Book summary “Power electronics handbook”: summary

Introduction

You need to convert and control electrical energy efficiently across modern industrial and renewable systems. Traditional linear conversion wastes significant power through thermal dissipation. You face the challenge of designing high-frequency, high-efficiency conversion systems using semiconductor switches.

Power electronics solves this problem by combining power engineering, semiconductor physics, and control theory. Power electronics enables efficient control and conversion of electrical energy using semiconductor switching devices and advanced control methods.

The book guides you through the fundamental components and system-level applications of power conversion. You will examine device characteristics, converter topologies, control strategies, and practical motor drive implementations.

• How do you select power semiconductor devices for specific switching frequencies?
• What topologies convert alternating current to direct current with minimal harmonic distortion?
• How do you design gate drive circuits to ensure reliable switching operation?
• Which control methods stabilize variable-speed electric motor drives under dynamic loads?

1. Power Semiconductor Devices

Power electronics deals with electronic circuits designed to control the flow of electrical energy at power levels much higher than the individual device ratings. You encounter this convergence of energy systems, electronics, and control constantly. 99.999% of the time electrical energy is available in a typical American household through such systems.

High efficiency in power electronics is achieved by utilizing semiconductor devices as ideal switches that operate with zero power loss.

Ideally, when a switch is on, it has zero voltage drop and carries any current. When off, it blocks all current regardless of voltage. Because device power is the product of voltage and current, it remains zero at all times. This creates a lossless circuit element crucial for reaching near 100% conversion efficiency. A power supply switch often must function 100,000 times each second. A modern 5 V power supply reaches a potential lifespan of 1,000,000 hours. You can fit 100 W of output in a package with volume under 15 cm3 at a price point of 0.30 dollars per watt.

Diodes and Thyristors

Semiconductor devices in power electronics act as restricted switches with specific limitations on current direction, voltage polarity, and blocking behavior. Power semiconductors fall into diodes, transistors like BJTs, MOSFETs, and IGBTs, and thyristors like SCRs and GTOs. Restricted switches combine ideal switch properties with defined operational boundaries across specific quadrants of voltage-current planes. A 2N2222A transistor features a 30 V collector-emitter breakdown voltage and 0.8 A collector current. An MTW20N50 MOSFET handles 20 A drain current and 500 V drain-source breakdown voltage. Power diodes display switching speeds ranging from 20 ns to 100 ms.

Increasing the forward-blocking voltage of a thyristor requires a thicker high-resistivity n-base region, which inherently slows down switching speeds and increases the forward voltage drop during conduction. The n-base must support large applied forward voltages when the switch is in its off-state or forward-blocking state. However, adding thickness to this region increases stored charge during conduction. This causes slower turn-on and turn-off times and a larger forward voltage drop compared to lower voltage devices carrying the same current. The n-base doping relies on a 1014 cm-3 phosphorus impurity concentration across an n-base region measuring 10s to 100s of mm in thickness. A 1 kV forward-blocking voltage device switches much more slowly than a 100 V device.

A thyristor is a bistable four-layer device that remains in its conductive state after the gate signal is removed, provided the anode current exceeds the latching current. When a positive gate current is injected, electrons and holes are injected across bases in a regenerative process until center junction J2 becomes forward-biased and the device latches. High energy levels for thyristors are defined by 1 kV or currents higher than 100 A. Once latched, the gate signal becomes superfluous. You can only turn the device off by reducing the anode current below the holding current or by commutating it away.

Transistors and Advanced Devices

The power MOSFET is a unipolar voltage-controlled device using a vertical channel structure that achieves high switching speeds exceeding a megahertz. Power MOSFETs position the source and drain on opposite sides of the silicon wafer to increase power ratings. They do not suffer from second breakdown voltages, making parallel device sharing possible. Their on-state resistance causes high conduction losses at higher power ratings. The power MOSFET device was introduced in 1975. MOSFETs maintain voltage power rating limits of 600 V and current rating limits of 40 A.

An n-channel enhancement-type MOSFET operates in the triode, saturation, or cut-off region depending on the gate-to-source and drain-to-source voltages relative to the device threshold voltage. When the gate-to-source voltage exceeds the threshold voltage, a channel is induced.

• Operate in the triode region when vDS < vGS - VTh and vGS > VTh
• Operate in the saturation region when vDS > vGS - VTh and vGS > VTh
• Operate in the cut-off region when vGS < VTh

The IGBT combines the low on-state resistance of a power BJT with the insulated gate input of a power MOSFET by integrating both physics within the same semiconductor region. The vertical cross-section of an n-channel IGBT is similar to a double-diffused power MOSFET except for a p-plus layer at the bottom which acts as the collector. Applying a positive gate voltage creates an n-channel for electron flow into the n-minus drift region. This forward-biases the collector junction and injects holes that cause conductivity modulation. Standard IGBTs feature a 1.2 kV rated voltage, 400 A rated current, and a 20 kHz rated frequency.

The hybrid Darlington configuration merges a MOSFET and BJT to achieve superior operating characteristics.

A MOS Controlled Thyristor merges a thyristor structure with a pair of auxiliary MOSFETs to provide fast turn-on and turn-off control via the gate. The MCT is modeled as a four-layer pnpn thyristor merged with a PMOSFET for turning on and an NMOSFET for turning off. A small percentage of PMOS cells provide turn-on current, while a large number of NMOS cells divert base current from the internal BJT to break regeneration and turn the device off. Each MCT contains 100,000 cells where 4% of cells are equipped with P-channel MOSFETs. A 50-A MCT at 25 C exhibits a 1.1 V forward drop.

Harris Semiconductors originally manufactured MCTs before the division was sold to Silicon Power Corporation.

2. Power Converters and Topologies

Power electronic systems convert electrical energy through rectifiers, DC-DC converters, inverters, and resonant soft-switching topologies.

Power diodes enable essential power conversion functions including full-bridge rectification, voltage clamping, and voltage multiplication. In bridge rectifier configurations, four diodes route current through a load in a single direction from an AC source without requiring an isolation transformer. Diodes also combine with capacitors to clamp voltage swings or multiply AC input signals into higher DC voltage levels. A full bridge yields an average rectifier output voltage of 2 Vm / pi, while voltage doubler and quadrupler circuits yield 2 Vm.

Inrush currents in diode rectifiers with capacitive filters can be extremely high because filter capacitors initially have zero charge, necessitating protective resistance or proper diode surge ratings. When a bridge rectifier is connected to the supply at its maximum voltage, the worst-case inrush current is limited by the transformer secondary resistance and equivalent series resistance of the filter capacitor. If transformers are omitted, an inrush-limiting resistor must be added.

• Ensure the Maximum Allowable Surge Current (IFSM) rating of employed diodes is higher than the calculated worst-case inrush current.
• Add an inrush resistor to control current when transformers are omitted.

A single-phase bridge rectifier connected to a 120-V 60-Hz source with a diode IFSM rating of 150 A for 8.3 ms requires a 1.13-ohm inrush-limiting resistor. Three-phase bridge rectifiers provide high power output with high transformer utilization and minimal output ripple without requiring additional filters. By conducting sequences across six diodes with a conduction angle of 2p/3, the output ripple voltage is significantly reduced in frequency and magnitude compared to single-phase systems. Three-phase bridge rectifiers exhibit a 4.2% output ripple voltage and an output ripple frequency equal to 6 times input frequency.

Active unity power factor rectifiers, such as the single-phase boost converter, shape the input current to follow a sinusoidal reference in phase with the supply voltage, complying with harmonic standards. By actively switching a transistor and controlling inductor current via hysteresis or pulse-width modulation, boost rectifiers achieve power factors near unity and low total harmonic distortion, overcoming the limitations of capacitive filter rectifiers. Low-power equipment and locomotive drives use PWM boost and bridge rectifiers controlled by integrated circuits like the UC3854 or MC33262 to maintain high power quality. Boost rectifiers achieve a 0.99 power factor and a 4.96% total harmonic distortion, operating below the 600 W power threshold for IEC 1000-3-2 Class D standards.

• Maintain a DC output voltage at least 10% higher than the peak source voltage to ensure proper dynamic control of the input current.
• Select hysteresis width and inductance to balance input current ripple against transistor switching frequency losses.

Rectifiers and DC-DC Converters

DC-DC switching regulators and choppers step down or step up dc voltages with high conversion efficiencies by operating semiconductor switches in on and off states. Buck converters reduce input voltage proportionally to the duty ratio in continuous conduction mode, while boost converters yield higher output voltages through parallel capacitors and series inductances. Isolated transformer variants like forward, push-pull, half-bridge, and full-bridge converters provide safety isolation and voltage scaling. These systems operate at a 100 kHz switching frequency with 10 ohm resistance and 25 microfarads minimum capacitance.

Practical transformers with leakage inductance cause severe voltage and current ringings in switch-mode converters, requiring snubber and clamping circuits to suppress resonances. Resonant circuits formed by transformer leakage inductance and parasitic capacitances of diodes and transistors cause ringing voltages. RC snubbers across transformers, switches, and rectifiers damp these oscillations, though large capacitors increase switching losses. A practical forward converter uses an R1C1 snubber across the transformer primary and R2C2/R3C3 snubbers across output rectifiers to damp parasitic ringing, operating with a 0.996 practical coupling coefficient and the 0.5 CV^2 f switching loss formula.

• Use a capacitor with 5 to 10 times the junction capacitance of the rectifier as a starting point for snubber iterations.
• Select a snubber resistor that provides a slightly underdamped operating condition.

Switched-capacitor multiquadrant Luo-converters are third-generation converters constructed solely of capacitors or inductors, achieving high power density and efficiency. They implement voltage-lift and current-amplification techniques. The conduction duty cycle does not affect the power transfer efficiency directly, though maximum output power occurs at a duty cycle of 0.5. Reference parameters use a source voltage of 48 V, a load voltage of 14 V, switch resistance of 0.03 ohms, capacitor resistance of 0.02 ohms, diode voltage drop of 0.5 V, frequency of 5 kHz, and capacitance of 5000 microfarads.

The four-quadrant switched-inductor Luo-converter integrates positive source voltages with load voltages of either positive or negative polarity using a single inductor and three switches. It performs across all four quadrants corresponding to dc motor forward and reverse operations in both motoring and regenerative braking states, with specific switch and diode configurations for each mode. Source voltage V1 and load voltage V2 are assumed to be +42 V and 14 V respectively.

Inverters and Resonant Topologies

Carrier-based pulsewidth modulation defines the switch states in voltage source inverters by comparing a sinusoidal modulating signal with a triangular carrier waveform. Sinusoidal PWM scales the fundamental output voltage linearly with the modulation index in the linear region, while a modulation index threshold of 3.24 leads to a totally square ac output voltage. In the overmodulation region, higher fundamental voltages are obtained at the expense of introducing low-order harmonics.

Space-vector modulation approximates reference voltage or current vectors using valid inverter switch states and zero states within each sampling period. By utilizing space-vector transformations, the technique maximizes fundamental voltage utilization in linear operating regions compared to standard SPWM and allows systematic selection of zero space vectors to reduce switching frequency. The normalized frequency fsn requires 6 as the integer multiple to maintain symmetry across sectors.

Quasi-resonant converters replace conventional power switches with resonant switches containing inductors and capacitors to achieve ZCS or ZVS in either half-wave or full-wave modes. In ZCS-QRCs, an inductor forces current to zero at turn-off, but full-wave operation is often preferred because it transfers excess energy back to the source and makes the voltage conversion ratio less sensitive to load variations. ZVS-QRCs use resonant capacitors to provide zero-voltage conditions, though half-wave mode is preferred in practice to avoid capacitive turn-on losses. The boundary between continuous and discontinuous conduction mode in SRC sits at 1.27g.

Extended-period quasi-resonant converters modify standard hard-switched topologies to achieve constant frequency operation and full soft-switching using small auxiliary switches and resonant components. By adding a small resonant inductor in series with the main switch and an auxiliary switch network, the circuit limits the di/dt of switch current and eliminates diode reverse-recovery current problems. Because resonance occurs outside the main inductor in a closed loop, standard averaged current mode control remains applicable using calculations with a 3.6 numerator constant and 25 ns fixed base delay time.

• Determine peak input line current Is(max) using power conversion equations.
• Design the resonant tank so that the peak resonant current exceeds the maximum inductor current.
• Turn on the main switch SW1 at zero current using the resonant inductor to limit di/dt.
• Initiate the first resonance half-cycle to discharge the resonant capacitor Cr through the auxiliary switch.
• Turn off the main switch SW1 at near zero voltage during the extended resonance period.
• Turn off auxiliary switch Q2 at zero-voltage and zero-current conditions after Cr charges to Vo.

3. Control Methods and Gate Drivers

Efficient operation of power converters requires specialized gate drive circuits and advanced control methods like sliding mode and fuzzy logic.

Gate Drive Requirements

Thyristors require precise positive gate voltage and gate current thresholds for successful triggering while avoiding undesirable thermal and electrical damage. You can protect your devices by following specific design rules:

• Determine the source short-circuit current and calculate the minimum series resistance RS to protect the source.
• Ensure the resistor R1 satisfies current limits to protect the gate-to-cathode junction.
• Incorporate resistor R2 across G and K terminals to clip VG below VGM and improve noise immunity.

Manufacturers specify threshold conditions like VGT and IGT alongside maximum permissible gate power dissipation PGM and nontrigger gate voltage VGD. A series resistor or Zener diode prevents positive overshoots from exceeding VGM. Power MOSFETs are voltage-controlled devices whose switching speeds are governed by nonlinear voltage-dependent capacitances and the Miller effect. Although ideally no dc current flows into the insulated gate, current is required to charge and discharge device capacitances like Cgs and Cgd during transitions.

You can ensure rapid turn-off and full turn-end by applying these gate drive techniques:

• Use CMOS logic or open-collector TTL with a pull-up resistor connected to a separate 10 to 15 V dc supply to ensure rapid turn-off and full turn-on.
• Employ complementary push-pull emitter followers to clamp the gate-to-ground for very fast turn-off speeds.

Linear and Sliding-Mode Control

Classical linear feedback design relies on Bode plots and root-locus techniques to ensure system stability, high low-frequency gain, adequate phase margin, and proper unity-gain crossover frequency. You should set high low-frequency gain to minimize output steady-state errors and place unity gain frequency nearly an order of magnitude lower than the switching frequency. Ensure positive phase margin greater than 30 degrees, preferably between 45 and 70 degrees, and keep gain margin greater than 30 dB where the phase reaches negative 180 degrees. Sliding-mode control leverages the variable structure behavior of power converters to force state trajectories along a prescribed sliding surface.

Sliding-mode control applied to PWM audio power amplifiers achieves high fidelity, flat passband response, and low distortion without phase delays. By treating the loudspeaker load as a perturbation and employing a fourth-order Chebyshev low-pass output filter, the sliding-mode controller forces the amplifier output to track audio input signals precisely.

Example 19.13 describes a low-cost PWM audio amplifier providing over 80 W into 8-ohm loads with a 50 V power supply and a switching frequency near 450 kHz.

You can implement this approach by executing these practical development steps:

• Define the error vector using output voltage tracking errors and its time derivatives.
• Apply a third-order Bessel polynomial pole placement to determine sliding surface coefficients.
• Implement a switching function using signum components to drive the half-bridge complementary switches.

Fuzzy logic controllers provide a heuristic approach for non-linear power converter control that does not rely on an accurate mathematical model and offers high robustness against disturbances. Simulated and experimental results on a unity power factor buck-boost rectifier showed that fuzzy logic controllers exhibit low sensitivity to system parameters and superior performance under load disturbances compared to PI controllers.

4. Motor Drives and Industrial Applications

Variable-speed control of electric motors and integration into renewable energy and automotive systems are achieved through specialized drive topologies and AI techniques.

Photovoltaic cells convert sunlight directly into electricity via the photovoltaic effect, exhibiting nonlinear current-voltage and power-voltage characteristics with a distinct maximum power point. Solar cells rely on pn junctions in semiconductor materials like monocrystalline, polycrystalline, or amorphous silicon. The outer atmosphere experiences a 1.373 kW/m2 solar energy constant, while peak incident sunlight on earth's surface at noon in the tropics reaches 1 kW/m2. Monocrystalline or polycrystalline silicon wafers have a thickness of 0.2 to 0.3 mm. A silicon solar cell under standard test conditions produces an open-circuit voltage of 0.6 to 0.7 V and a short-circuit current of 20 to 40 mA for every square centimeter of cell area.

Grid-connected photovoltaic systems utilize utility-interactive inverters featuring pulse-width modulation to convert dc power into synchronized, utility-grade ac power. Typical grid-connected PWM inverters operate in a frequency range from 2 kHz up to 20 kHz. To ensure safety, you must follow specific protection practices:

• Monitor grid voltage, impedance, frequency variation, or harmonic increases to detect islanding conditions.
• Force inverter output outside protection tolerances upon grid failure.
• Isolate the inverter energy system from the grid within the specified maximum combined operation time of 2 seconds after grid failure.

Wind power extraction depends on wind speed and turbine design characteristics, utilizing advanced permanent-magnet or induction generators managed by power electronics. Until recently, typical average wind turbine sizes ranged from 300 to 600 kW, with newer units reaching 1 to 3 MW. Wind energy generation costs dropped from 35 cents per kWh in 1980 to less than 5 cents per kWh in 1999, backed by a capital cost of $600 per kW and a 95% availability factor for grid-connected wind plants.

Vector control maintains constant rotor flux linkage in induction motors by controlling instantaneous phase currents as space vectors, achieving dynamic performance equivalent to commutator motors. Six-step switching is a common commutation method for three-phase star-connected PM BLDC motors where two phases are active and one rests during each interval. You can implement this method using specific operational steps:

• Sense physical pole edges using Hall-effect sensors.
• Connect the three phases in a star shape.
• Activate two phases at a time in the sequence defined by rotor position to produce continuous rotation.

Hysteresis-band current control via pulse-width modulation regulates winding current amplitude to maintain it within a desired tolerance band around a reference value. Small machines using HBCC reach a maximum switching frequency of 100 kHz, whereas machines rated up to several hundred kilowatts have a typical minimum switching frequency of 15 kHz. High-speed electronic switching circuits require dead time and antiparallel flyback diodes to prevent short-circuit shoot-through and destructive inductive voltage spikes. You must construct these circuits carefully:

• Ensure upper and lower switches in the same leg are never turned on simultaneously.
• Insert a dead time of a few microseconds where active switches are turned off before any switch turns on.
• Provide antiparallel freewheel diodes to maintain the current path during dead time.

Renewable Energy and Automotive Systems

Direct Torque Control directly and independently controls flux linkage and electromagnetic torque by selecting optimal inverter switching modes without coordinate transformations. ABB introduced DTC technology in 1994 and released their first DTC product, the ACS600, in 1995. This induction motor drive yields typical torque response times of 2 ms. The Regenerative DTC based ACS600 Multidrive covers a power range from 200 kVA to 4000 kVA.

Multilayer feedforward artificial neural networks act as universal function approximators capable of learning complex nonlinear input-output mappings through training. An induction motor speed estimator utilized 11,000 input-output training data sets to configure network weights. Fuzzy logic systems utilize linguistic rule-bases and fuzzy inference mechanisms to approximate nonlinear functions based on expert knowledge or input-output data clustering. A brushless DC motor drive system control table established 63 linguistic rules for its fuzzy logic implementation.

The growing electrical power demands of modern vehicles require moving from traditional 12 V systems to dual-voltage systems incorporating a 42 V high-voltage bus.

Dual-voltage architectures retain a 14 V bus while introducing a 42 V high-voltage bus powered by a 36 V lead-acid storage battery to support high-power loads. Active suspension demands 12 kW of peak power, while heated catalytic converters require 3 kW. Sensorless PM BLDC motor operation eliminates physical rotor-position sensors by integrating back-emf waveforms and comparing them with a fixed reference, functioning smoothly down to a minimum speed of a few hundred revs/min. You execute sensorless operation through distinct steps:

• Obtain switching signals for the inverter from the motor back emf.
• Integrate the back emf waveforms.
• Compare the integrated outputs with a fixed reference to determine switching signals.

Servo motor drives require high dynamic response, achieved through attributes such as a high torque-to-inertia ratio, low winding inductance, and direct-coupled shaft sensors. Brushed pancake ironless motors provide an armature inductance of 100 mH, and motors are allowed to draw 2 to 3 times their continuously rated current during acceleration and deceleration. Microstepping subdivides a stepper motor's regular step angle by partially energizing multiple windings simultaneously with calculated fractional current combinations. Stepper motor drives apply microstepping to subdivide step angles by typical factors of 10 to 100 at stepping rates reaching 28,800 steps/s. You implement microstepping through structured operational sequences:

• Calculate phase current references based on static torque-angle characteristics.
• Energize multiple windings partially using combinations of currents ranging from zero to full rated value.
• Regulate phase currents via high-bandwidth current controllers to match references.

Switched-reluctance motors operate on the reluctance principle where doubly salient stator and rotor poles seek minimum-reluctance and maximum-inductance positions. An angular error limit of 1 degree can decrease SR motor torque production by almost 8% of the maximum output at high speeds, while standard SR motors achieve maximum speeds greater than 15,000 rpm. You control SR motors by following precise operational steps:

• Switch excitation current pulses to each stator phase according to rotor position.
• Turn on phase current shortly before the rising inductance region begins.
• Turn off phase current shortly before the maximum inductance region is reached.

Synchronous reluctance motors utilize conventional sinusoidal three-phase stator windings combined with specialized reluctance rotors that distort flux distribution without rotor excitation. PSpice symbols can be made self-contained by embedding model attributes and netlist expressions directly via the TEMPLATE attribute, removing the need for external subcircuit files. You build these self-contained symbols using a defined procedure:

• Select the symbol by clicking into it.
• Select the Edit Symbol function to start the symbol editor.
• Invoke the attribute input window using the Part Attributes dialog.
• Enter the required attributes including REFDES, PART, MODEL, TEMPLATE, and electrical parameters.

PSpice attribute lists for induction motors include a mutual inductance Lm of 51.25 mH, stator leakage inductance Lsl of 2.18 mH, rotor leakage inductance Lrl of 2.89 mH, moment of inertia J_mot of 0.1, and 4 poles. A systematic five-step procedure guides circuit partitioning by mapping component delivery forms against technical constraints and packaging technologies. You partition circuits by executing specific steps:

• Create a comprehensive list of electrical, mechanical, and thermal technical user requirements.
• Create a simple component characterization map identifying dominant component attributes.
• Strategically group components by delivery form considering electrical and thermal limits.
• Reassociate components into different groupings using user requirements as constraints and engineering experience.
• Map the groupings of components to packaging technologies iteratively.

Circuit partitioning and technology selection must be driven by a full-cost model that combines material costs, production costs, and business overhead rather than relying solely on bill of materials or substrate area cost. Automated technologies experience a typical cost decrease factor of 20% per 10-fold increase in production volume. You calculate facility production economics by following explicit steps:

• Determine total wages, equipment and facility depreciation, and facility overhead.
• Determine the number of production technologies in the facility.
• Determine technology costs by calculating the ratio of overhead parameters to production technologies.
• Add scaling factors for volume dependency.

10 Key Ideas

1. High efficiency in power electronics is achieved by utilizing semiconductor devices as ideal switches that operate with zero power loss.
2. A thyristor is a bistable four-layer device that remains in its conductive state after the gate signal is removed, provided the anode current exceeds the latching current.
3. The power MOSFET is a unipolar voltage-controlled device using a vertical channel structure that achieves high switching speeds exceeding a megahertz.
4. The IGBT combines the low on-state resistance of a power BJT with the insulated gate input of a power MOSFET by integrating both physics within the same semiconductor region.
5. Power electronic systems convert electrical energy through rectifiers, DC-DC converters, inverters, and resonant soft-switching topologies.
6. Quasi-resonant converters replace conventional power switches with resonant switches containing inductors and capacitors to achieve ZCS or ZVS in either half-wave or full-wave modes.
7. Sliding-mode control leverages the variable structure behavior of power converters to force state trajectories along a prescribed sliding surface.
8. Vector control maintains constant rotor flux linkage in induction motors by controlling instantaneous phase currents as space vectors, achieving dynamic performance equivalent to commutator motors.
9. Multilayer feedforward artificial neural networks act as universal function approximators capable of learning complex nonlinear input-output mappings through training.
10. Dual-voltage architectures retain a 14 V bus while introducing a 42 V high-voltage bus powered by a 36 V lead-acid storage battery to support high-power loads.