Power Electronics for Machine Drives

Introduction and Context

Power electronics is the interface between an electrical supply and an electric machine. It converts voltage, current, frequency, and power flow so that a motor can deliver the required torque and speed efficiently. Modern mechatronic systems—including robots, electric vehicles, machine tools, pumps, compressors, and automated production lines—depend on controlled semiconductor converters rather than direct connection of motors to fixed-frequency supplies.

A typical drive contains four functional layers: a source, a power converter, a machine, and a control system. The converter may rectify alternating current (AC) into direct current (DC), regulate a DC voltage, invert DC into variable-frequency AC, or return energy from the machine to the supply during braking. Semiconductor switches perform these functions at high speed, while sensors and digital controllers determine switching commands.

The central engineering challenge is to obtain accurate torque and speed control while limiting losses, harmonics, electromagnetic interference (EMI), acoustic noise, and stress on the motor and supply.

Learning Objectives

After studying this chapter, the reader should be able to:

1. Explain the operation and selection of power semiconductor switches.
2. Analyze uncontrolled and controlled rectifiers, DC choppers, voltage-source inverters, and current-source converters.
3. Explain pulse-width modulation (PWM), DC-link operation, and regenerative power flow.
4. Estimate converter output quantities, switching losses, and harmonic effects.
5. Identify power-quality, EMC, protection, thermal, and safety requirements.
6. Compare silicon, silicon-carbide, and gallium-nitride devices.
7. Apply these concepts to the design of a practical machine-drive system.

1. Semiconductor Switches

An ideal power switch has zero voltage when conducting, zero current when blocking, and instantaneous transitions. Real devices have finite on-state resistance or voltage drop, leakage current, switching time, parasitic capacitance, and thermal limitations.

1.1 Principal device types

Power diode. A diode conducts when forward biased and blocks reverse voltage. It is widely used for rectification and as a freewheeling path across inductive loads. Reverse-recovery current can produce voltage overshoot and EMI.

Thyristor (SCR). An SCR is triggered into conduction by a gate pulse but normally turns off only when its current falls below the holding current or when forced commutation is applied. It is robust and economical for high-power line-commutated converters, but its switching speed and controllability are limited.

Power MOSFET. A MOSFET is voltage-controlled, has very high switching speed, and is especially effective at low and medium voltage. Its conduction loss is approximately determined by the drain-source on-resistance, R_DS(on).

IGBT. An insulated-gate bipolar transistor combines a MOS-controlled gate with bipolar conduction. It offers relatively low conduction loss at medium and high voltage and is common in industrial inverters. It generally switches more slowly than a MOSFET.

GTO and IGCT. These gate-controlled thyristor devices are suited to very high power. They require substantial gate-drive energy but can handle high voltage and current.

Wide-bandgap devices. Silicon-carbide (SiC) MOSFETs and diodes, and gallium-nitride (GaN) transistors, permit higher switching frequency, temperature, and electric field than conventional silicon devices. Their advantages are accompanied by higher cost, stringent layout requirements, and potentially greater common-mode EMI.

1.2 Static and dynamic ratings

Important ratings include blocking voltage, continuous and pulsed current, gate voltage, switching frequency, junction temperature, short-circuit withstand time, reverse-recovery behavior, and safe operating area (SOA). The device must be selected with margin for supply variation, regenerative overvoltage, fault transients, and thermal conditions.

For a switch with approximately constant on-state voltage V_on, the conduction loss is

For a resistive approximation, the corresponding expression is

where I_avg is average conducting current, I_rms is rms current, and R_on is the effective on-state resistance. The first model is useful for IGBTs and diodes; the second is often useful for MOSFETs. Both assume the device parameters remain approximately constant over the operating range.

A first-order estimate of switching loss is

where f_s is switching frequency and E_on, E_off are turn-on and turn-off energies under specified voltage, current, gate-drive, and temperature conditions. Datasheet energy values must not be transferred blindly between operating points.

The junction temperature can be estimated by

where T_j is junction temperature, T_a ambient temperature, P_loss total device loss, and R_θJA thermal resistance from junction to ambient. For a mounted power module, a junction-to-case plus case-to-heatsink model is usually more accurate.

2. Rectifiers and the DC Link

A rectifier converts AC into DC. In a conventional industrial drive, a diode bridge produces an uncontrolled DC voltage, followed by a DC-link capacitor and a three-phase inverter.

For an ideal three-phase six-pulse diode bridge supplied by a balanced sinusoidal source with line-to-line rms voltage V_LL, the average DC voltage under continuous current is approximately

The approximation assumes negligible source impedance, ideal diodes, and sufficiently smooth DC current. A capacitor-input drive often charges near the line-voltage peak, so its no-load DC-link voltage can approach

The difference arises because the first expression describes the average output of a current-continuous six-pulse rectifier, whereas the second describes capacitor charging near the voltage peak.

A controlled bridge uses thyristors, with firing angle α, to regulate the average voltage. In the ideal continuous-current case,

For 0° < α < 90°, the converter normally operates as a rectifier. With suitable current direction and system arrangement, 90° < α < 180° permits inversion and regenerative power transfer. Commutation overlap and source inductance reduce the actual voltage.

The DC link decouples the input converter from the output inverter. Its capacitor absorbs short-term power imbalance and reduces voltage ripple. Excessive capacitance improves voltage smoothness but increases inrush current, stored energy, physical size, and fault severity.

The energy stored in a DC-link capacitor is

where C is capacitance in farads and V_dc is capacitor voltage in volts. This stored energy requires discharge resistors, pre-charge circuits, and safe isolation procedures.

3. DC Choppers

A chopper converts one DC voltage into another by high-frequency switching. For an ideal buck converter operating in continuous conduction,

where D is duty ratio, defined as on-time divided by switching period, and V_in and V_o are input and average output voltages. The ideal relation neglects switch, diode, inductor, and capacitor losses.

A boost converter has, under continuous conduction,

A buck-boost converter can reverse polarity in its basic form. Practical machine drives use half-bridge, full-bridge, or four-quadrant choppers. A four-quadrant converter can control both voltage polarity and current polarity, enabling motoring and regenerative braking in both directions.

Inductors limit current ripple. In a buck converter, the approximate peak-to-peak inductor-current ripple is

where L is inductance and f_s switching frequency. This formula assumes ideal switching and continuous conduction. Increasing inductance or switching frequency reduces ripple, but increases cost or switching loss, respectively.

4. Inverters for AC Machine Drives

An inverter synthesizes AC voltage or current from a DC source. A three-phase, two-level voltage-source inverter has three legs, each with an upper and lower controlled switch and antiparallel diodes. The two switches in one leg must never be commanded on simultaneously; a short dead time is inserted to prevent shoot-through.

The inverter does not create energy. It regulates the instantaneous relationship between the DC link and machine terminals. The motor’s inductance smooths current, while the controller adjusts voltage frequency, phase, and amplitude.

4.1 Six-step operation

In six-step operation, each leg changes state at defined electrical angles. This produces a quasi-square-wave voltage containing a fundamental component and characteristic harmonics. Six-step operation provides good DC-bus utilization but causes torque ripple, acoustic noise, and harmonic current.

4.2 Sinusoidal PWM

In sinusoidal PWM, a sinusoidal reference is compared with a high-frequency triangular carrier. For a two-level inverter in the linear modulation region, the fundamental phase-voltage amplitude is approximately proportional to the modulation index m_a. The exact relationship depends on whether phase or line voltage is being specified and on the adopted PWM convention; therefore the inverter topology and definition must always be stated.

For sinusoidal references

where V_m is reference amplitude and ω_e is electrical angular frequency in rad/s. The carrier comparison generates gate commands. Dead time causes voltage error, particularly at low speed, and often requires compensation.

4.3 Space-vector PWM

Space-vector PWM represents the three-phase reference as a rotating vector in the stationary α–β plane. During each switching period, the inverter applies adjacent active vectors and zero vectors for calculated dwell times. Compared with sinusoidal PWM, it generally provides better DC-bus utilization and systematic switching-state control. It is widely used in field-oriented control of induction and permanent-magnet machines.

4.4 Hysteresis current control

Hysteresis control switches when current error reaches upper or lower limits. It has fast current response and simple implementation, but switching frequency varies with operating conditions and machine back electromotive force. Variable frequency may complicate EMI filtering and acoustic-noise prediction.

5. Machine-Drive Control and Regeneration

Scalar volts-per-hertz control maintains an approximate voltage-to-frequency ratio to preserve air-gap flux in an induction motor. It is inexpensive but offers limited dynamic torque control.

Field-oriented control transforms three-phase currents into rotating d–q components. The d-axis current primarily controls flux and the q-axis current primarily controls torque, although exact decoupling depends on the machine model and control quality. Direct torque control instead selects inverter voltage vectors based on torque and flux errors.

During regenerative braking, the machine acts as a generator. Mechanical energy flows into the DC link. If the front end cannot return energy to the AC supply, the DC-link voltage rises. A braking chopper connects a resistor when the voltage exceeds a threshold, converting energy into heat. A four-quadrant active front end can instead generate controlled AC currents and return energy to the grid with near-unity power factor.

A regenerative system must coordinate speed reduction, current limits, DC-link voltage, thermal capacity, and grid-code requirements. Regeneration is not automatically safe: an open-circuit or weak grid may leave no path for returned energy.

6. Harmonics and Power Quality

Power converters draw nonsinusoidal current unless specifically controlled. Harmonic current causes additional transformer and motor losses, neutral conductor heating, voltage distortion, interference, and possible resonance with power-factor-correction capacitors.

Total harmonic distortion of current is commonly defined as

where I₁ is the rms fundamental current and I_h is the rms current at harmonic order h. THD is a waveform metric, not a complete measure of power quality; displacement power factor, voltage distortion, unbalance, transients, and interharmonics may also matter.

A six-pulse rectifier produces characteristic current harmonics of order

where k is a positive integer, under ideal balanced conditions. In practice, impedance, imbalance, firing errors, and discontinuous current create additional components.

Mitigation methods include line reactors, passive filters, active filters, multipulse transformers, active rectifiers, and appropriate cable and transformer sizing. Filter design must consider resonance, fundamental-frequency reactive power, fault current, and changing load conditions.

7. EMC, Layout, and Protection

Electromagnetic compatibility requires that a drive neither emit unacceptable interference nor malfunction in the presence of interference. Fast dV/dt and dI/dt produce conducted and radiated emissions through parasitic capacitances and inductances.

Good practice includes:

• minimizing commutation-loop area;
• placing decoupling capacitors close to switching devices;
• separating power and signal wiring;
• using shielded motor cables with high-frequency termination;
• bonding enclosures and heatsinks appropriately;
• controlling gate-drive impedance and voltage;
• using common-mode chokes and differential filters where required;
• measuring emissions with standardized methods rather than relying only on simulation.

Protection commonly includes semiconductor fuses, circuit breakers, current sensors, desaturation detection, overvoltage clamps, snubbers, gate interlocks, undervoltage lockout, thermal monitoring, ground-fault detection, and DC-link pre-charge. A protection circuit must act faster than the device’s short-circuit withstand time.

Isolation is essential where control electronics, user-accessible interfaces, and high-voltage power circuits have different safety domains. Creepage, clearance, insulation coordination, protective earthing, discharge time, and lockout/tagout procedures must follow applicable standards and institutional safety rules.

8. Worked Example: DC-Link and Braking Resistor

An ideal three-phase drive is supplied by a 400 V line-to-line rms source. Estimate the average DC-link voltage of a six-pulse diode bridge using the continuous-current approximation. If a braking resistor must absorb 20 kW at a DC-link voltage of 700 V, estimate its resistance and current.

For the rectifier,

This is an average value for a current-continuous rectifier. A capacitor-input link may charge closer to √2 × 400 = 566 V, before considering tolerances and transients.

For the braking resistor,

and

The resistor must be selected for pulse energy, duty cycle, thermal environment, insulation, and a suitable switching threshold. The calculation does not establish a complete design because motor inertia, braking time, and worst-case regenerated energy are also required.

9. Applied Case: Regenerative Servo Axis

Background

Consider an illustrative packaging machine whose servo axis repeatedly accelerates a carriage and then decelerates it over a short cycle. A conventional diode-front-end drive operates correctly during acceleration but repeatedly trips on DC-link overvoltage during deceleration.

Problem analysis

The kinetic energy of the carriage and reflected motor components flows into the DC link during braking. Since the diode bridge cannot return energy to the grid, the capacitor voltage rises. A braking resistor can dissipate the energy, but frequent cycles may exceed its thermal rating. In addition, the machine requires low position error and minimal disturbance to neighboring equipment.

Engineering approach

The design team first records speed, torque, DC-link voltage, braking duration, and cycle repetition. The maximum regenerative energy is estimated from the mechanical energy change,

where J is reflected rotational inertia and ω₁, ω₂ are initial and final angular speeds. Losses and energy recovered by other loads are then included. A braking chopper and resistor are sized for both peak power and average thermal power. Because the duty cycle is high, an active front end is evaluated as an alternative.

Outcome

The braking resistor prevents overvoltage for occasional stops, but the active front end provides superior energy recovery for continuous production. It also reduces low-order input harmonics and permits controlled power factor. The final choice depends on installation cost, grid requirements, maintenance, safety, and the value of recovered energy.

Learning points

1. Regenerative power must have a defined energy path.
2. Peak braking power and average braking energy are different design quantities.
3. A converter solution must be evaluated at system level, not by semiconductor ratings alone.
4. Harmonic performance, EMC, thermal behavior, and safety can determine the preferred topology.

10. Wide-Bandgap Devices and Contemporary Development

SiC devices are particularly attractive in high-voltage drives, renewable-energy converters, and electric-vehicle traction inverters because their lower switching and conduction losses can reduce heatsink size. GaN devices are promising for high-frequency, lower-voltage applications such as compact power supplies and auxiliary converters.

Higher switching speed is not automatically beneficial. It can increase bearing currents, motor insulation stress, common-mode current, and radiated emissions. Successful implementation requires low-inductance packaging, carefully controlled gate loops, suitable dead time, active gate control, and accurate measurement of switching waveforms.

Digital control platforms now combine microcontrollers, DSPs, FPGA logic, and model-based commissioning. Data-driven diagnostics can identify bearing faults, insulation degradation, abnormal current signatures, or cooling problems. Artificial intelligence may assist condition monitoring, but safety-critical protection should remain deterministic, validated, and independently supervised. Cybersecurity is also relevant when drives are networked: unauthorized parameter changes can create hazardous motion or equipment damage.

Design Best Practices and Limitations

A sound drive design begins with the machine’s torque-speed envelope, supply constraints, braking duty, environment, and required dynamic response. The engineer should then select topology, semiconductor technology, modulation method, filter, cooling system, sensors, protection, and control hardware together.

Models are useful but incomplete. Ideal-switch simulations do not predict parasitic-induced overshoot, thermal cycling, EMI, bearing currents, or manufacturing variation reliably. Prototype testing should include steady-state efficiency, overload, short-circuit response, regenerative operation, temperature rise, conducted and radiated emissions, acoustic noise, and abnormal supply conditions.

Standards and regulations vary by application and jurisdiction. Designers must consult current requirements for machinery safety, EMC, electrical installations, functional safety, grid connection, and semiconductor insulation coordination. Compliance cannot be inferred solely from a component datasheet.

Chapter Summary

Power electronics enables controlled energy conversion between electrical supplies and machines. Diodes and controlled switches form rectifiers, choppers, and inverters; the DC link separates input and output power processing and stores energy. PWM controls the fundamental voltage and current applied to AC machines, while field-oriented and related methods provide dynamic torque and flux control. Regeneration requires either a braking resistor or an energy-return path through an active front end.

Converter design must address conduction and switching losses, thermal limits, harmonics, power factor, EMC, protection, insulation, and safety. Silicon remains important, while SiC and GaN offer higher-frequency and higher-temperature capability with stricter layout and EMC demands. Professional design combines equations, simulation, measurement, standards compliance, and system-level judgment.

Key Terms

Power semiconductor: A semiconductor device designed to control substantial electrical power.

Rectifier: A converter that changes AC into DC.

Inverter: A converter that changes DC into controlled AC.

Chopper: A DC–DC switching converter.

DC link: The intermediate DC circuit connecting converter stages.

PWM: Pulse-width modulation used to synthesize a controlled average waveform.

Modulation index: A ratio describing reference-wave amplitude relative to the carrier or available inverter voltage.

IGBT: An insulated-gate bipolar transistor used widely in medium- and high-power drives.

MOSFET: A fast voltage-controlled transistor with resistive conduction behavior.

Wide-bandgap device: A semiconductor based on materials such as SiC or GaN.

Regeneration: Conversion of mechanical or electrical energy back toward the supply or an energy absorber.

Shoot-through: A destructive short circuit across a converter leg.

Dead time: Deliberate delay preventing simultaneous conduction of complementary switches.

THD: Total harmonic distortion relative to the fundamental component.

EMC: Electromagnetic compatibility between equipment and its environment.

Desaturation protection: Fast detection of abnormal semiconductor voltage during conduction.

Four-quadrant converter: A converter capable of controlling voltage and current polarities in all operating quadrants.

Space-vector PWM: A modulation method based on inverter switching vectors in a stationary reference plane.

Active front end: A controlled rectifier capable of regulating input current and returning energy to the grid.

Review Questions

Short Answer Questions

1. What are the principal functions of a DC link in a machine drive?
2. Why is dead time required in an inverter leg?
3. Distinguish conduction loss from switching loss.
4. What is the purpose of a braking resistor?
5. Name two advantages and one limitation of SiC devices.

Long Answer Questions

1. Explain the operation of a three-phase diode rectifier and discuss the assumptions behind the average DC-voltage relationship.
2. Compare sinusoidal PWM, space-vector PWM, and hysteresis current control.
3. Explain the sources and consequences of harmonics in a variable-frequency drive.
4. Describe a complete protection strategy for a voltage-source inverter.
5. Discuss how an active front end improves regenerative operation and input power quality.

Analytical / Application Questions

1. A drive supplied from a 415 V line-to-line rms source uses a six-pulse rectifier. Estimate its average DC voltage using the ideal continuous-current model, and explain why a capacitor-input measurement may differ.
2. A motor decelerates rapidly and raises the DC-link voltage. Identify the possible energy paths and state what measurements are needed before sizing a braking resistor.
3. A high-frequency SiC inverter exhibits excessive common-mode current and motor-bearing damage. Propose electrical, mechanical, layout, and measurement changes to investigate and mitigate the problem.

Multiple Choice Questions with Answers

Question 1

Which device is normally turned off by reducing its gate drive and is especially suited to high-frequency switching?

A. Power MOSFET

B. SCR

C. Line-commutated diode

D. Transformer

Correct Answer: A. Power MOSFET

Question 2

Under ideal continuous-current conditions, the average output of a three-phase six-pulse diode bridge is approximately which expression?

A. 0.5 V_LL

B. 1.35 V_LL

C. √2 V_LL only in all conditions

D. 2 V_LL

Correct Answer: B. 1.35 V_LL

Question 3

What is the primary purpose of dead time in a voltage-source inverter leg?

A. To increase DC-link capacitance

B. To prevent simultaneous conduction of the two switches

C. To eliminate all harmonics

D. To reverse motor rotation

Correct Answer: B. To prevent simultaneous conduction of the two switches

Question 4

For an ideal buck converter, if the duty ratio is 0.4 and the input voltage is 500 V, what is the average output voltage?

A. 125 V

B. 200 V

C. 500 V

D. 900 V

Correct Answer: B. 200 V

Question 5

During regenerative braking with a diode front end, where is excess DC-link energy commonly dissipated?

A. In a braking resistor

B. In the gate driver only

C. In the line-frequency transformer core only

D. In the PWM carrier

Correct Answer: A. In a braking resistor

Question 6

Which harmonic orders are characteristic of an ideal six-pulse rectifier current?

A. h = 2k

B. h = 3k only

C. h = 6k ± 1

D. h = k/6

Correct Answer: C. h = 6k ± 1

Question 7

Which control method represents inverter switching states as vectors in a stationary reference plane?

A. Six-step thermal control

B. Space-vector PWM

C. On-off thermostat control

D. Mechanical commutation

Correct Answer: B. Space-vector PWM

Question 8

A braking resistor absorbs 10 kW at 500 V. What resistance is required under the ideal resistive model?

A. 2.5 Ω

B. 25 Ω

C. 50 Ω

D. 250 Ω

Correct Answer: B. 25 Ω

Question 9

Which feature is generally associated with an active front end?

A. Inability to control input current

B. Regeneration to the AC supply

C. Operation without semiconductor switches

D. Removal of all switching losses

Correct Answer: B. Regeneration to the AC supply

Question 10

What is a major design consequence of increasing semiconductor switching speed?

A. EMI and parasitic effects may become more significant

B. Thermal design becomes unnecessary

C. The DC link stores no energy

D. Motor insulation is never stressed

Correct Answer: A. EMI and parasitic effects may become more significant