter prototype circuit are given in thissection.
B.1.1 Power Stage PCB
The two PCBs used to generate the positive and negative dc bus are shown inFig. B.1 and Fig. B.2 respectively. The boards are made from 70 µm coppertracks on both the top and bottom layers. Each PCB contains a half-bridgeconverter topology, gate-drive circuitry and a pre-charge soft-start mechanismas well as current sense resistors.
157
Stellenbosch University https://scholar.sun.ac.za
APPENDIX B. PRINTED CIRCUIT BOARDS 158
Inductor
Pre-chargecircuit
MOSFET
Gate-drivecircuit
Current senseresistors
Buscapacitors
Inductor
Pre-chargecircuit
Gate-drivecircuit
Buscapacitors
MOSFET Current senseresistors
Figure B.1: PCB for the top-half of the dc-dc converter circuit producing+175 V.
Stellenbosch University https://scholar.sun.ac.za
APPENDIX B. PRINTED CIRCUIT BOARDS 159
Inductor
Pre-chargecircuit
MOSFET
Gate-drivecircuit
Current senseresistors
Buscapacitors
Inductor
Pre-chargecircuit
MOSFETCurrent sense
resistors
Gate-drivecircuit
Buscapacitors
Figure B.2: PCB for the bottom-half of the dc-dc converter circuit producing-175 V.
Stellenbosch University https://scholar.sun.ac.za
APPENDIX B. PRINTED CIRCUIT BOARDS 160
B.1.2 Controller PCB
The two-layer PCB containing the dc-dc converter control circuit is shown inFig. B.3. The tracks on both the top and bottom layers are made from 35 µmof copper. The PCB includes a DKE10-15A 10 W dc-dc regulated dual output(+15 V and -15 V) converter to power the control circuit from one of the 12 Vbatteries.
+5VRegulator
DKE10A15Converter
PWMSignals
PWMSignals
Carrier WaveformGeneration
Figure B.3: Control PCB of the dc-dc converter.
Stellenbosch University https://scholar.sun.ac.za
APPENDIX B. PRINTED CIRCUIT BOARDS 161
B.2 Inverter Printer Circuit Boards
The PCBs used for the inverter prototype circuit are given in this section.
B.2.1 Power Stage PCB
The PCB used to implement the power stage of the inverter circuit is shownin Fig. B.4. The board is made from 70 µm copper tracks on both the topand bottom layers. The PCB contains the three-phase half-bridge invertertopology located inside the IGBT module, gate-drive circuitry and the dc buscurrent sensor. The PCB includes a DKE10-15A 10 W dc-dc regulated dualoutput (+15 V and -15 V) converter to power the gate-drive circuitry from a12 V bench supply.
IGBT moduleGate-drivecircuitry
Three generator statorterminals
Allegro currentsensor
DC busterminals
Figure B.4: PCB for the power stage of the inverter circuit.
Stellenbosch University https://scholar.sun.ac.za
APPENDIX B. PRINTED CIRCUIT BOARDS 162
B.2.2 Controller PCB
The two PCBs containing the control circuitry used to control the IGBTs onthe power board are shown in Fig. B.5 and Fig. B.6 and contain the digitaland analog control circuits, respectively. Both PCBs are two-layer boards andthe tracks on the top and bottom layers are made from 35 µm of copper. Theanalog PCB includes a DKE10-15A 10 W dc-dc regulated dual output (+15 Vand -15 V) converter to power the control circuitry on both the analog anddigital PCBs from a 12 V bench power supply.
Magnetic pick-upsignal processing
Control boardinterface
DAC dsPIC
Figure B.5: PCB for the digital inverter control circuit.
Stellenbosch University https://scholar.sun.ac.za
APPENDIX B. PRINTED CIRCUIT BOARDS 163
dsPIC PCB placement
Carrier waveformgeneration
PWM SignalsPWM Signals
CS <3 AC
Figure B.6: PCB for the analog inverter control circuit.
Stellenbosch University https://scholar.sun.ac.za