An unfortunate consequence of our preoccupationwith things electrical is that the problems ofheat sinking and thermal management are frequentlyignored until forced on us by sound, sight, orsmell. The insatiable need to make things smaller– and the possibility of doing so by using higherfrequencies and new components and materials –aggravates the problem of heat transfer, becausesuch improvements in power densities are seldomaccompanied by corresponding improvements inefficiency. Thus we are stuck with the task ofgetting the same heat out of a smaller volumewhile disallowing any increase in temperature.

The topics we have addressed so far – powercircuits, control, and components – do not coverall the issues encountered in designing a powerelectronic system. Among those we have deferredare: (i) providing gate and base drives to thepower semiconductor switches; (ii) using forcedcommutation to turn off SCRs; (iii) controllingthe transient voltages and currents that accompanyswitching in practical circuits; (iv) contendingwith EMI created by fast switching waveforms; and(v) providing a thermal environment that allowssystem components to operate within theirtemperature ratings. We address these five topicsin Part IV.

Magnetic components such as inductors andtransformers are present in the vast majority ofpower electronic circuits. Inductors store energyin the conversion process, filter switching ripple(as part of input and output filters, forexample), create sinusoidal variations of voltageor current (paired with capacitors, as in resonantconverters), limit the rate of change of current(as in snubber circuits), and limit transientcurrent.

In this chapter we introduce polyphase systems,their advantages relative to single-phase systems,and how they are represented. We then use thesingle-phase bridge circuit as a building block tocreate polyphase rectifiers and inverters. We alsodiscuss the operating characteristics of thesecircuits from both the ac and dc sides.

When a power semiconductor switch is either onor off, its power dissipation is relatively small.However, the transition from one state to theother is often not so benign, and can imposesimultaneous high voltage and high current on theswitching device. Special efforts are oftenrequired to ensure that the device will survivethis most stressful part of its operating cycle.We have already seen in Fig. 22.4(b) how a snubbercircuit consisting of$C_s$ and$R_s$ can moderatethe effects of parasitic inductance.

The phase-controlled dc/ac converter introducedin Chapter 4 requires that the external ac systembe a voltage source, typically the ac utilityline. This condition is necessary because thephase-controlled converter uses the reversal ofthe ac voltage to drive the commutation process.Therefore the ac frequency in these circuits isconstrained to be that of the ac source. In thischapter we remove the restriction that the acsystem be a voltage source, realizing that bydoing so we must use means other than linecommutation to turn devices off. These circuits,which we call switched-modedc/ac converters, use fully controllabledevices such as MOSFETs and IGBTs, and switch atfrequencies that are higher (often much higher)than the ac-side frequency.

Pulse-width-modulated switching converters arepower circuits in which the semiconductor devicesswitch between on and off states at a rate that isfast compared to the frequencies of the input andoutput waveforms. Control of the converter isexercised by varying the ratio of on-time to totalswitching period of the controlling switches,thereby controlling the width of the pulsesapplied to the output. This is called pulse-width modulation orPWM control.

A basic understanding of device physics, andhow the physics relates to device behavior, arevaluable assets for both interpreting a device’sspecification and its successful application.

A power circuit is typically composed of only afew kinds of components (other than its source andload): switches, and energy storage elements suchas capacitors and inductors (or transformers). Inits ideal form, each of these components islossless and capable of operating at anyfrequency. In the ideal switch, for instance, thevoltage across it is zero when it is on, thecurrent flowing through it is zero when it is off,and the transition between these two states occursinfinitely fast. Although we did not explicitlydiscuss the means by which the actual switches areturned on and off (a topic covered in Chapter 23),an ideal switch responds infinitely fast to itsdrive signal and requires zero drive power.

The rectifier circuits discussed so far areuncontrollable; that is, their output voltage is afunction of system parameters and cannot beadjusted in response to parametric changes, suchas variations in load$(I_d)$ or ac voltage$(V_s)$.

Converter circuits in the class known asswitched-capacitorconverters (SCCs) comprise only switchesand capacitors. They operate on the principle ofcharge transfer, wherein capacitors are charged inone switching state, then reconfigured in a secondstate to deliver charge to the output. Dependingon the circuit topology and how the switches areconfigured and controlled, the converter functioncan be either step-up or step-down.

Power semiconductor devices are distinguishedby the high current densities at which theyoperate while on, and the high voltages they mustwithstand when off. These requirements haveserious consequences for both their physicalstructure and electrical behavior. In this chapterwe consider their structural and behavioraldepartures from the ideal devices studied inChapter 16.

In Chapter 12 we introduced the process oflinearization for certain classes of nonlinearaveraged-circuit models. This allowed us to obtainLTI circuit models for small perturbations ofaveraged values from their constant values innominal, steady-state operating conditions. TheseLTI models then served as the basis for stabilityevaluation and control design in the examplesconsidered in Chapter 12.

A rectifier converts ac to dc. A basicrectifier circuit produces dc in the electricalengineering sense, that is, unipolar current flow.It does not produce dc in the mathematical sense,that is, a waveform that is constant in time andwhose spectrum consists of a single zero-frequencycomponent. A rectifier’s output containsconsiderable ac content. These ac componentsresult in fluctuations, called ripple, about the averagevalue of the dc output. Eliminating this rippleand obtaining an approximation to “pure” dcrequires insertion of a filtering process afterthe basic rectification function.

The preceding chapter presented severalexamples of perturbations from nominal operationof power electronic systems. These examplesmotivated the need for dynamic modeling of powerconverters, in order to understand suchperturbations and to design appropriatecontrollers. We introduced the idea of circuitaveraging as a means of obtaining simple andinformative dynamic models in circuit form.