Acoustics Leo Beranek PDF

ACOUSTICSACOUSTICS Leo L. Beranek 1984 Edition Acoustic Laboratory Massachusetts Institute of Technology Bolt Beranek and Newman, Inc. 1993 Edition 975 Memorial Drive, Suite 804 Cambridge, MA 02138Library of Congress Catalog Card Number: 86-70671 International Standard Book Number: 0-88318-494-X Copyright©1954, 1986, 1990, 1993, 1996, by the Acoustical Society of America All rights reserved. No part ofthis publication may be eproduced, stored in a retrieval system, or transmitted, in any form or by any means (electronic, mechanical, photocopying, recording, or otherwise) without the prior written permission of the publisher. Published by the Acoustical Society of America though the American Institute of Physics, Ine '500 Sunnyside Blvd, Woodbury, New York 11797 Printed in the United States of America PREFACE Acoustics isa most fascinating subject. Music, architecture, engineering, science, drama, medicine, psychology, and linguistics all seek from it answers to basic questions in their fields. In the Acoustics Laboratory at MLL-T. students may be found working on such diversified problems as auditorium and studio design, loudspeaker design, subjective percop- tion of complex sounds, production of synthetic speech, propagation of sound in the atmosphere, dispersion of sound in liquids, reduction of noise from jetaireraft engines, and ultrasonic detection of brain tumors. The annual meetings of the Acoustical Society of America are veritable five- ring shows, with papers and symposia on subjects in all the above-named fields, Opportunites for employment are abundant today because man ‘agement in industry has recognized the important contributions that acoustics makes both to the improvement of their products and to the betterment of employee working conditions. There is no easy road to an understanding of present-day acoustics. First the student must acquire the vocabulary that is peculiar to the subject. Then he must assimilate the Iaws governing sound propagation and sound radiation, resonance, and the behavior of transducers in an Acoustic medium, Last, but certainly not of least importance, he must Jearn to understand the hearing characteristics of people and the reac- tions of listeners to sounds and noises. ‘This book is the outgrowth of course in acoustics that the author has taught to seniors and to first-year graduate students in electrical engineering and communication physies. The basic wave equation and some of its more interesting solutions are discussed in detail in the first part of the text. The radiation of sound, components of acoustical systems, microphones, loudspeakers, and horns are treated in sufficient detail to allow the serious student to enter into electroacoustic design. ‘There is an extensive treatment of such important problems as sound in enclosures, methods for noise reduction, hearing, speech intelligibility, and psychoacoustic criteria for comfort, for satisfactory speech intelligi Dility, and for pleasant listening conditions The book differs in one important respect from conventional texts on acoustics in that it emphasizes the practical application of electrical circuit theory in the solution of a wide variety of problems. Wherever possible, the background of the electrical engineer and the communication physicists utilized in explaining acoustical conceptsPREFACE ‘The high-fidelity expert will find the chapters onloudapesker enclosures, horus, and rooms particularly interesting because they show how the per formance of loudspeakers either in baffles oF attached to horns may be accurately and simply calculated. ‘These chaptere alco ilustrate the necessity of considering in design the overall system, including the amplifier, the loudspeaker, the baflle or horn and considering also the room in which they are to be operated. Numerical examples and sum- rary charts are given to facilitate application ofthis material to music- reproduction systems, Tn view of the incressed interest in noise control, the author has pt this subject in mind in waiting Chapters 1, 2, 4, and 10 to 18. ‘These chapters served as the basis of a special summer program on noise reduction at MILT. in 1958. ‘The material of Chapters 11 and 1 is new, and it is hoped that it will be of value to those interested in noise and its effect on human beings, Tn short, the engineer or scientist who wishes to practice in the field of coustios and who does not intend to confine his efforts to theoretical tatters must know the material of this text Problems for each chapter are included at the end ofthe text for use by thestudent. References to collateral reading in English are given in the text, although no attempt has been made to give a bibliography of the primary sources of material. Suggestions to instructors for best use of the text aro given immodintely after this preface. ‘The author wishes to express his deep appreciation to Francis M. Wiener and Rudolph H. Nichols, Jr for their assistance in the detailed review and editing of the text and the preparation of some original material. Many members of the Acoustics Laboratory at M.L:T. have read one or more chapters and have given valuable assistance to the author. Of these, particular mention is made of Mary Anne Summer- field, Walter A. osenblith, Kenneth N. Stevens, Jerome R. Cox, Jordan J. Baruch, Joanne J. English, and Norman Doeling ‘The illustrations are due to the highly capable and untiring efforts of Clare Twardsik. ‘The author is deeply indebted to his typist, Eliasbeth H. Jones, to his secretary, Lydia Bonazzoli, and to his wife, Phylis, who ‘made it possible for him to complete the text within a reasonably short span of time. Leo L. Beranex SUGGESTIONS FOR INSTRUCTORS ‘This text is divided into thirteen chapters, comprising thirty-two parts. Bach patt is intended to he approximately 1 week's work, although this will vary among students owing to differences in their previous training. If the entire class expects to take a full year of acoustics, the parts should be taught in sequence, with the exception of Part XXVIII, Meas- urement of Acoustic Levels, which may be referred to in associated labor ratory experiments and demonstrations throughout the course. If only ‘8 purt of the eluss plans to continue throngh both terms, thefundamental ‘material should be taught in the first term and the more applied material in the second, One suggested division, in this case, is as follows: Birst Term Second Term Part I. Introduetion Part V. Energy Density ‘and Intensity Part I. ‘Terminology Part IX. Circuit’ Theorems, Part HI. ‘The Wave Bqua- ergy and Power Part XI. Diroctivity Index Part IV tions of the and Directivity Wave Equation Factor Part VI. Mochanieal Cire Patt XIV, General Characters ae isties of Micro- — phones Pare VIt. Acoustical Part XV. Pressure gin) Microphones, Part VIL “Pransducers Part XVI. Gradient and Com- Part X.Directivity bination Micro Patterns phones oe Part XVII. Design Factors Af Part XL aiaion Impede leas aera ve ator Loudspeakers Part XIU. Acoustic Part XX. Bass Reflex Enclo- Hlements cae Part XVI. Basie Theory of Part XI Hora Driving Direot-radiator Units Lanulspeakers Part NXUL Hornsvill SUGGESTIONS FOR INSTRUCTORS First Term Second Term, Part XIX. Simple Enclosures Part XXV, Sound Transmis- sion through Walls Part XXII, Sound Fields in between Enclosures Small Regularly Part XXVI. Noise Control Pro- cedures and Noise Shaped Brelonures ‘ Part XXIV. Sound Fields in Pitt XXVIT Acmatic Tanemi Large Irregularly ‘at Seen Shaped Bnelowures PAC XXIX. Reciprocity of Transducers Part XVII. Measurement of Par XXXI. Speech Acoustic Levele Intelligiblity Part XXXII. Psychoacoustie Part XXX, Hearing Criteria A course in acousties should be accompanied by a set of well-planned JIaboratory experiments. For example, the material of the first few chapters will be more significant if accompanied by a laboratory experiment, (on noise measurement. ‘This will familiarize the student with the measurement of sound pressure and with the use of a frequency analyzer He will appreciate more fully the meaning of sound pressure, sound intensity, decibels, sound energy density, and power level; and he will understand the accuracy with which noise can be measured, A suggested minimum of 10 experiments, listed both numerically for a year's course and by term, is as follows: First Term Second Term No. 1. Noise measurement, No. 3. Free-feld calibration of ‘microphones No. 5. Design and testing of a loudspeaker baffle No. 8. Prediction and control of noise in a ventilating No.2. Measurement of the cone stants of an electro- ‘mechanical transducer No.4. Measurement of free-field response of a loudspeaker system No. 6. Study of sound fields ina No, 9, Audiometsic testing of small rectangular hearing enclosure No. 10. Application of psychoacoustic eriteria in the design of an auditorium, No.7. Study of sound fields in a large irregular enclosus An assignment of two problems per week should provide sufficient application of the material of the text, ‘The short list of problems for each chapter should he supplemented by timely problems derived from the instructor's experience, CONTENTS PREFACE SUGGESTIONS FOR INSTRUCTORS CHAPTER 1. INTRODUCTION AND TERMINOLOGY rant I Introduction pane IL Terminology CHAPTER 2. THE WAVE EQUATION AND SOLUTIONS vant IIL The Wore Equation vant IV Solutions of the Wave Equation pant V_ Energy Density and Intenaty CHAPTER 3. ELECTRO-MECHANO-ACOUSTICAT. CIRCUITS Pant VI" Mechanieal Cirewite pat VIL Acoustical Cireuite part VIII Transducers rant IX Circuit Theorems, Bnergy, and Power CHAPTER 4, RADIATION OF SOUND vane X- Directiity Patterns vant XI Directivity Index and Directivty Factor CHAPTER 5. ACOUSTIC COMPONENTS pax XIT Rodiation Impedances vant XIII Acoustic Blomenta CHAPTER 6, MICROPHONES rant XIV. General Characeritice of Microphones pane XV Preseure Microphones vant XVI Gradiont and Combination Microphones CHAPTER 7, DIRECT-RADIATOR LOUDSPEAKERS vant XVII Basic Theory of Divect-radtotor Loudspeakers Pant XVIII Design Factors Affecting Direc-radiator Lowdepeater Performance 6 16 a a 62 0 1 a o 109 16 16 128, 14 a 150 18, 183 1832 CONTENTS JAPTER 8, LOUDSPEAKER ENCLOSURES vane XIN Simple Bnolonures pant XX Baswreler Bnelosures CHAPTER 9, HORN LOUDSPEAKERS. pant XXE_ Horm Driving Wits rawr -XXIL Horns CHAPTER 10. SOUND IN ENCLOSURES ant XXII Sound Fields in Small Regularly Shaped Enclosures pant XXIV Sound Fielda in Large Irregularly Shaped Bnelosures rant XXV_ Sound Tronemiseion throuph Walls between Enclosures CHAPTER 11, NOISE CONTROL pant XXVI Procedures and Sources vant XXVIL Acoustic Pronemission Pathe CHAPTER 12, ACOUSTIC MEASUREMENTS pane XXVIT Measurement of Acoustic Lele vant XXIX Reciprocity Calibration of Transducers CHAPTER 13. HEARING, SPEECH INTELLIGIBILITY, AND PSYCHOACOUSTIC CRITERIA pane XXX Hearing Pant XXX Speech Inteligibility rant XXXII Poychoacoustic Criteria PROBLEMS. DIX J, DECIBEL CONVERSION ‘TABLES APPENDIX Il, CONVERSION FACTORS ‘APPENDIX IL APP! INDEX 208 208 239 259 250 268 285 285 298 at 392 saz 45 361 361 317 388 388 406 az 431 464 469 479 PREFACE TO THE PAPERBACK EDITION With the advent ofthe compact dise, with miniature high-fidelity systems ambulating everywhere, and with emphasis on combination voice and | in'a medium Gee Fig: 2 where the incremeefromlefciorahtatacpace rate | wound bremure» cheng in apece at 2 a op/or (re Pag 2) pace rato gradp = 128 4 522 4 2 where i,j and ae unit velar in the and diectins, respectively, and Pit fe remire ats pone Assume that the sides of the box are completely frictionless; i.e, any viscous drag between gas particles inside the box and those outside is negligible. ‘Thus the only forces acting on the enclosed gas are due to the pressures at the faces of the box. The difference between the forces acting on the two sides of our tiny’ box of gas is equal to the rate at which the force changes with distance times the ineremental length of the box: Forge seting bo acoerat he box Inthe | Fore ating ta aelrte the box in the positive x dretion = ~ (32.2) ay a | postive dieton = ~ [1 (222) ay ae te) | 4 (2249) see + (22 2) ae ay] an) Note that the positive gradient causes an acceleration of the box in the negative direction of 2 + Nonvector derivations of the wave equation are given in Rapleigh, “Theory of Sound,” Vol. 2,pp. I-15, Macmillan & Co.,[ih, London, 1806; P.M. Morse, “Vibra tion and Sound," 2d ed, pp. 217-225, MeCiraw-Hill Book Company, Ine., New York, 1948; L. B. Kinsler and A. R. Frey, “Fundamentals of Acoustic” pp. 118-137, John Wiley & Sons, Ine, New York, 1950; R. W. B. Stephens and A. F, Bave, “Wave Motion and Sound,” pp. 32-13, 400-406, Edward Amold & Co,, London, 1950; and other pes, A vector derivation of the wave equation ie given in two papers that must be read fogethor: W.J. Cunningham, Application of Vector Analysis tothe Wave Equation, J Acoust Soe. Amer, 22: 61 (1950); nd R. Vo Hartley, Note on “Application of ‘Vector Analysis to the Wave Equation,” J. Acous, So. Amer, 22: 511 (1950).18 THE WAVE EQUATION AND SOLUTIONS — (Chap. 2 f--2 aw fener a By Newton's aw, the foree per nit volume (/7) of Ba. (2) must be equal tothe tine rie of change othe momentum prunit vlan othe box. We have already assured tat our box i dorm packet thatthe sas of fhe gas within it lays constants "That is fo Neyo % sa a Dg Vat ae send FB an average velocity of the gus | where qis the average vector velocity of Jn the "box" in the direction, oi the | the gas in the box,” 9” is the average space average of the instantaneous den- | density of the gas in the box, and sity of the gas inthe box, and MC = p'V | AF = p'V isthe total mass of the gas in fi the otal mess ofthe gas in the bor, | the box, D/Dt ia not a simple partial erivative but represents the total rate of the change of the velocity of the particular bit of gas in the box regardless of its position, Pa 29 ee) where tg and gar the component She set paride vest hanes in density ofthe gas due | "iF te veto prt vlsty nema nd wove isemal enough, then | enough, the stee! change of womestam stares danty 9h appro | fe pres n the ean be apa tnsnly ual tbe average dency, | indy the ate of change of omens Thee, Samat dned pointy Dq/DP = 9/95 and the instantaneous" ewity. pea be throsinated bythe average dealy po thn, ae “at p e 2 (2a) ‘The approximations just given are generally acceptable provided the sound pressure levels being considered are below about 110 db re 0.0002 microbar. Levels above 110 db are so large as to create hearing dis- comfort in many individuals, as we shall see in Chap. 13 at the end of this book, The Gas Law. If we assume an ideal gus, the Charles-Boyle gas law applies to the box. It is Py =RT es) where P is the total pressure in the box, V is the volume equal to Ar By 42, T is the absolute temperature in degrees Kelvin, and R is a ‘constant for the gas whose magnitude is dependent upon the mass of gas, Pact 11 THE WAVE EQUATION 0 chosen. Using this equation, we can find a relation between the sound pressure (excess pressure) and an incremental change in V for our box. Before we can establish this relation, however, we must know how the temperature 7’ varies with changes in P and V and, in particular, whether the phenomenon is adiabatic or isothermal ‘At audible frequencies the wavelength of a sound is long compared with the spacing between air molecules. For example, at 1000 eps, the wavelength 4 equals 0.34 m, as compared with an intermolecular spacing of 10° m, Now, whenever a portion of any gas is compressed rapidly, its temperature rises, and, conversely, when it is expanded rapidly, its temperature drops. At any one point in an alternating sound field, therefore, the temperature rises and falls relative to the ambient, temperature. This variation oceurs at the same frequency as that of the sound wave and is in phase with the sound pressure, ‘Let us assume, for the moment, that the sound wave has only one frequency. At points separated by one-half wavelength, the pressure and the temperature Quctuations will be 180° out of phase with each other. Now the question arises, Is there sufficient time during one-half an alternation in the temperature for an exchange of heat to take place between these two points of maximally different temperatures? Tt has been established? that under normal atmospheric conditions the speed of trave) of a thermal diffusion wave at 1000 eps is about 0.5 1n/se ‘and at 10,000 epsit is about 1.5 m/see. The time for one-half an alterna- ‘ion of 1000 eps is 0.0005 sec. In this time, the thermal wave travels a distance of only 0.00025 m. This number is very small compared with one-half wavelength (0:17 m) at 1000 eps. At 10,000 eps the heat travels, 7.5 X 10-*m, which is a small distance compared with a half wavelength (1.7 X 10? m), It appears safe for us to conclude, therefore, that there is negligible heat exchange in the wave in the audible frequency range. Gaseous compressions and expansions of this type are said to be adiabatic, For adiabatic expansions, the relation between the total pressure and the volume is known to bet PV? = constant 26) where + is the ratio of the specific heat of the gas at constant pressure to the specifie heat at constant volume for the gas. This equation is 130 4 mass of gas i choren so that ite weight in grams is equal to ite molecular weight Genowa to chemists as the gram-molecular weight, or the mole, then the Yolume of thia mace st O°C and 0.76 m Hg is the sue for all gates and equals 0.02242 m?. ‘Then R= 8314 watt-see per degree centigrade per gram-moleculae Weight, If the mass of gas chosen ie times its molecular weight, thea e = 8.314 ‘Soe LL, Beranck, "Acoustic Measurements,” p. 49, John Wiley & Sous, Tne, Now York, 1940 "MW. Zemaasky, “leat and "Thermodynamics," 24 ed, pp. 104-114, MoGran- Hil Book Company, Inc, New York, 1,EQUATION AND SOLUTIONS — (Chap. 2 20 THE WAVE ‘obtained from the gas law in the form of Eq. (2.5), assuming adiabatic conditions. For air, hydrogen, nitrogen, and oxygen, é¢., gases with diatomic molecules, yal Expressing Eq, (2.6) in differential form, we have av an Let P=Petp V=Vetr (28) where Ps and Vs are the undisturbed pressure and volume, respectively, fand p and r are the incremental pressure and volume, respectively, owing ry ay vagesioe as ee opiet Beara oo _ o Fis, 22. Change in volume of the box with change in position. From (a) snd (6) I acen that the inesomentel change im volume of the box equals + = (96,/9¢) S to the presence of the sound wave. Then, to the same approximation as that made preceding Eq, (2.4) and because p « Po and 1 Vo, Ra-F 29) The time derivative of this equation gives Ag-72 ax ‘The Continuity Equation. The continuity equation is s mathematical expression stating that the total mass of gas in a deformable “box” must remain constant. Because of this law of conservation of mass, we are able to write a unique relation between the time rate of change of the ineremental velocities at the surfaces of the box. aot 11 THE WAVE EQUATION a Ea See a | et ibae ‘The diflerence of the two qvantitios above multiplied by the area ay 3z ives | the increment in volume + ot, Warsyar ane | Yeave eam) ved a1 Differentiating with respect to time | Diflrentiating with respict to. time veld, yields, 132) | a Fa vodivg 2.30) Where aj the instantincons particle | where it the instantaneous particle velocity velocity The Wave Equation in Rectangular Coordinates One-dimensionol Derivation Threedimensional Devnation "The one-dimensional wave equation ig | The three-dimensionsl wave equation obtained by combining the equation of | ix obtained by combining the equation of ‘mation (24a), the gus las (2.10), and the | motion (2-40), the gas law (2.10), andthe continuity equation (2.134), Combing | eontinuty equation (2.125). Combine: tom of (210) and (2.14) gives tion of (210) and (219) gives Dierotte (1 with pect o | Dierntate (2.4) wih rape to wre eet) a Piderevte a) with apes to | Ta the degen of ech sd of -% ace | 2 . ~ di gat ~ mar 220822 THE WAVE EQUATION AND SOLUTIONS — [Chap.2 One-dimensional Derivation ‘hveeimensionat Derivation Replacing the div (grad p) by Wp, wo get vip = podiv ¥ tp = pediv 8 (a7) interchangesbilty of the 2 | where vis the operator called the Lapla- and £ derivatives, and combining (2.152) | cian. Combining (2.199) and (2.17), we tnd (2162), we get ee ~ 52% ei v essay Let us, by definition set, we eat (2.19 oe (2.19) ‘We shall see later that cis the speed of propagation of the sound wave in the medium, We obtain the onedimensional wave | We obtain the three-dimensional wave cenwation caution re) vende aay Jn rectangle coordinates vip 2 4 OF 4 OP pe ME MP any We could als ave sited p and | We sould ao have eliminated p and retained in whih ease we would have | rsttnedq, im whith ease we would have 533 (222%) oy Law where 4g = grad (div q) when there is a (2220) | Co rotation in the medium, Equations (2.20) and (2.22) apply to sound waves of “small” magni- ‘tude propagating in a source-free, homogeneous, isotropie, frictionless gas at rest, The Wave Equation in Spherical Coordinates, ‘The one-dimensional ‘wave equations derived above are for plane-wave propagation along one dimension of a rectangular coordinate system. In an anechoic (echo- free) chamber or in free space, we frequently wish to express mathe- ‘matically the radiation of sound from a spherical (nondireetional) source of sound. In this ease, the sound wave will expand as it travels away. from the source, and the wave front alvrays will he a spherical surface. To apply the wave equation to spherical waves, we must replace the ‘operators on the left side of Eqs. (2.20) and (2.22) by operators appro priate to spherical coordinates, Assuming equal radiation in all directions, the wave equation in one= dimensional spherical coordinates is ap oe (223) Pert IV] SOLUTIONS OF THE WAVE RQUATION 23 Simple differentiation will show that (2.23) can also be written apr) _ 1 HG) ar Bae Ce) Itis interesting to note that this equation has exactly the same form as Eq (2.204). Hence, the same formal solution will apply to either equa tion except that the dependent variable is p(z,¢) in one case and pr(r,t) in the other case. ‘Example 21. In the steady stat, that ia, 2u/2¢ = jou, determine matheratially hhow the sound pressure in a plane progreetive sound wave fone ‘ould be determined fom ineasurement of partsle veloeity alone Solution. From Bq. (2-4a) we find in the steady state that where p and ware now rms Lively,” Written in diferent of the sound pressure and particle velocity, respme= form, 1 the particle velocity is 1 em /e0, wit 1000 radians/ste, and ais 0 em, then {8p = =30.005 X 1000 1.18 x a0 31059 newton mn? ‘We shall have an opportunity ia Chap, 6 of this test toate @ practical application ‘of these equations to the mesturement of particle vlocty hy a velocity microphone pant IV Solutions of the Wave Equation 2.3. General Solutions of the One-dimensional Wave Equation, ‘The ‘one-dimensional wave equation was derived with either sound pressure or particle velocity as the dependent variable. Particle displacement, or the variational density, may also be used as the dependent variable. ‘This can be seen from Eqs. (2.4a) and (2.18a) and the conservation of mass, which requires that the product of the density and the volume of a small box of gas remain constant. That is, 2'V = poVo = constant, (2.25) and so pave = —Vae (2.26) Let p= pte (2.27) where pis the incremental change in density. Then, approximately, from Eqs. (2.8) and (2.26), per = —Vep (2.28)a THE WAVE EQUATION AND SOLUTIONS —[Chep.2 Differentiating, ar _ Vode ee a asetl so that, from Eq. (2.132), 2 au et nese (2.29) Also, we know that the particle velocity is the time rate of change of the particle displacement. watt e220) Inspection of Eqs. (24a), (2.180), (2.29), and (2.30) shows that the pressure, particle velocity, particle displacement, and variational density are related to each other by derivatives and integrals in space and time. ‘These operations performed on the wave equation do not change the form of the solution, as we shall see shortly. Since the form of the solution is not changed, the same wave equation may be used for determining density, displacement, or particle velocity a well as sound pressure by substituting p, or &, or u for p in Eq. (2.20a) o p, &, oF q for p in (2.200), assuming, of course, that there is no rotation in the medium, General Solution. With pressure as the dependent vatiable, the wave ‘equuataon 38 ap _ lap Bet ae (2a) ‘The gencral solution to this equation is a sum of two terms, p-t-an(+3) 82) where fi and fs are arbitrary functions, We assume only that they have continuous derivatives of the first and second order. Note that hecause tand x occur together, the first derivatives with respect to z and t are exactly the same except for a factor of +e. ‘The ratio z/c must have the dimensions of time, so that e is a speed. From ¢ = yPe/pe (Eq. (2.19)] we find that 108\" 18, in airat an ambient pressure of 10° newtons/im? and at 22°C. ‘This quan- lity is nearly the same as the experimentally determined value of the speed of sound 344.8 [see Eq. (1.8)], 80 that. we recognize c as the speed at which a sound wave is propagated through the ait, From the general solution to the wave equation ‘observe two very important facts: ven in Eq, (2.82) we Pat IV] SOLUTIONS OF THE WAVE EQUATION 25 1. The sound pressure at any point x in space can be separated into ‘two components: an outgoing wave, f(t — z/c), and a backward-traveling, wave, fall + 2/0) 2, Regardless of the shape of the outward-going wave (or of the backward-traveling wave), it is propagated without change of shape. To show this, let us assume that, at ¢ = f,, the sound pressure at x — 0 is f(t). At a time ¢ = ty + t the sound wave will have traveled a distance x equal to tye m. At this new time the sound pressure is equal to = fills + te ~ bec/e) = full). In other words the sound pressure has ropagated without change. ‘The same argument can be made for the backward-traveling wave which goes in the —z direction. It must be understood that inherent in Eqs. (2.31) and (2.92) are two assumptions. First, the wave is a plane wave, i, it does not expand laterally. ‘Thus the sound pressure is not a function of the y and coordinates but is a function of distance only along the x coordinate. Socond, it is assumed that there are no losses of dispersion (scattering of the wave by turbulence or temperature gradients, et.) in the air, so that. the wave does not lose energy as it is propagated. Dissipative and dis- persive cases are not treated analytically in this book, but are discussed briefly in Chaps. 10 and 11 Steady-state Solution. In nearly all the studice that we make in this ‘text we are concerned with the steady state. As is well known from the theory of Fourier series, a steady-state wave can be represented by a Jinear summation of sine-wave functions, each of which is of the form VO = V2 orl cos (st + 6 (2.330, For example, if ys is sound pressure, we write 2 = Spt) = Y V2 [pel 08 (ot + 0.) (2.380 where w, = 2rf,;f, = frequency of vibration of the »th component of the wave; 9, is the phase angle of it; and /2 |¢| (or |p") is the peak magnitude of the component. Because the wave is propagated without change of shape, we need consider, in the steady state, only those solutions to the wave equation for which the time dependence ateach point in space Js sinusoidal and which have the same angular frequencies w, as the source Borrowing from electsical-cieuit theory, we represent & sinusoidal funtion with a frequency w hy the real part of a complex exponential function, Thus, ata fixed point in space 2, we have the sound pressure, YG) = VIRe lolze"] (2.340) PUz,t) = V2 Re [plz )e (2.340)26 INE WAVE EQUATION AND SOLUTIONS — (Chap. 2 where p(x) is a complex function (fe, it hes a real and an imaginary part) that gives the dependence of p on z. The product of +/2 times the magnitude of p(z) is the peak value of the sinusoidal sound pressure function at z, ‘The phuse angle of p(z) is the phase shift measured from some reference position, Generally we omit writing Re although it always must be remembered that the real part must be taken when using the final expression for the sound pressure, In the steady state, therefore, we may replace fs and fs of Eq. (2.82) by a sum of functions each having a particular angular driving frequency Ww, s0 that PD = Y pales) = Y VERE lipyremore + promenjeee] (2.35) ‘The part of Eq. (2.85) within the brackets is the same as that within the brackets of Eq. (2.34). ‘The factor +/2 is introduced s0 that later p.* and p_' may represent complex rms functions averaged in the time dimension. ‘The + and ~ subscripts indicate the forward and backward traveling waves respectively. tis apparent that the first term of Hq. (2.85) represents an outward- traveling wave whose rms magnitude |p,*| does not change with time { or position z. A similar statement may be made for the second term, which is the backward-traveling wave. It is customary in texts on acousties to define a wave number k, ee kat. Blo (2.36) Also, let us drop Re and the subscript » for convenience. Any one term of Eq. (2.35), with these changes, becomes PEED = V2 [Dla)er] = V2 (Dyer + peer) (2.37) Similarly, the solution to Eq. (2.22a), assuming steady-state conditions, ula) = VF (get + u_eitatey (2.38) tis understood that the real part of Eqs. (2.37) and (2.38) will be used in the final answer. The complex magnitudes of p, and p-or wy and ware determined from the boundary conditions ‘The complex rms pressure and particle velocity are found directly from Bqs. (2.37) and (2.38) by eanceling -/Ze™ from the right-hand sides, When the remaining fonetion is converted into magnitude and phase angle, the magnitude ts the quantity that would be indicated by a rms sound pressure meter. Note, however, that when we take the real part of P(e.) 0F w(t), the quantity +/2 e must bein the equation if the proper values for the instantaneous pressure and particle velocity are to be obtained Pat lV] SOLUTIONS OF THE WAVE EQUATION 27 Brample 22. Assume tha for the steady state, at x point x = 0, the sound pres rare in a onedimensional outward-raveing wave has the recurrent form shown by the dotted curve in the sketch below. This wave form i ven by the real part of the equation P10.) = VF (er 4 2 (e) What are the partite velocity and the particle displacement «function of time st z= 5m? () What ate the rms valies of these two quantities? (c) Are the ‘ins values dependent upon 2? ° 3005 ‘oor Tas seconds, Solution. a. We have for the solution of the wave equation giving both = and ¢ (sco Eq. (2.371 nia = Va Prom Ba 2s) wee hat = ch omen wie) = jh wat = 3 wed And from Hq, (280) we have PB (4. sansa 4 2B grimuerae Sow a8" ies Aba = Sm, 2/2 = 5/348 = 00145 see ata tat) = 23 (aesoneosian 4 detmaresieny sod tt = 2 (igen sto etme) ‘aking the apart ofthe two preci equation size) = 2 co (ots — ut) +200 cts 2.9 ean = 23 phgnin ou 0.) + 52 sn ns —270)] Note hat enh rm nthe parte dplcemen i on of in pase withthe velocity and Use the wave shape is different. As might be expected, differen tistion ‘nphasioes the28 THE WAVE EQUATION AND SOLUTIONS —(Chap.2 ‘Thoee equations are plotted below: q q 2. The rms magnitude of a sine wave is equal to ite peak value divided by VB. ‘This may be verified by squaring the cine wave and Gnding the average value over one cyele and then taking the square root of the rat. If two sine waves of diferent square root of the sums of the squares of the individual rms magoitudes, 99 that iby VEER ~ 0011 m/tc0 tn de VGR) + Ga) <8 x10 Thoms values w and é, are independent of x fora plane progressive sound wave, 24. Solution of Wave Equation for Air in a Rigidly Closed Tube. For this example of wave propagation, we shall consider a hollow cylindrical Devin ston Holiow yd tube fig emiston 7 = t t tele re, 23. Rigidly terminated tube with rigid side walls. ‘The velocity at = value of 72 ny com al 9 /ce has tube, closed at one end by a rigid wall and at the other end by a flat vibrating piston (soe Fig. 2.3). ‘The angular frequency of vibration of PatiVj SOLUTIONS OF TH WAVE EQUATION 29 the piston is w, and its rms velocity is ue. We shall assume that the diameter of the tube is sufficiently small so that the waves travel down the tube with plane wave fronts. In order that this be true, the ratio of the wavelength of the sound wave to the diameter of the tube must be ‘greater than about 6. Fo. 24, Porton of the tuhe showing the diection and magnitude of movement ofthe ‘ie particles sv fnction of LAL position , the particle volocity and displace- ent ere maxituin, At position B, they are zero. Particle Velocity. ‘The form of solution we shall select is Eq. (2.38). Set tis equal to the rims velocity of the vibrating piston at x = 0, and set Lequal to the length of the tube, The boundary conditions are Atz = 0, u(0,) = V2 we, so that aye Ata = thu = 0, s0 that ef uel = 0 239) Remember that sny = Se Hence = xahsin Al (2.40) and wet ; wn ean) whieh gives ws lad) = VB te LS ny sin k(t — 2} un wk 49) Note that the «/% and the time exponential have been left out of Eq (243) so that both ve and w are complex rms quantities averaged ove time, Refer to Fig, 24. If the length Land the frequency are held constant30 THE WAVE EQUATION AND SOLUTIONS —{Chap. 2 the particle velocity will vary from a value of zero at x = 1 to a maxi mum at |~ z= 2/4, that is, at J — x equal to one-fourth wavelength. In the entire length of the tube the particle velocity varies according to a sine function. Between the end of the tube and the d/4 point, the oscillatory motions are in phase. In other words, there is no progressive phase shift with z. This type of wave is called a standing wave because, in the equation, z and cf do not occur as a difference or & sum in the argument of the exponential funetion. Hence the wave is not, propagated In the region between 1 — z= /4 and lz =/2, the particle velocity still has the same phase except that its amplitude decreases sinusoidally. At I— x =/2, the particle velocity is zero, In the region between | — x = 2/2 and 1 ~ x = the particle velocity varies with 2 according to a sine function, but the particles move 180° out of phase with those between 0 and 2/2. This is seen from Eq. (2.43), wherein the sines of arguments greater than © are negative. If we fix our position at some particular value of x and if Lis held constant, then, as we vary frequency, both the numerator and denominator of Eq. (2.43) will vary. When A is some multiple of x, the particle velocity will become very large, except at x= 0 or at points where E( ~ x) isa multiple of x, that is, at points where — z equals multiples of X/2. ‘Then for kl = nz (2.44) Equation (2.43) would indicate an infinite rms velocity under this condition. In reality, the presence of some dissipation in the tube, which was neglected in the derivation of the wave equation, will keep the particle velocity finite, though large. ‘The rms particle velocity 1 will he zero at those parts of the tube where k( — 2) = nefandnisan integer or zero. That is, (2.45) In other words, there will be planes of zero particle velocity at points along the length of the tube whenever 1 is greater than X/2. Some examples of the particle velocity for 1 slightly greater than various multiples of 4/2 are shown in Fig. 2.5. Two things in particular are apparent from inspection of these graphs. First, the quantity n deter mines the approximate number of half wavelengths that exist between the two ends of the tube. Secondly, for a fixed wu, the maximum velocity 1 For the type of source we have assumed and no diaspation, this ease breaks down for Hl = ne Pat IV] SOLUTIONS OF THE WAVE EQUATION 31 of the wave in the tube will depend on which part of the sine wave falls atx = 0. For example, if 1 — nd/2 = /4, the maximum amplitude in the tube will be the same as that at the piston. If] — nh/2 is very near zero, the maximum velocity in the tube will become very large. Let us choose a frequency such that n = 2 as shown, Two factors determine the amplitude of the sine function in the tube. First, atz = 0 the sine eurve must pass through the point we. Second, atx = I the sine curve must pass through zero. It is obvious that one and only one sine wave meeting these conditions can be drawn so that the amplitude is determined, Similarly, we could have chosen a frequency such that a2) — FH aut x0 a Fic. 25. Variation ofthe particle velocity (x for =0, 482 function ofthe distance along the tue of Fig 2 for thee frequencies, ce, forthree wavelengths. ALx = 0, therms particle velocity the particle velocity is zero. The period T= 1/f n « 2, but where the length of the tube is slightly less than two half wavelengths, Tf this ease had been asked for, the sine wave would have started off with a positive instead of a negative slope at x = 0. Sound Pressure. ‘The sound pressure in the tube may be found from the velocity with the aid of the equation of motion (Eq. (2.4a)], whieh, in the steady state, becomes p= ~iepfude (2.46) ‘The constant of integration in Eq, (2.46), resulting from the integration of Eq, (24a), must be independent of 2, because we integrated with32 TH WAVE EQUATION AND SOLUTIONS — [Chep. 2 respect to x. ‘The constant then represents an increment to the ambient. pressure of the entire medium through which the wave is passing. Such An increment does not exist in our tube, so that in Eq, (2.46) we have set the constant of integration equal to zero. Integration of Eq. (2.46), after we have replaced u by its value from Eq. (2.42), yields pe) = “Ine VE na HL 2 ean) P= ~ineuy 2H 2) 4s) Note that the /% and the time exponential have been left out of Eq. 2.48) so that both p and tz are complex rms quantities averaged in time. ‘The rms pressure p will be zero at those points of the tube where i(U = 2) = nx + 7/2, where n is an integer or zero, (2.49) The pressure will equal zero at, one or more planes in the tube whenever 1 s greater than )/4. Some examples are shown in Fig. 2.6. Here again, quantity n fs equal to approximate number of half wavelengths in tube Refer once more to Fig. 2.5 which is drawn for ¢= 0. ‘The instan- aneous particle velocity is at its maximum (as a function of time). By comparison, in Fig. 2.6.at ¢ = 0, the instantaneous sound pressure is zero. At a later time = 7/4 = 14f, the instantaneous particle velocity has vecome zero and the instantancous sound pressure has reached its m: num. Equations (2.42) and (2.47) say that whenever k(0 ~ 2) is a small number the sound pressure lage by one-fourth period behind the yarticle velocity. At some other places in the tube, for example when 1 ~ 2) lies between d/4 and 2/2, the sound pressure leads the particle relocity by one-fourth period. ‘To see the relation between p and u more clearly, refer to Figs. 2.5 and 2.6, for the case of n = 2. In Fig. 2.5, the particle motion is to the right whenever w is positive and to the left when it is negative. Hence, at the 2/2 point the particles on either side are moving toward each other, 80 hat one-fourth period later the sound pressure will have built up to a naximum, as can be seen from Fig. 2.6, At (I — z) = 4/2, the particles wre moving apart, so that the pressure is dropping to below barometric ws ean be seen from Fig. 2.6 Figures 2.5 and 2.6 also reveal that, wherever along the tube the magni- ude of the velocity is zero, the magnitude of the pressure is a maximun und viee versa. Hence, for maximum pressure, Eq. (2.45) applies. Specific Acoustic Impedance. It still remains for us to solve for the ipevifie acoustic impedance Z, at any plane x in the tube. Taking the Part IV] SOLUTIONS OF THE WAVE EQUATION 33 ratio of Eq. (2.48) to Bq. (2.43) yields = —iowe cot Ml! = 5X, mks rayls (2.50) where X, is the reactance. Where we have set 1- (21) ‘That is, U is the astance between any plane x in Fig. 2.3 and the end -———— wa ——___ ry] oT ft aan ME rel Y : omer | POEUIET, 0 sl Fa, 2.6, Variation of the sound pressure p(s) as «function of the distance along the tube for three frequencies, i.e, for three wavelengths, Atz = 0, the rms particle Velocity is we and at x =f, itis zero, ‘The period 7 equals 1/f of the tube at L The —j indicates that at low frequencies where cot kl! & 1/k’ the particle velocity leads the pressure in time by 90° and the reactance X,is negative. Atall frequencies the impedance is reactive and either leads or lags the pressure by exaetly 90° depending, respec tively, on whether X, is negative or positive. ‘The reactance X, variuM THE WAVE EQUATION AND SOLUTIONS — (Chop. 2 us shown in Hig. 2.7, If the value of kl’ is small, we may approximate the cotangent by the first two terms of a series 1a cot ht = yp — (2.52) ‘This approximation is valid whenever the product of frequency times the distance from the rigid end of the tube to the point of measurement is very small. If the second term is very small, then it may be neglected with respect to the frst, 2 ft fo ne vty i : "hI ' 1 | ! ! ! 1 | ! ! } | b I | | +e t 0 i i wl A 1 I | fo ry eM 0 Fro, 27. Tho specific acoustic reactance (Prne/tem) along the tube of Fig. 2.3 for a particular frequency, sc, » particular wavelength where 3(4/2) isa litle less than the fue length For this ease, the numberof zeros is 3, and the number of poles is Let us see how small the ratio of the distance to the wavelength \ must be if the second term of Eq. (2.52) is to be 3 per cent or less of the first term. That is, let us solve for I'/A from 2al" d 3) By 5 0.03 (2.53) which gives us ve : 5 = 0.05 (2.54) In other words, if cot ki’ is to be replaced within an accuraey of 3 per cent by the first term of its series expansion, I’ must be less than one-twentieth wavelength in magnitude, Assuming I" < 4/20, Eq. (2.50) becomes 1 1 pot _ re ae 2m ikem —i5h = a0 7 pat = jo, HB rays (2.55) Heneo, the specific acoustic impedance of a short length of tube can be Part IV} SOLUTIONS OF THE WAVE EQUATION 35 represented as a ‘capacitance” elle specific acoustic compliance, of mes; nitude C, = U'/poc*. Note also that C, = '/7Ps, because of Eq. (2.19). The acoustic impedance is of the same type, except that an area factor sppears so that u-2-— pet Bu ~ jo(V Joo) ~ jul, where V = I'S is the volume and S is the area of cross section of the tube. C, is called the acoustic compliance and equals V/pec?. Note also that C4 = V/yPo, from Eq. (2.19). ‘mks acoustic ohms (2.56) Example 2.8. A cylindrical tube isto be used in an acoustic device as an impedance clement. (a) ‘The impedance desired is thet of a compliance. What length should it have to yield w reactance of 1.4 X 10% mks ray xt an angular frequency of 1000 radiana/sec? (6) What is the relative magnitude of the fir. and second terms of Bq. (252) for this case? Solution. ‘The reactance of auch a tbe is P @ en 14x10 he Henee, = 0.1m, o e renee, the second term is about 3 per eent ofthe fist term. 2.6, Freely Traveling Plane Wave. Sound Pressure. If the rigid termination of Fig. 2.3 is replaced by a perfectly absorbing termination, a backward-traveling wave will not occur. Hence, Eq. (2.37) becomes Plat) = VE pyemern 7) where p. is the complex rms magnitude of the wave. This equation also applies to a plane wave traveling in free space. Particle Velocity. From Eq, (2.4a) in the steady state, we have 1 op un-- 12 58) Japa 3 58) Hence, V2 Pe pcre = PEL u(t) = rr a pot (2.59) ‘The particle velocity and the sound pressure are in phase. This is mathematical proof of the statement made in connection with the qualitative discussion of the wave propagated from a vibrating wall in Chap. 1 and Fig 11 Specific Acoustic Impedance. The specific acoustic impedance is 2, =F = pe mks rayls (2.60)36 THE WAVE EQUATION AND SOLUTIONS — [Chap.2 ‘This equation says that in a plane freely traveling wave the specific acoustic impedance is purely resistive and is equal to the product of the average density of the gas and the speed of sound, This particular ‘quantity is generally called the characteristic impedance of the gas because its magnitude depends on the properties of the gas alone. It is a quantity that is analogous to the surge impedance of an infinite electrical line. For air at 22°C and a barometric pressure of 10° newtons/m?, its magnitude is 407 mks rayls. 2.6. Freely Traveling Spherical Wave. Sound Pressure. A solution to the spherical wave equation (2.24) is vi ec + acm) a 261) where A, is the magnitude of the rms sound pressure in the outgoing wave at unit distance from the center of the ephere and A_ is the same for the reflected wave. If there are no reffecting surfaces in the medium, only the first term of this equation is needed, i.e., Pert) Pen) = VEAL pu 0) Varticle Velocity. With the aid of Eq. (2.48), solve for the particle velocity in the r direction, ure) VE oo( 43h) VEL (14 be 203) Specific Acoustic Impedance. ‘The specific acoustic impedance is found from Eq. (2.62) divided by Eq. (2.63), kr peckr BR Sore 7 (90° = tan mks (2.64) oe ~ i kr rayls (2.64) Plots of the magnitude and phase angle of the impedance as a function of kr are given in Figs. 2.8.and 2.9. The real and imaginary parts, R, and X,, are plotted in Fig. 2.10. Por large values of kr, that is, for large distances or for high frequencies, this equation becomes, approximately, Z.% poe mks rayls (2.65) The impedance here is nearly purely resistive and approximately equal to the characteristic impedance for a plane freely traveling wave. In other words, the specific acoustic impedance a large distance from a spherical source in free space is nearly equal to that in a tube in which no reflections pcour from the end opposite the source. Pat IV} SOLUTIONS OF va = 1 ee 14 ete oe | See tncretmceste, CHEV | gt REE a - £ a gL | od | okt oe | ry t coe * ‘ood “oa 05 10 5 10 20 (erat feat r/R) Fo, 28. Plot of the magnitude of the specific acousticsimpedance ratio !Z/ow in a spherical feeely traveling wave as a funetion of kr, where is the wave number equal tow/ror 2e/sand ris the distance from the center of the spherical source. [21] isthe ‘magnitude of ratio of pressure to particle velocity in a spherical free traveling weve, sand pr ie the characteristic impedance ef wit so Ao * LUI iz 7} + itt 4 « a . t : 5° ani ol | lt" ercraegeetis NN £40) particle velocity 3 Bisiecaate Ey : {ty | i | Pa 9 - ih — + . ee oo * “ood “ox 08° "10, 5° 10 = Uirytorfee(200) Fie, 29. Mot ofthe phate angle, n degre ofthe pee acousticmpedance rt Zulosc in a spherical wave a3 8 function of kr, where kis the wave number equal to wfeor 2r/a, und ris the distance from the centr of the spherical source.38 THE WAVE EQUATION AND SOLUTIONS — [Chap.2 Equations (2.62) and (2.63) are significant hecause they reveal the difference between the responses of a microphone sensitive Lo pressure and a microphone sensitive to particle velocity as the microphones are brought close to a small spherical source of sound at low frequencies. AS ris made smaller, the output of the pressure-responsive microphone will double for cach halving of the distance between the microphone 29) 1 0s| 03] 02 oa 005} 03] 0.02] Po 0.005 0.008] 002] Irapedance con ‘aalogy 0.0005} 0.0008] 0.0002] 00014 = ire oor * G05 “02 03 ‘10 510 20 be Fio, 2.10. Teal and imaginary parts of the normalized spociie acoustic impestance Z.Joue of the air loul ons plating sphere of radius r loeated in free spac. Fe {quency is plotted on a norinalized seale where kr = 2xjr/e = 2er/d.” Note als that the ordinate is equal to Zu/owS, where Zw i the mechaninl impedance, nil to ZaS/no, whore Zs the acoustic impeance. The quantity 8 is the aren for which the impedance is being determined, and py ig the characteristic impesanc: of the edie center of the spherical source. Expressed in deeibols, the output increases Gdb for cach halving of distance, For the velocity-responsive microphone, the output variation is not so simple. Only at sufficiently large distances (kr* >> 1) does the output increase 6 dh for each halving of distance. For shorter distances the second term inside the parentheses on the right-hand side of Eq. (2.63) becomes large, and the magnitude of w increases at a rate exceeding +6 db for each halving of distance. For Pat VJ SOLUTIONS OF THE WAVE EQUATION very short distances (Kr? < 1), the rate of increase of w approaches : limit of +12 db for each halving of distance. It is for this reason tha the voice of a radio crooner sounds “bassy”” when he sings very neat tot velocity-sensitive microphone which was designed to have ite best response when located a large distance from the source of sound ‘Another significant thing is to be learned from Eq, (2.64). At low fre quencies it is very difficult to radiate sound energy from a small loud speaker. A small loudspeaker may be likened to a pulsating balloon o! some small radius r. The specific acoustic impedance Z, of the air pre sented to each square centimeter of the balloon is given by Eq. (2.64) and Fig. 2.10. At low frequencies, the impedance becomes nearly purely reactive, and the resistance becomes very, very small, Hence, the powe! radiated by a small loudspeaker becomes very small. At high fre. quencies, kr > 2, the impedance Z, becomes nearly purely resistive and has its maximum value of pc, s0 that the power radiated for a given value of p, reaches its maximum. ‘The important steady-state relations derived in this chapter are sum. marized in Table 2.1. ‘TABLE 21. General and Steady-state Relations for Small-signal Sound Propa. gation in Gases ame Genel equation | Steuyatate euatin Wave equations 718 Joy ya 22) X= Bar apr) _ 1 apr) oF 7a ae Equation of motion oa Bm Mfume Jude grad p= Jao Diplacement Incremental density Incremental temperature40 THE WAVE EQUATION AND SO TIONS — [Chap. 2 parr V Energy Density and Intensity 2.7, Bnergy Density. Energy density is an important concept in acoustics because, in dealing with sound in enclosures, it is necessary to study the flow of energy from a source to all parts of the room. The energy density, ie., watt-seconds per unit volume, is greater near the source than farther away and is the variable that appears in the equations describing the acoustical conditions. On the other hand, the ear and most sound-level meters respond to rms sound pressure. We need to ascertain, therefore, the relation between energy density and sound pressure in sound fields. ‘The energy density associated with the small “box” of gas at any par~ ticular instant is the sum of the kinetic and potential energies per unit volume of the air particles in the box. ‘The kinetic energy density due to the excess pressure of the sound wave Drx is 1 Mut _1 Dax = 5 ye = yee (2.66) where w is the average instantaneous velocity of the air particles in the ox, pe is the average density, and 3f/V is the mass per unit volume. ‘The potential energy density due to the sound wave Dre may be found from the gas law. For very small changes in the volume of the box, we may write [see Eqs. (2.8)] (2.67) (2.9) and substitute the resulting expression for dz, the potential energy density becomes = pap lw Pa 25P, Des (2.08) When the sound pressure p is equal to zero, the potential energy due to the sound wave must be zero. The arbitrary constant of integration is therefore also equal to zero. ‘The total energy density due to the sound wave D = Dax + Des, oF Dw = 3 (on + 5) (2.69) ‘This equation is true at any instant at a given point in space 2.8, Energy Density in Plane Waves. Energy Density in a Plane Pree-progressive Wave. From Eqs. (2.57) and (2.59) we have seen that the pressure and particle velocity in a plane free-progressive (outgoing) Part VI ENERGY D PNSITY AND INTE 41 wave are equal to (zt) = Re V2 pret = /2 Ips cos [k(ct — 2) + 6] (2.70) V2 Me psc = V2 Pol Ee Feel cos h(ct ~ 2) + 6] (2.71) u(z,t) = Re where py = |pyle. ‘The instantaneous energy density for such a wave in the steady state is, from Lage. (2.68) and (2.19), equal to D(z) = |psl* DG pan) cost Ia(ct — 2) + 8] = Dpal? 1+ cos 2at ~ 2/e + Of) mc 2 WAL (1 + cos 2u(t — 2/0 + 6/4) @72) “This equation says that for a plane fre-progressive wave, at all times, the Kinetic and potential energy densities are equal at a given point in space but that they vary with position or with time sinusoidally from zero to twice their average value. ‘The situation here is different from that fora pendulum where the kinetic energy and the potential energy vary in opposite phase, &.c, one ie a maximum when the other iz a minimum Here, energy is being transported away from the source. Conversely, the pendulum is a conservative system. ‘When averaged over either a length of time equal tol = 1'/2 = 1/2f or a distance in space 2 ~ /2~ o/2f, we find the average energy density to be equal to Ipal? 2 13) al wattce/m 273) D, where |ps| is the magnitude of the rms value (in time) of the sound pressure measured at any point in the sound wave. Note also that pect = 7Ps as stated before. Inspection of Eq. (2.60) shows that we may let Pe pot where 1, is the rms value (in time) of the velocity at any point in the wave, Then, uy (274) Dara = |1sl¥00 (275) ‘Equations (2.73) and (2.75) give the relations among rms sound pressure, particle velocity, and energy density Energy Density in a Plane Standing Wave. From Eqs. (2.42) and (2.47) we have that piety = VE need EM cow — 2) aaa42 THE WAVE EQUATION AND SOLUTIONS — (Chap.2 where @ is the phase angle of us. 608 (ot + 0) sin k= 2 UE) = VE ug) SRE Sin BE 2) em In this case, the kinetic and potential energy are 90° out of time phase ‘The situation is analogous to that for a pendulum becuse in both eases the systems are conservative. ‘The instantaneous energy density for such a wave in the steady state is, from Eqs. (2.69) and (2.19), equal to 08 2(ut + @) cos 2(l — 2) T= cos 2E D(z) = |udlioe 2.78) ‘When averaged over either a length of time f equal to 7/2 or a distance in space equal to 4/2, we find the average energy densily to be equal to Dog = ltl, wateane/mt ex where us| is the magnitude of the rms velocity of the piston at 2 = 0. This equation shows that, for « constant value of jus, the average energy density varies from |uy/?p9/2 to infinity depending on the value of H = 2el/r A better way of representing the average energy density is in terms of the rms pressure. If, by definition, we let the rms value of the pressure be related to the mms velocity uy at x = O by the formula pi = one (280) we have (zt) = V2 [pil sin (wt + 6) cos k(t — 2) (281) Then Eq. (2.79) becomes = iat Dow = PO (282) Here, {pi| is the magnitude of the rms value (in time) of the mazimum value (in space) of the sound pressure. If we measure the rms value of the sound pressure in space by moving a microphone backward and forward over a wavelength and averaging the varying output in a rms rectifier, then [Poul = [psl/-V2 and Dig = Weal’ yattace/mt (283) where |pow| is the magnitude of the rms value of the sound pressure averaged in both space and time, Note that Eq. (2.83) is identical to By. (273) Part V1 ENERGY DENSITY AND INTENSI 3 Example 24. Calculate the average energy density in a plane free-progressive Sinusoidal sound wave with a maximum particle displacement of 0.01 em at a quency of 100 eps. Soluion, From Eq, (2.80) we find that the ras yrtile velovity 1 tga wine 80 2e x 100 x OL > one = 00 X 0.01 9.645 ma /sce VP x10 ‘The average energy density is given by Ka. (275), Dag = (0.0485)* X LB = 234 X 10-9 wattseo 2.9. Energy Density in a Spherical Free-progressive Wave. ‘The energy density in a spherical free-progressive wave can be shown to be equal to 81) where [p, is the magnitude of the rms value (in time) of the sound pressure at a point a distance r from the center of the spherical souree. If the produet of the distance r and the frequency is large (2k? >> 1), the average energy density is the same as for plane free-traveling oF standing wave, as can be seen from Eqs. (2.73) and (2.83). Near the sonree, however, the eneray density hecomes very large. ‘This occurs because the impedance {sce Eq. (2.64)] becomes largely reactive and the stored energy becomes high. 2.10. Sound Energy Flow—Intensity, Later in this text we make frequent reference to the flow of sound energy through an acoustie system, Because of the law of the conservation of energy, the total acoustie energy starting from a source must be completely accounted for in the system. At any part of an acoustic system, we should be able to state the amount of energy flowing through that part per unit time, and it should equal the power emanating from the source minus any intervening losses. In Part IT we defined intensity as the average time rate at which energy is flowing through unit area of the acoustie medium. In the mks system, the units of intensity are watts per square meter. The intensity is actually the product of the sound pressure times the in-phase component of the particle velocity. General Equation for Intensity. We can find the average intensity J 1 given direction at a given point in the medium by performing the operation T= Reptgcos ¢ (285) SL. B Kinder and A. TR. Frey, “Fundamentals of Acoustss,” pp, 167-168, John Wiley & Sons, Ine, 1980. 1 The average power supplied by an clrtrical generator to a circuit equals the voltage times the in-phase component of the current. "Tht power ean be shown to quel Re (E*1), where B and 7 are the complex ema voltage and curtont, respectively.a THE WAVE EQUATION AND SOLUTIONS — [Chap.2 where p* is the complex conjugatef of the rms sound pressure p, qis the complex rms particle velocity in the direction the wave is traveling, and @ is the angle between the direction of travel and the direction in which the intensity is being determined. ‘The symbol Re indicates that the real part of the product is to he taken. Intensily in a Plane Free-progressive Wave. For a plane free-progressive sound wave the intensity equals = Re pot Pt co T= Re pom Be é 2.80) Another way of looking at the question of intensity for & plane progressive wave is to say that all the energy contained in a column of gas equal in Jength to em must pass through unit area in 1 see. Hence, the intensity is T= Dang 005 4 287) So, regardless of whether the intensity is determined from (2.86) or (2.87), wwe get for & plane free-progressive wave that Pal oo g = 1 = BE cose Intensity im a Plane Standing Wave. In a plane standing wave the pressure and particle velocity are 90° out of phase in time [see Fqs. (2.76) and (2.77)] so that the real part of the product ptu is zero. Hence, for a plane standing wave, ius love cos (2.88) T=0 (289) Physically, tbis means that as much sound energy returns to the source as travels away from it. Intensity in a Spherical Free-progressive Wave. For a spherical progres cive wave, we got the pressure p from Eq. (2.62). By definition, let pe a em (2.00) Then, Plot) = V2 pact eo) ‘The quantity p, is equal to the complex rms pressure at any point a distance 7 from the center of the source. Hence, the particle velocity u(r) at any point a distance r is Em (14 jbo asm HIE pis represented by [ple then pe is [ple Similarly, if p i represente hy e+ jy, then 9" i represented by px ~ Je ur. Ps Vi ENERGY DENSITY AND INTENSITY 45 or the complex rms particle vel we B15] Substitution of the sound pressure at, p,, and Eq, (2.98) into Bq, (2.85) vields ) a I= Rept iF Be cos & eo) ‘owe here, as before, ¢ is the angle between the direction of travel of the wave and the direction in which the intensity is being determined. ‘Wo can derive these results in a different way. Equation (2.08) states that, for kr large, p and u fora spherical wave are nearly in time phase and P(r) = pucu(r) a8 shown by Eq. (2.05). Hence, for kr large, we see fron. Eq, (2.88) that in a spherical wave for large distances I = |ur|?pec 008 6. ‘The total power at any radius r is equal to W = er! = 4r"|pl?/pee. Hence, for a spherical wave, Ww T=, forg=0 (2.95) By the law of conservation of energy, W is independent of r if there are uio losses in the gas so that the intensity varies inversely as the square of the distance r. From Eq. (2.90) we see also that: the square of the rms magnitude of the sound pressure at any point varies inversely with the square of the distance r. Hence, because the intensity / at any point varies similarly, it is directly proportional to the square of the sound pressure at that. point. ‘This result agrees with that shown in Eq. (2.94) Example 2.5. A spherical sound source is radiating sinusoidally into free space | watt of acoustic power at 1000 eps. Caleulate (a) intensity in the direction the wave js traveling; (6) sound pressure; (e) particle velocity; (d) phase angle between (0) and (¢); (2) energy density; and (f) sound preseure level at @ point 30 em from the ‘eonter of the source, (Assume 22°F and 0.751 m Hg.) Solution. a. ‘The intensity maybe found fo 2.95). >, The rins sound pressure eomes from Ig. (2:94). Ind = Vase = VORB XAT = 18.97 nowtone/m* ‘ The rms particle volocity is given by Bg. (2.98), be = (Or x 1000/344.8)(08) = 5.46 be VTE A 1897 VT + HR tel = Fe VUES 1897 VT BB aaz4 mine46 THE WAVE EQUATION AND SOLUTIONS — (Chap. 4. The phase angle @ between pe and ue may be found from Faq. (264) 0 = 90" — tant kr = 90" — 796" = 10.4" The energy density is given by Hq. (2.84) BEC + ah) = pata ( Pow pact AI * aunt) ~ TEx TO + +2162 X 10-4 wattsee/n* J. The sound pressure love is found from Hq. (1.18). oe 18.97 SPL = 20 log 5 OTT = 119.5 db re 2 X 10-4 nowton/m (re 2 X 10-4 mierobar) ‘This sound pressure level is about 15 db higher than the highest level tht is meas tured at 25 1t above a full aymphiony orchestra. In other words, 1 watl of acoustic power ervatcs a very high sound pressure level atl ft from the souree CHAPTER 3 ELECTRO-MECHANO-ACOUSTICAL CIRCUITS parr VI Mechanical Circuits SS 3.1, Introduction. ‘The subject of electro-mechano-acoustics (some. times called dynamical analogies) is the application of electrical-cireuit theory to the solution of mechanical and acoustical problems. In classical mechanics, vibrational phenomena are represented entirely by differential equations. ‘This situation existed also early in the history of telephony and sadiv. As telephone and radio communication developed, it became obvious that a schematic representation of the elements and their interconnections was valuable. These schematic diagrams made it Possible for engineers to visualize the performance of a circuit without laboriously solving its equations. The performance of radio and television systems can be studied from a single sheet of paper when sich schematic diagrams are used. Such a study would have been hopelessly 4ifficult if only the equations of the system were available. There is another important advantage of a schematic diagram besides its usefulness in visualizing the system. Often one has a piece of equipment for which he desires the differential equations. ‘The schematic iagram may then be drawn from visual inspection of the equipment. Following this, the differential equations may be formed directly from the schematic diagrams. Most engineers are trained to follow this procedure rather than to attempt to formulate the differential equations directly. Schematic diagrams have their simplest applications in circuits that contain Jumped elements, é¢., where the only independent variable is time. In distributed systems, which are common in acoustics, there may be as many as three space variables and a time variable. Here, a schematic diagram becomes more complicated to visualize than the differential equations, and the classical theory comes into its own again. There are many problems in acousties, however, in which the elements are Jumped and the schematic diagram may be used to good advantage. a48 ELECTRO-MECHANO-ACOUSTICAL CIRCUITS (Chap. 3 Four principal requirements are fulfilled by the methods used in this text to establish schematic representations for acoustic and mechanical devices. They are’ 1. The methods must permit the formation of schematic diagrams from visual inspection of devices 2. They must be capable of such manipulation as will make possible the combination of electrical, mechanical, and acoustical elements into one schematic diagram. 3. They must preserve the identity of each element in combined circuits 60 that one can recognize immediately a force, voltage, mass, inductance, and so on. 4. They must use the familiar symbols and the rules of manipulation for electrical circuits Several methods that have been devised fulfill one or two of the above four requirements, but not all four. A purpose of this chapter is to present a new method for handling combined electrical, mechanical, and acoustic systems. It incorporates the good features of previous theories ‘and also fulfils the above four requirements. ‘The symbols used conform with those of earlier texts wherever possible.!“* 3.2. Physical and Mathematical Meanings of Circuit Blements. ‘The circuit elements we shall use in forming a schematie diagram are those of clectrical-circuit theory. These elements and their mathematical meaning are tabulated in Table 3.1 and should be learned at this time. There are generators of two types. ‘There are four types of circuit elements: resistance, capacitance, inductance, and transformation. There are three generic quantities: (a) the drop across the circuit element; (0) the flow through the circuit element; and (¢) the magnitude of the circuit element. Attention should be paid to the fact that the quantity ais not restricted to voltage ¢, nor b to electrical current i. In some problems a will represent force f, or velocity wu, or pressure p, or volume velocity U. In those cases b will represent, respectively, velocity u, or force f, or volume °B, Gehlshoj, “lectromechanical and Hectroscoustieal Anslogies,” Academy of ‘Technical Sciences, Copenhagen, 1947 *F. A. Firestone, A Now Anelogy between Mechanieal and Hlectrcal Systems, J. Acoust. Soe. Amer, 4: 249-267 (1988); The Mobility Method of Computing the Vibration’ of Linear’ Mechanical and Acoustical Syatems: Mechanical-clectrical Analogies, J. Appl. Phyr., 8: 373-387 (1938). *H. F, Olson, “Dynamical Analogies” D, Van Nostrand Company, Inc, New York, 1988. “WP. Mason, Electrical and Mechanical Anslogies, Bell Syslem Tech, J., 20: 405-414 (1941). A. Bloch, Electro-mechanical Anslogies and Their Use for the Analysis of Mechani- cal and Bleetrosmechanieal Systems, J. Ina. Elec. Eng., 92: 157-169 (1945), {Among the four circuit elements, the first three are two-poles, This list is exhaus- tive, The transformation element isa four-pole, There are other loses fonr-poes Which one might have chosen in addition, e., the ideal gyrator, Part. VI) MECHANICAL CIRCUITS 49 velocity U, or pressure p. Similarly, the quantity ¢ might be any appropriate quantity such as mass, compliance, inductance, resistance, etc, The physical meaning of the circuit elements ¢ depends on the way in which the quantities a and b are chosen, with the restriction that ab thas the dimension of power in all eases, ‘The complete array of alterna- tives is shown in Table 3.2. ‘TABLE 3.1. Mathematical and Physical Significance of Symbols ] T ‘Mooring somtt | tame : | Constant-drop | The quantity a is independent of tor. ‘The arrow points to the positive terminal ofthe generator |Constant-ow |The quantity b is independent of | generator what is connected to the generator. ‘The arrow points in the ireation of positive fow Resistancetype | a sement Sent fend f ba fom | ee — ee clement gst $ type element, be a 138 aa ‘An important idea to fix in your mind is that the mathematical opera~ tions associated with a given symbol are invariant. If the element is of the inductance type, for example, the drop @ across it is equal to the time derivative of the flow b through it multiplied by its sizec. Note that this rale is not always followed in electrical-cireuit theory because there con- ductance and resistance are often indiscriminately written beside the symbol for a resistance-type element, The invariant operations to be associated with each symbol are shown in columns 3 and 4 of Table 3.1.| Part VI) MECHANICAL CIRCUITS 51 CHANO-ACOU 3.3, Mechanical Circuits. Mechanical-circuit elements need not “kK always be represented by electrical symbols. Since one frequently draws " a mechanical circuit directly from inspection of the mechanical device, bis. ‘more obvious forms of mechanical elements are sometimes useful, at least until the student is thoroughly familiar with the analogous cireuit. We shall accordingly devise a set of ‘‘mechanieal” elements to be used as an introduction to the elements of Table 3.1 em Ma leney TABLE 3.3. Conversion from Mobility-type Analogy to Impedance-type Analogy, or Vice Versa Acoustical | ats = [MECHANICAL NULOGES {AOOUSTIAL MALDGIES Impedance analogy P u ovis pe [_trowtnce ype | Meity ype | tnpetonn Spe ca ee (a i Ec a ry ToS we PS BS att! on Ra Je= Me ‘and Acoustical Circuits Impedance ‘analogy f T ‘Mechanical Mesh aus (nobiy Boe) ech to was empress) {us 2 poe In electrical circuits, a voltage measurement is made by attaching the Jeads from a voltmeter across the two terminals of the element. Voltage is a quantity that we can measure without breaking into the circuit, ‘To measure electric current, however, we must break into the circuit because this quantity acts Ghrough the clement. In mechanical devices, on the other hand, we ean measure the velocity (or the displacement) without disturbing the machine by using « capacitive or inertially operated vibra~ tion pickup to determine the quantity at any point on the machine. It is not velocity but force that is analogous to electric current. Force can- i not be measured unless one breaks into the device. Values for a, & and ¢ in Electrical, Mechani ese TABLE 3.2 : ans yas oie ° ’ 1 PreferrRO-M CHANO-ACOUSTICAL CIRCUITS (Chap. It becomes apparent then that if a mechanical element is stetly analogous to an electrical element it must have a velocity difference appeating between (or acrou) is wo terminals ands free acing Cough it Analogounly als, the produet of ie rma foree/ in newlane ard the inphase eompovent of the nn velocity w in meter por second Je the oF ———r—C— rrr torreeponds to a voltage and a force to a ourant, the mobilgnine tmalogy. tis also known as the “averse” asalogy ‘Many text teachin addition a "dieet” analogsy Tt the opposite of the mobility analogy in that force is made to correspond to voltage and rh ———F_———— Impedance type analogy To fain the student with bath consepte ‘all examples will be given here both in mobility-type and impedance-type analogies. Mechanical Impedance Zy, and Mechanical Mobily tw. ‘The mechan- Oe —r———i— tt mechanical device, "We commonly use the symbol Ze or mechani impedance, where the subscript M stands for “mechanical.” "The unit tre nenton-acconda per meter, or mka mechaneal hina "The mechanical mcilty isthe inverse of the mechanical impedance, 1Linthe complex ata volrtytatce nea gees eaat a ooh devieesWeommonly ure the symbol fs for mechanical mating Te Unite are meters por seco per newton, or mia rascal 14. cal mohms.+ —— Je selon a omelet opto that force. ‘The units the kilogram, AUN ght ase LL ———_N convection ip needed to et it in motion” Howevty, the Je. Me force acting on t mase and tho renllant acederalon {ArH reckoned with reapet to ube earth neta rae) co thet Inreality the second terminal of masse the earth he mechanival symbol used to represen mast is shower i Fig. 31 ‘The upper end of the mase moves with a velocity wth respect fo te ground. ‘The }shaped configuration represents, the "aesnd” terminal Of the mass end has aero velocity. "The force cam be measured by a tuilable device inorted between the point I and the next elomen cr reverator connecting toi Maan Mw obeys Newton second aw that au) Sy = Mu @y {The word “mohm” stands for mobility ohm. The vty are meters pee second Pact VE) MECHANICAL CIRCUITS 53 ‘whore /() is the instantaneous force in newtons, Mw is the mass in kilograms, and 1(0) is the instantaneous velocity in meters per second, In the steady state [see Eqs. (2.38) to (2.35)], with an angular frequency «sequal Lo 2r times the frequeney of vibration, we have the special ease of Newton's second law, joM we 82) where j = \/=i as usual and f and w are rms complex quantities, The mobility-type analogous symbol that, ; we use as a replacement for the mechanical I symbol in our circuits is a capacitance type. [>| re Wisshown in Fig. 320, Themathematical fu eae Py Say 4 ‘operation invariant for this symbol is found from Table 3.1. In the steady state we have am or = L658) ty toe tpt jue oN (a) 0) ‘This equation is seen to satisfy the physical Fra. 3.2. (a) Moblity-ype law given in Bq, (3.2). Note the similarity $4 (impedance type ym- in appearance of the mechanical and analo- coussvmbolsin Figs. 8.1 and 3.2a, Inelectrical eireuits the time integral of the current through a capacitor is charge, The analogous quantity here is the time integral of force, which is momentum, ‘The impedance-type analogous symbol for a mass is an inductance. It is shown in Fig. 3.25, ‘The invariant operation for steady state is 4a = jucb orf = juM yn. Ialso satisfies Bq. (82). Note, however, that in this analogy one side of the mass element is not neces- 1) Ju, sarily grounded; this often leads to confusion. In electrical circuits the time integral of the voltage across an inductance is fluxctums. The analogous quantity here is momentum, Mechanical Compliance Cy. A physics} structure is 2ll Yur said to be a mechanical compliance Cw if, when itis acted ie. 83. Me. on by a force, its displaced in direct proportion to the chasiealsymbol force. The unit is the meter per newton. Compliant for a mishan~ elements usually have two apparent terminals ‘The mechanical symbol used to represent a mechanical compliance isa spring. It is shown in Fig. 3.3. ‘The upper end of the element moves with a velocity u: and the lower end with a velocity us “The force required to produce the difference hetween the velocities u; and tu: may be measured by breaking into the machine at either point } or point 2. Just as the same vrrent would be measured at either end of an element in an electrical cirewit, 20 the same force will be found here at either end of the compliant element. Cw54. RO-MECHANO-ACOUSS AL CIRCUITS [Chap.3 Mechanical compliance Cy, obeys the following physical ia, ant foa or 0 Z fue ea) where Cy is the mechanical compliance in meters per newton and u(t) is the instantaneous velocity in meters per second equal to us — u, the difference in velocity of the two ends, In the steady state, with an angular frequency w equal to 2x times the frequency of vibration, we have, i jae 5) where f and u are taken to be rms complex quantities. ‘The mobility-type analogous symbol used as 2 replacement for the mechanical symbol in our circuits is an inductance. Tt is shown in Fig, 3.4a. ‘The invariant mathematical operation f © that this symbol represents is given in Table 3.1. In the steady state we have w= jeoluf G6) _ ae gee Ls Ly In electrical circuits the time integral of the voltage acrose an inductance is flus-tusns esaty-toe —Ingetne-tye ‘The analogous quantity here is the time in- () (0) ‘tegral of velocity, which is displacement, Fro. 34. (2) Mobility-type ‘This equation satisfies the physical law ni) npotaeetype given in Eq, (QS). Note the wimaiy in fei fr athe Sompir cance the mechanical and analogous symbols in Pig 8.8 and 3. ‘The impedance-type analogous symbol for a mechanical compliance is a capacitance. It is shown in Fig. 3.45. ‘The invariant operation for steady state isa = b/jue, or f = u/jwCy. It also satisfies Bq. (3.5). In electrical circuits the time integral of the current through capacitor is the charge. The analogous quantity here is the displacement, Mechanical Resistance Ry, and Mechanical Responsiveness tw. A physical structure is stid to be a mechanical resistance Ry if, when it is acted on by a force, it moves with a velocity direetly proportional to the force. ‘The unit is the mks mechanical ohm. ‘We also define here a quantity ry, the mechanical responsiveness, that, is the reciprocal of tw. ‘The unit of responsiveness is the mks mechanical mohm, ‘The above representation for mechanical resistance is usually limited to viscous resistance. Frictional resistance is excluded because, for it, the ratio of force to velocity is not a constant. Both terminals of resistive ements can usually be located by visual inspection, Pact Vi) MECHANICAL CIRCUTTS 55 ‘The mechanical element used to represent viscous resistance is the fluid dashpot shown schematically in Fig. 3.5, ‘The upper end of the element. moves with a velocity w and the lower with a velocity us. ‘The force required to produce the difference between the two velocities u; and ts may be measured by breaking into the machine at either point 1 oF point 2 Mechanical resistance Ry obeys the following physical law, foram de where f is the force in newtons, 1 is the difference between the velocities tu, and uz of the two ends, Hy is the mechanical resistance in mechanical ohms, ie, newtons/(meter per second), and ry is the mechanical responsiveness in mks mechanical mohms, i.e., meters per second per newton. ‘The mobility-type analogous symbol used to replace the mechanical symbol in our circuits is a resistance, It is shown in Fig. 3.6a, ‘The ody tye gta al be (a) 0) al be Fa. 35, Mechanical Fic.3.6. (@) Mobility-typeand Fra. 3.7. Mechenical symbol for mechanical @) impedaneetype symbols for eymboi for a constant (iseous) resistance. mechanical resistance. velocity generator, invariant mathematical operation that this symbol represents is given in Table3.1. Incither the steady or transient state we have we raf G8) In the steady state w and f are taken to be rms complex quantities. ‘This equation satisfies the physical law given in Ea. (3.7) ‘The impedance-type analogous symbol for a mechanical resistance is shown in Fig, 3.00. It also satisfies Bq. (3.7). ‘Mechanical Generators. ‘The mechanical generators considered will be cone of two types, eonstant-velocity or constant-force. A constant. relocity generator is represented as a very strong motor attached to a shuttle mechanism in the manner shown in Fig. 8.7. ‘The opposite ends of the generator have velocities uy and us. One of these velocities either 1 oF ua, is determined by factors external to the generator. The differ-56 ELECTRO-MECHANO-ACOUS L CIRCUITS (Chap. 3 ence between the velocities w; and wy, however, is velocity w that is inde pendent of the external load connected to the onerator. ‘The symbols that we used in the two analogies to replace the mechanical symbol for a constant-velocity generator are shown in Fig. 38. The invariant mathematical operations that these symbols represent are also given in Table 3.1. ‘The tips of the arrows point to the “positive” terminals of the generators. ‘The double circles in Fig. 3.8a indicate that the internal mobility of the generator is zero. ‘The dashed line in Fig. 3.80 indicates that the internal impedance of the generator is infinite. ‘A conslant-force generator is represented here by an electromagnetic transducer (e.g, a moving-eoil loudspeaker) in the primary of which an electric current of constant amplitude is maintained. Such a gencrator produces a foree equal to the product of the current é, the flux density B, and the effective length of the wire I cutting the flux (f = Bli). This device is shown schematically in Fig. 3.8. ‘The opposite ends of the aaa, 4 fu == ee (a) (b) (a) (8) generator have velocities u and ws that are determined by factors external to the generator. ‘The force that the generator produces and that may be measured by breaking into the device at either point 1 or point 2 is a constant force, independent, of what is connected to the generator. ‘The symbols used in the two analogies to replace the mechanical symbol for a constant-force generator are given in Fig. 3.10. ‘The invariant mathematical operations that these symbols represent are also given in Table 3.1. ‘The arrows point in the direction of positive flow. Here, the dashed line indicates infinite mobility, and the double circles indicate zero impedance. Levers. simpie vever. It is apparent that the lever is a device closely analogous to a transformer. ‘The lever in its simplest form consists of a weightless bar resting on an immovable fulerum, so arranged that a downward force on one end causes an upward force on the other id (see Fig. 8.11). From elementary physics we may write the equation ‘of balance of moments around the fulerum, Suds = fils Past VI} MECHANICAL CIRCUITS a7 of, if not balanced, assuming, small displacements, 9) Also, .10) ‘The above equations may be represented ¥ the ideal transformers of ws nl ), rh ipa pe FLOATING LevER. As an example of a simple floating lever, consider a weightless bar resting on a fulcrum that yields under force. ‘The bar is, seta HE) OG} Aa Mosity ype Impedance. oe 2 (@) () Fie, 3.11. Simple lever. Fro, 3.12, (a) Mobility-typo and (®) imped- fance-type symbols for a simple lever, so arranged that a downward force on one end tends to produce an upward force on the other end, An example is shown in Fig. 3.13. ‘To solve this type of problem, we first write the equations of moments. ' 4 et 4 ‘Summing the moments about the center support gives Lf hfe at Mobis consti 7, move vp a8 Sonn ery hfe (B11) Fic. 3.18. Floating lever. the end support gives (i + bf When the forces are not balanced, and if we assume infinitesimal displacements, the velocities are related to the forces through the mobilities, so that (3.12)ICAL CIRCUITS (Chap. 3 mau QE fora =0 so that (3.13) and, finally, 14) This equation may be represented by the analogous circuit of Fig. 3.14. The lever loads the generator with two mobilities connected in series, each of which behaves as a simple lever when the other is equal to zero. It will be seen that this is a way of obtaining the equivalent of two series masses without a common zero-velocity (ground) point. This will be illustrated in Example 3.3, (uit) et 13. 3.14. Mobility-type symbol for a Fic, 3.16, Sixelement mechanical de- floating, lever. view. Example 8.1. The mechanical dovieo of Fig. 3.15 consists of piston of mass Mwy iding on an oil surface inside a eylinder of mass Afw;.. This eylinder in urn slides in an oiled groove cut ina rigid body... ‘The sliding. (viscous) resistances art Iewy and Re respectively. ‘The cylinder is held by a spring of compliance Cw. ‘The mechan cal generator maintains a constant sinusoidal velocity of angular frequency wy whose rms magnitude is xim/see. Solve for the force J produced by the generator. Solution. Although the force will be determined ultimately from an analysis of the mobility-type analogous circuit for this mechanical deviee, itis frequently useful to raw & mechanical-circuit diagram. This interim step to the desiel circuit will be especially helpful to the student who is inexperienced in the use of analogies. Its use virtually eliminates errors from the final eieuit ‘To draw the mechanical circuit, note fst the junction points of two oF more ele- nents, This locates all element terminals which move with the same velocity. "Phere ue in this example two velocities, w and us, in addition to “ground,” or zero velocity These two velocities are represented in the mechanical-cireuit diagram by the veloetica of two imaginary rigid bars, land 2 of Fig, 3.16, which oscillate in a verlical direction The cirevit drawing is made by attaching al clement terminals with velocity 1 to tho iret bar and all terminals with velocity wz tn the second hat. All terminals with zero Past Vi) MECHANICAL CIRCUITS 59 velocity are drawn to a ground bar. Note that & mass always has one termiaal on ground.t "Three elements of Fig. 3.15 have one terminal with the velocity a the generator, the mass Mus, and the viscous resistance wi, These are attached to bar 1. Four elements have one terminal with the velocity us the viscous resstaacet ‘Ru: and Ra, the mass Mus, and the compliance Cy, These ate attached to bar 2 Five elements have one terminal with zero velocity: the generator, both masecs, the ‘viscous resistance Ry, and the compliance Cx. Fic. 3.16. Mechanioal cireut for the device of Fig. 3.15, We are now in a position to transform the mechanical circuit into « mobility-type analogous circuit, This i accomplished simply by replacing the mechanical elements with the analogous mobilty-type elements. The circuit becomes that shown in Fig. 3.17, Remember that, in the moblity-type analogy, force “flows” through the elements and velocity is the drop across them, The resistors must have lower case v's written alongside them. As defined shove, rw = 1/Ry, aad the unit is the mks ‘mechanical mob, ‘The equations for this circuit are found in the usual manner, using the rules of Table 3.1. Let us determine zy = u/f, the mechanical mobility presented to the 8.17. Mobility-type analogous eireuit for the device of Fig. 8.15. enerator. ‘The mechanical mobility of the three elements in parallel on the right= Dband side of the echematie diagram is a ho Theta * tan * Jat om r Jour + Bas + se Including the element ry: the mechanical mobility for that part ofthe circuit through which Je flows is, then, oeetaeetesssssenae = ao™ r SoMa Rare + ‘Note that the input mechanical mobility ew is given by “Tih { An exception to this rule may oeeur when the mechanical device embodies one or more floating levers, a8 we just learned.60 BLEGTRO-MECHANO-ACOUSTICAL CIRCUITS (Chap-3 and f= apcity, 7 Moe Substituting /. al fs into the second equation preceding gives us the input mobility tse) 2.355) mt —— 1 Jes + Neus + 5h he res is fm Zyu — newtons 1) Example $2. Ava further © masses of Zand 4 kg shown in Fi maple of « mechanical eireuit, let us vonsider the two 3.18, They are assumed to rest on a frietionless uo mp +R OORT ese ‘ist iebonese ‘race Fa, 3.18. Three-dement mechanical deviee. plane surface and to be conncetod together through a generator of constant velocity that it also free to slide on the Fretiouloas plane surface, Let ite velocity be uy = 2.08 1000 ean one Draw the mohility-type analogous eieuit, and determine the force f produced by the generator. Also, determine the mobility presente to the generator, Solution. ‘The masses do not have the sume velovity with respect to ground. The difference hetweon the velocities of the two mass is ux. The element represeating mass is that shiown in Fig. 3:20 with one end grounded and the other moving at the Velocity of the mas, — the moblity-type vieeait for this example is shown in Fig. 2.19. The velority up equals ay + is, where wy i t Mont the velocity with respect to ground of My), and uz is that us Theta) + Weed Some Fic. 3.19, Mobility-ty = dellandvs nonin erent for the Maan device of Figs 3.18, ai 4X10? — go nomtone tain a+ ve > ny The j indieates that the time phase of the force is 0° leading with respect to that uf the Velocity of the generator. Part Vi) MECHANICAL CIRCUITS 6 CObviowsly, when one mass is large compared with the other, the foree is that neces sary to mov the sinaler one alone. ‘This example revels the only type of ease whieh masses can be in verias without the introduction of floating levers. At most, ‘only two masses ean Ie in series boeansr a common grou is necessary ‘The mobility presented to the generator ie ee (0), = Ma tan “ FF ane Gol wiM ves = ja seg = —FES X 1 oe a8) Example 3, An example of « mechanical devi embodying a floating lever ie shown in Fig. 3:20.” The maine atached pointe 2 and 3 may be sansa to be GAs fe: o Mee * te : (2) Mie @) (oy Fic, 8.20, (a) Mechanical device em- Fis. 8.21, (a) Mobility-type analogous cr- hdying a floating lever. (@) Mechan- cuit for the deviee of Fig 8.20. (0) Same ‘eal diagram of (a). The eompliances as (a) but with transformers removed, ff the springs are very large so that all Sf feand fz go to move M ys and Ms resting on very cormpliant springs, The driving force fi will be assurned to have a Srequeney well above the resonance frequencies of the masses and their spring supports so that Algo, assure that anne is attached 0 the weightless lever bar at point 1, with #62 ELECTRO-MECHANO-Acoust AL CIRCUITS (Chap. 3 mobility Solve for the total mobility presentnd to the constant force generator fi Solution, By inspection, the mobility-type analogous eireit is drawn as shown in Fig. §.21@ and 0, Solving for ew = ui/fy we get i apes @.19) Fat Fa + Me] Note that if f-» 0, the mobility is simply that of the mass Man. Also, if +0, the mobility is that of Muy and Mya, that is, hie F Mand 8.20) t is possible with one or more floating levers to have one of more masses with no sound terminal pant VII Acoustical Circuils 3.4. Acoustical Elements. Acoustical circuits are frequently more lifficult to draw than mechanical ones because the elements are less easy o identify. As was the case for mechanical cireuits, the more obvious orms of the elements will be useful as an intermediate step toward draw- ng the analogous circuit diagram. When the student is more familiar vith acoustical circuits, he will be able to pass directly from the acoustie levice to the final form of the equivalent circuit. In acoustic devices, the quantity we are able to measure most easily vithout modification of the device is sound pressure. Such a measure- nent is made by inserting a small hollow probe tube into the sound field the desired point. This probe tube leads to one side of a microphone iaphragm. ‘The other side of the diaphragm is exposed to atmospheric ressure. A movement of the diaphragm takes place when there is a ifference in pressure across it. ‘This difference between atmospheric ressure and the pressure with the sound field is the sound pressure p. Because we can measure sound pressure by such a probe-tube arrange- nent without disturbing the device, it seems that sound pressure is nalogous to voltage in electrical circuits. Such a choice requires us to onsider current as being analogous to some quantity which is propor- ional to velocity. As we shall show shortly, good choice is to make urrent analogous to volume velocity, the volume of gas displaced pe econd Part VIE ACOUSTICAL CIRCUITS 63 A strong argument can be made for this choice of analogy when one considers the relations governing the flow of air inside such acoustio devices as loudspeakers, microphones, and noise filters. Inside a certain type of microphone, for example, there is an ar eavity that connects to the outside air through a small tube (eee Fig. 3.22). Assume, now, that the outer end of this tube is placed in a sound wave. The wave will cause a movement of the air particles in the tube. Obviously, there i a junction between the tube and the cavity at the inner end of the tube at point A. Let us atk ourselves the question, What physical quantities are Continuous at this junction point? int the sound pressure just inside the tube at 4 is the same as that in the cavity just outside A. That is to say, we have continuity of sound Pressure. Second, the quantity of air leaving the inner end of the small tube in a given interval of tm isthe quantity that enters the cavity thestme interval of time. Thatis, the mass per second of gas leaving the small tube equala the iass per second of gas entering the volume, Because the pressure is the same at both ee places, the density of the gas must also be are, the same, and it follows that there is eon. Einuity of volume velocity (cubie meters per secuud) at this Junction. Analogously, in ete the case of eesti, there sconinaty of ad electric current at a junction, Continuity F ed cavity con- of volume velocity must exist even if there 3eag, W tbe outer are several tubes or eavities joining nearone tonal area The jncton point. A violation of the law of conserva- Plane between the tube and the tion of mass otherwise would occur. ee We conclude that the quantity that flows through our acoustical elements must be the volume velocity U in cubie meters per second and the drop across our coustieal elements must be the pressure pin newton per square meter. This conclusion indicates that the impedance type of analogy is the preferred analogy for acoustical circuits. The product of the effective sound pressure p times the in-phace component of the effec: tive volume velocity U gives the acoustie power in watts. Tn this part, we shall discuss the more general aspects of acoustical circuits. In Chap. 5 of this book, we explain fully the approximations involved and the rules for using the concepts enunciated here in practical problems. Acoustic Mass M4, Acoustic mass is a quantity proportional to mass but having the dimensions of kilograms per meter‘. It is associated with 44 mass of air accelerated by a net force which acts to displace the gas without appreciably compressing it. ‘The concept of acceleration without compression is an important one to remember. "It will assist you in distinguishing acoustic masses from other elements64 ELECTRO-MECHANO-ACOUSTICAL CIRCUITS (Chap.3 ‘The acoustical element that is used to represent an ucoustic mass is a tube filled with the gas as shown in Pig. 3.23 ‘The physical law governing the motion of a mass that is acted on by a force is Newton's second law, (Q) = Mw du(t)/dt, ‘This law may be txpromed in acoustical term as follows, Med pw = Bree alo '3 0 = Ma 21) here p() = instantancous diference between presures in netons per square meter existing at each end of a mass of gas of Mu kg undergoing acceleration Mu/S? = acoustic mass in kilograms per meter‘ of the gas undergoing acceleration. ‘This quantity is nearly equal to the mass of the gas inside the containing tube divided by the square of the cross-sectional area, To be more exact we must note that the gas in the immediate vicinity of the snds of the tube also adds to the mass, Hence, there are ‘cud curvecliuns” which must be considered. ‘These cor rections are discussed in Chap. 5 (pages 132 to 139) U(® = instantaneous volume velocity, of the gas in cubic meters per second across any cross-sectional plane in the tube. ‘The volume velocity Ud) is equal to the linear velocity u(®) multiplied by the cross-sectional area S, In the steady state, with an angular frequency w, we have M. p= joMU (8.22) where p and U are taken to be rms complex quantities. aye at |e ‘rai vet Impedance type Moby type (a) ) Fic. 3.24. (a) Impedance-type and (2) Pra. 8.25, Bnclosed volume of air V with mobility-type symbole for an acoustic opening for entrance of pressure varia the impedance-type analogous symbol for acoustic mass is shown in Fig, 3.24a, and the mobility-type is given in Fig. 3.24). In the steady state, for either, we get Eq. (3.22). ‘The arrows point in the direction of positive flow or positive drop, Poot VU ACOUSTICAL GIREUETS 6 Acoustic Compliance Cy. Acoustic compliance is a constant quantity having the dimensions of meter* per newton, It is associated with a volume of air that is compressed by a net force without an appreciable average displacement of the center of gravity of air in the volume. In other words, compression without acceleration identifies an acoustic complianee. ‘The acoustical clement that is used to represent an acoustic eompliance is a volume of air drawn as shown in Fig. 3.25. ‘The physical law governing the compression of a volume of air being acted on by a net force was given as f(0) = (1/Cu)fu(t) dt. Converting from mechanical to acoustical terms, Oye SOAS fooge wae or me) (3.23) where p(@) = instantaneous pressure in newtons per square meter acting. to compress the volume V of the air, CyS* = acoustic compliance in meters* per newton of the volume of the air undergoing compression. ‘The acoustic compliance is nearly equal to the volume of air divided by ‘yPo, as we shall see in Chap. 5 (pages 128 to 131). U(® = instantaneous volume velocity in cubie meters per second of the air flowing into the volume that is undergoing compression. ‘The volume velocity U(C) is equal to the linear velocity u(0) multiplied by the cross-sectional area S. In the steady state with an angular frequency a, we have u | * iat 3.28) C, uv > a te ta fe Be ee ee ae quantities. Umpene-tpe wipe ‘The impedance-type analogous element for” (gy () P66 ‘TRO-MECHANO-ACOUSTICAL CIRCUITS Chap. 3 ‘The acoustic clement used to represent an acoustic resistance is a fine-mesh sezeen drawn as shown in Fig. 8.27 ‘The reciprocal of acoustic resistance is the acouse tic responsiveness rs. The unit is the mks acoustic mohm with dimensions meter per second per newton. 1.327, Finemesh _ The physical law governing dissipative effects ina Sc wblover mechanic! system vas given by 0) = Raw, oF ia Ssanacousticalsym- terms of acoustical quantities, bol for eceetiolse - teal quantities, Pe) = RU = LUD (8.25) where p(®) = difference between instantaneous pressures in newtons per square meter across the dissipative element. In the steady state p is an rms complex quantity. Ra = Ru/S* = acoustic resistance in acoustic ohms, ie, newton- seconds per meter 14 = rwS? = acoustic responsiveness in acoustic mohm: meter* per newton-seconds U@ = instantaneous volume velocity in eubic meters per second of the gas through the cross-sectional aren of resistance. In the steady state U is an rms quantity. ‘The impedance-type analogous symbol for acoustic resistance is shown in Fig. 3.280 and the mobility-type in Fig. 3.280, Acoustic Generators, Acoustic generators can be of either the constant- volume velocity or the constant-pressure type. The prime movers in our uv » m7 trite th roe Ly ie of} Lot (@) ir) () 6) Fic. 3.28 (a) Impedancetype symbol Fra, 9.29. a) Impedancetype an fo acoustic estanen aod @S'mabity=moitetgye apes or eos type symbol for acoustic reponsivenesa. pressure generator, acoustical circuits will be exactly like those shown in Figs. 3.7 and 8.9 exeept that us often will be zero and 1, will be the velocity of a small piston of area S. Remembering that w= a — us, we see that the generator of ig. 8.7 has a constant-volume velocity U'= uS and that of Fig. 3.9 « constant pressure of p = f/S. ‘The two types of analogous symbols for acoustic generators are given in Figs, 3.29 and 3.°% ‘The arrows point in the direction of the positive Past VIS ACOUSTICAL CIRCUITS 7 terminal or the positive flow. As before, the double eircles indicate zero impedance or mobility and a dashed line infinite impedance or mobility. Mechanical Rotational Systems. Mechanical rotational systems are handled in the same manner as mechanical rectilineal systems. ‘The following quantities are analogous in the two systems, Rectilineal systems Rotational systems J = force, newtons 1 = torque, newton-m 1 = velocity, m/see 9 = angular velocity, radians ace = displacement, m 4 = angular displacement, radians Zu = fie = mechanical impedance, mks Zp = 7/2 = rotational impedance, mks rmechanieal ohms rotational obme su =e/f= mechanical mobility, mks zg = 6/7 = rotational mobility, mks mechanieal mohme rotational mobs Ry = mechanical resistance, mks me- Re = rotational resistance, mks rota- chanical ohms tional ohms ry = mechanical responsiveness, mks rx = rotational responsiveness, mks 0- mechanical mohms tational mohms Mog = mass, kg In = moment of inertin, kgm? Cur = mechanieal compliance, m/aewlon Cy = rotstionel compliance, radians/ Wy = mechanical power, watts Wa = rotational power, watts Bxample 84, ‘The acoustic device of Fig. 3.31 consists of three cavities Vs, Vs and Vs, two finesmesh sereens Its» and Ru» four short lengths of tube 7, Ts, Ts, and Ta, and a constant-pressure generator. Because the air in the tubes is aot confined, it Pot of (@) (b) own : Fro, 3.80, (a) Impedance-type and (@) Fra. 831. Acoustic device consisting of imobility-type symbols for a constant-vol- four tubes, thre eavities, and two screens time velocity generator. driven by's coustant-pressure generator, experiences negligible eampression. Because the ai in each ofthe cavities is confined, it experiences little average movement. Let the force of the generator be ‘f= 10 cos 1000 newtons the radius of the tube @ = 0.5 om; the longth ofeach of the four tubes t = 5 em; the ‘volume of each of the three eavities V = 10/em#;and the magnitude ofthe two acoustic resistances Ra = 10 mks acoustic ohms. Neglecting end corrections, solve for the volume velocity Usat the end of the tube 7. ‘Solution. Remembering thst there is continuity of volume velocity and pressure at the junctions, we ean draw the impedance-type analogous eireuit from inspection It is shown in Fig. 3.32. The bottom line of the schematic diagram represents laimospheric pressure, which means that here the variationel pressure p is equal to ero, At each of the junctions of the elements 1 to 4, «different variational pressure fan be observed. ‘The end of the fourth tube (7) opens to the atmosphere, which requires that M aa be connected directly tothe bottom line of Fig. 8.32 Note that the volume velocity of the gas leaving the tube 1, is equal to the sum8 rho-Me HANO-ACOUSTICAL CI CULTS. (Chap a of the volume veloities of the gas entering ¥; and 7 ‘The volume veloity of the tess Teaving 7. isthe same as that flowing through the sereen Mea and is equ to the sum of the volume velocitin of the gas entering Vs and 7. ‘One test of the validity of an analogous circuit is its beliavior for direct current, If one removes the piston and blows into the end ofthe tube 7, (Fig. 3.31), a steady Row of air from Tis observed. Some resistance to this fow will he offered by the two seteens Mas and Las. Similarly in the sehematie diagram of Fig, 3.32, steady pees. sure p will produce a steady tow U through Ma resisted only by Rea: and Ita. to, 3.32, Impeslance-type analogous eireut for the acoustie deviee of Fig. 3.31 Asan aside, let us note that an acoustic compliance ean occur in a circuit without one of the terminals being at ground potential only if it iy produced by an elastir diaphregm. For example, if the resistance Ra, in Fig. 3.31 were replaced by an impervious but clastic diaphragm, the element Ra; in Fig. 3.32 would be replaced by a compliancestype element with both terminals shove ground potential. In this ‘ease a steady flow of air could not he maintained through the device of Fig. 3.31, a= «an alao be seen from the eirewit of ig. 3.82, with Ray replaced by a compliance Determine the element sizes of Fig. 3.32 J 2 10-*cos 1000 pag = perm ae reset Mas = Max = Mas = Mac = 22 = 138X998 «750 kame = 7.15 X 10 m*/newton v1 Ca = Cn = Ca = =e Ray = Ray = 10 mks ncoustic ohms ie-cineuit theory, we solve for Uy indircetly, First, arbi- tarily let Uy = 1 mut/se, an the ratio p/Us. Pe = juM alls = 7.5 X 10 newtons/m* Uy = feCaps = “5.36 X 10° m*/a0e WUT Us = 006 i = (Ras + jalfac + py = 9.46 + 514.6 X 108 Us = jaune = “0.1088 + 530.77 10° Uy Us m 01842 + 50.77 x 10-7 Ba = (ay + joM as) + Bs = 17.37 + 32.001 x 108 Oy = jeCasps = 0.1496 + 51.242 X 10° U 2 Us $0, = 0.602 4 51.919 x 10" jeMaW + ps = 1898 +7261 X10 = F fore = 1 ‘The desired value of Uo is 0.1273 cos 10008 15.95 + 92.01 x 1 4.88 10° cos (10001 ~ 90°) 4.88 X 10-* sin 10000 Jn other words, the impedance is principally that of the four aoouste masses in series 0 that Ue Ings p by nearly 90°. a Ueno Pact VII) Acous: AL crneurrs co) Example 8.5. A Iichnbolts resonator is frequently sed asa means foro ‘an undesired froquency component From an acoustic aystern. An example is given it Fig. 3.392, A constant-foree generator @ produces a serie of tones, among whieh is tance is 500 mks acoustic ohms. If the tuhe haa a cross-sectional 1, = ly = 5.0m, fy = Lem, V = 100 em?, and the cross-sectional aren of Isis 2 em’, ‘what frequeney is eliminated from the #93 @ M, M Mas a Dire cn @) Pia. 8.83. (a) Acoustic device consisting of a constant-force generator G, piston P, tube T with length f+ 1, microphone M, and Helmolta resonator 8 with volume V and connecting tube as shown, (B) linpedanee-type analogous circuit for the deviee of (a) Solution. By inspection we may draw the impedance-type analogous circuit of ‘ig. 4.386, The element sizes are 18 x 0.05 3x10 May = Ma B= = 118 kg/mé Ply 3IBX 001 59 bg ns May = Bb = BOM 50 ke 10 eo 7 * m/newton Co = 5B,” Ta cage ~ TS x TOA me Re) = 500 mks neoustic ohms tis obvious that the volume velocity Us of the transducer Af willbe zero when the shunt branch is at resonance, Hence, eh = I — 1540 satin one Vitals ABB f= 285 epRO-MECHANO-ACOUSTICAL ELRCULTS: (Chap. 3 parr VIL Transducers A transducer is defined as a device for converting energy from one form. to another. Of importance in this text is the electromechanical transducer for converting electrical energy into mechanical energy, and vice versa. ‘There are many types of such transducers. In aeousties we are concerned with microphones, earphones, loudspeakers, and vibration pickups and vibration producers which are generally linear passive roversible networks. ‘The type of electromechanical transducer chosen for each of these instruments depends upon such factors as the desired electrical and mechanical impedances, durability, and cost. It will not be possible here to discuss all means for electromechanical transduction. Instead we shall Timit the discussion to electromagnetic and electrostatic types. Also, we shall deal with mechano-acoustic transducers for converting ‘mechanical energy into acoustic energy. 3.6. Blectromechanical Transducers. Two types of electromechanical transducers, electromagnetic and electrostatic, are commonly employed in loudspeakers and microphones. Both may be represented by transformers with propertico that permit the joining of mechanical aud elec ‘rieal cireuits into one schematie diagram. Electromagnetic-mechanical Transducer. This type of transducer can be characterized by four terminals. Two have voltage and curr associated with them. ‘The other two have velocity and force the measurable properties. Familiar examples are the moving loudspeaker or microphone and the variable-reluctance earphone or microphone. ‘The simplest type of moving-coil transducer is a single length of wire in uniform magnetic field as shown in Fig. 3.34. When a wire is moved upward with a velocity was shown in Fig, 3.34a, a potential difference e will be produced in the wire such that terminal 2 is positive. If, on the other hand, the wire is fixed in the magnetic field (Fig. 3.34) and a current iis caused to flow into terminal 2 (therefore, 2 is positive), a force will be produced that acts on the wire upward in the same direction as that indicated previously for the velocity. ‘The basic equations applicable to the moving-coil type of transducer are fa Bi (8.26a) e= Blu (8.260) where i = electrical current in amperes J = “open-circuit” force in newtons produced on the mechanical circuit by the current i B = magnetic-flux density in webers per square meter Pact VIN TRANSDI RS nm I= effective Jength in meters of the electrical conductor that moves at right angles across the magnetic lines of force of flux density B 1 = velocity in meters per second = “open-circuit” electrical voltage in volts produced by a velocity u ‘The right-hand sides of Eqs. (3.26) have the same sign because when w and f are in the same direction the electrical terminals have the same sign. + i 2 J x (lle HAle if Veloty x wth postive The current produces a ‘recon upward fove f ating upward @ @) Fro, 84, Simplified form of moving-eil transducer consisting of single length of wire cutting a magnetic field of flux density J. (a) The conductor is moving ver- tically at constant velocity so as to generate an open-circuit voltage across terminals Vand 2. (b) A constant current i entering terminal 2 to produce force on the eon ductor in & vertical direation. ‘The analogous symbol for this type of transducer is the “ideal” transformer given in Fig. 3.35. The “windings” on this ideal transformer have infinite impedance, and the transformer obeys Eqs. (3.26) at all frequencies, including steady flow. _{ B!1_{ The mechanical side of this symbol necescarily is of the mobility type if current flows in the primary. te ‘The invariant mathematical operations which this symbol represents are given in Table 3.1, They lead Fre. 3.35. Analo- directly to Eqs. (3.26). The arrows point in the ‘gous symbol for ‘ directions of postive flow or positive potent. eepromagnetic Electrostatic-mechanical Transducer. This type of ca ae Fig. x aa transducer may also be characterized by four termi- 'e mechanical side nals. Av two of them vatagesand eurentacan be yeh the MODIY mmeasured. "At the other te, forts and velostes can be measured. ‘This transducer is satisfactorily described by the fol~ lowing mathematica! relation, = (8270) fan (27)72 ELECT RO.MEGHANO- ACOUSTICAL CIREULTS [Chap 3 where ¢ = “open-circuit” electrical voltage in volts produced by displacement §, displacement in meters of @ dimension of the piezoclectrie device = clectrical charge in coulombs stored in the diclectrie of the piezoelectric deviee. J= “open-cireuit" force in newtons produced by an electrical charge 4, upling coefficient with dimensions of newtons per coulomb or volts per meter. It is a real number when the network is Tinear, passive, and reversible. ‘An example is a piezoelectric erystal microphone such as is shown in Fig. 8.36. A force applied uniformly over the face of the crystal causes an inward displacement of magnitude & As. result of this displacement, 4 voltage © appears across the electrical terminals 1 and 2. Let us assume that a positive displacement ware (nWard) of the erystal causes termi- wiveyaai nal 1 to become postive. For small € ‘Sceneried displacements, the induced vollage is stituted proportional to displacement, ‘The Tue” inverse of thi effect occurs when no $e!" external force acts on the erystal face a but an electrical generator is con- © crete nected to the terminals 1 and 2. If O ws the external generator is connected s0 that terminal 1 is positive, an internal duces ented ee ei areatal 8 Force f is produced which acts to expand the crystal. For small displacements, the developed force f is proportional to the electric charge g stored in the dielectric of the crystal. Equations (8.27) are often inconvenient to use because they contain charge and displacement. One prefers to deal with current and velocity, which appear directly in the equation for power. Conversion to current and velocity may be made by the relations ua B= jot (3.280) a dt jg (3.288) 1 The coupling coetMcient is frequently defined differently in advanced texts on sectrostatic-mochanical transilucres, For example, in some teats it ib defined as the squure root of CyCar'/8, where Ciy and Cy are defined after By, (7). Phe muthor does not intend that the definition for coupling coeficient in this text should pe adopted ag standard; rather, the teri ie used simply for convenience in the Jiseussion, Part VIN TRANSDUCERS 3 so that Eys. (8.27) become (8.29a) (3.290) Unfortunately, the usual analogous symbol for this type of transformer is not as simple as that for the electromagnetic-mechanical type. Two possible forms are shown in Fig. 3.37. ‘The mechanical sides are of the impedance-type analogy. Let us discuss Fig. 3.37a first, The element (%, is the mechanical compliance of the transducer. In order to measure C',, a sinusoidal driving force f is applied to the transducer terminals 3 and 4, and the ; a resulting, sinusoidal displacement g—tos ismeasured. Duringthismeasure- , & ment the electrical terminals are ce 7 short-circuited (e= 0). Averylow 2 “ driving frequency is used so that (a) the mass reactance and mechan- 4 Grit y ical resistance can be neglected, 1H ° The clement C’; is the electrical f We sew capacitance of the crystal mea, . sured at low frequencies with the (o) mechanical terminals open-cir pio, 3.37, Two forma of analogous apm- cuited (u = 0). Application of a bols for pisoslctse anedutor, ‘The Current it. the primary will ectaned ‘ale te of the pens produce a voltage across the con- ‘¥?* denser (%, of ¢/jwC',. This in turn produces an open-circuit force ‘wanedicer (2) if oa @) Bit fe t Bie % Ba ‘a renin’ Ly & () Fie. 8.48. Combined electrostatie-electromagnetic transducers. ial diagram of the device. (2) Analogous circuit with mobilities on mechanical sie, ‘The @ operator ia used for the piezoelectric transduce? (c) Same as (3), except that 24 replaces the three parallel mobilities as shown by (). D POWER 89 ie Ys Loud Im Ce Rex? ote (ny (@) Block mechan- (e) Dual of. ) Because the circuit of (2 has infinite mobility, 6) simplifies to thi form. -(@) and (A) Solution of U by superponition,90 ELECTRO.MECHANO-ACOUSTICAL CIRCUITS [Chap. 3 Fig. 3470, Solving, we get Aw fate =X 10-*p 8280027 X 10-9 = 3078 x 10" fo = jallumn = ~834 X10" f= fot fem 814 x 10% Sa Ly = 10.78 x 10-4 — 9 x 10 In| = ESOT X10" 2 185 newton/mt = 189.8 db re 0.0002 mierobar st = ato Example 310. Two transducers, ones piezoelectric erystal and the other a moving coil in a magnetic fel aro connected to.a mane Ms f 1.543 kg as shown in Fig. 3.480 Determine the stored electrical energy inthe condenser Cy at 100 eps for the following constants: = volt Re = 10 ohms Bom i weber/m? 130m Cy = 4X 107 Farad Maas = LOkg. Cua; = 10" m/newton ‘r= 1.28 X10" = 628 radians/eco Solution, The transducers are shown schematically in () of Fig. 3.48, A further simplification of this diagram is shown in (e). Let us determine the value of zu first. We note that the dual of (dis given by (e), fhe t= jth 4) hy = s1620 + 968) ~ j1503 ~ 0 In other words, the mobility ie infinite at 100 eps, Hence, eireit (¢) simplifies to that shown in (), By superposition, #; can be broken into two parts and 4) given by the two circuits (9) and (A), 0 that f= sf — ¥. 28 x07 Ho ~ 1X30 XO 680 f= it! ~ it = ja X 10-4 lial = 147 X 10° an The voltage drop across the enpacitor is Ied = ATX 10-4 108 = 59 volts The electric stored encegy on the capacitor Ce i B4Crlelt X 10-4 X 8.48 X 108 = 7 10-4 wathnee CHAPTER 4 RADIATION OF SOUND In order fully to specify a source of sound, we need to know, in addition to other properties, its directivity characteristics at all frequencies of interest. Some sources are nondirective, that is to say, they radiate sound equally in all directions and as such are called spherical radiators. Others may be highly directional, either because their size is naturally large compared to a wavelength or because of special design. ‘The most elementary radiator of sound is a spherical source whose radius is small compared to one-sixth of a wavelength. Such a radiator is called a simple source oF a point source. Ita propertion aro eposified by the magnitude of the velocity of its surface and by its phase relative to some reference. More complicated sources such as plane or curved radiators may be treated analytically as a combination of simple sources, each with its own surface velocity and phase. ‘A particularly important consideration in the design of loudspeakers and horns is their directivity characteristics. This chapter serves as an important basis for Inter chapters dealing with loudspeakers, baffies, horns, and noise sourees. The basic concepts governing radiation of sound must be grasped thoroughly at the outset. It is then possible to reason from those concepts in deducing the performance of any particular equipment or in planning new systems, Examples of measured radiation patterns for common loudspeakers are given here as evidence of the applicability of the basic concepts, pant X Directivity Patterns "The directivity pattern of a transducer used for the emission or for recep- tion of sound is a description, usually presented graphically, of the response of the transducer as a function of the direction of the transmitted ot incident sound waves in a specified plane and at a specified frequency. a92 RADIATION OF SOUND (Chop. 4 The beam width of a directivity pattern is used in this text as the angular distance between the two points on either side of the principal axis where the sound pressure level is down 6 db from its value at @ = 0. 4.1. Spherical Sources." A spherical source is the simplest to eon sider because it radiates sound uniformly in all directions. As we saw from Eq, (2.62), the sound pressure at a point a distance rin any direction from the center of a spherical source of any radius in free space is equal to V2 AS jury pin) = ¥ (4) where A, is the magnitude of rms sound pressure at unit distance from the center of the sphere, _ os ro or as ner = so Dr=0DR cy Fro, 4.1, Directivity pattern for a nondirectional source, Such a pattern is drawn on a particular plane intersecting the center of the source. ‘The directivity index DI defined in Pare XI) equals 0 db at all angles, Directivity Pattern. On a polar diagram, the directivity pattern on any plane surface intersecting the center of such a spherical source is given in Fig. 4.1. It is obviously a nondirectional source. Simple Source (Point Source). For the special case of a very small source, whose radius a is small compared with one-sixth wavelength (that. 2d ed,, pp. 311-826, MeGraw-Hill Book 2P. M. Morse, “Vibration and Sound, Company, Ine., New York, 1948. +1, B. Kinsler and A. Frey, “Pondamentals of Acoustics," pp. 168-173, John Wiley & Sons, Ine., New York, 1950, * Morse, op. rit, pp. 312-313, Pact Xi pine STIVETY PAY Ns 93 is, tw & 1), the velocity at the surface of the sphere is [see Eq, (2.63)] V2A, ¢ wat) Ln pea joxfa puts 42) A sourve for which (his formula is valid is called a simple sonrce. Substitution of Bq. (4.2) into Bq. (4.1) yields Pan) (43) Us = rms volume velocity in eubie meters per sevond of the very small source and ix equal to (40°), rms ound pressure in newLons per square meter ata distance 1 from the simple source Strength of a Simple Source.* ‘The rms magnitude of the total sir flow at the surfare of a simple source in cubie meters per second (or eubie centimeters per sevond in the egs system) is given hy Us and iy called the strength of a simple source.t Intensily al Distance r. Ata distance r from the center of a simple sonree the intensity is given by P walts/m? ay When the dimensions of a source are much smaller than a wavelength, the radiation from it will be much the same no matter what shape the radiator has, as long as all parts of the radiator vibrate substantially in phase. The intensity at any distance is directly proportional to the square of the volume velocity and the frequency. 4.2. Combination of Simple Sources.‘ ‘The basie principles governing the directivity patterns from loudspeakers can be learned by studying combinations of simple sources. ‘This approach is very similar to the consideration of Huygens wavelets in opties. Basically, our problem is to add, vectorially, at the desired point in space, the sound pressures arriving at. that point from all the simple sources. Let us see how this method of analysis is applied. Two Simple Sources in Phase, ‘The geometric situation is shown in Fig. 4.2. Itis assumed that the distance r from the two point sources to the point 4 at which the pressure p is being measured is large compared ‘with the separation 6 between the two sources. “The spherical sound wave arriving at the point p from source 1 will have fm sonar (ext the peak mnagitude of the otal ae Hos i taude is used. tn the ‘IL F. Olam, "Elements of Aconstical Engineering,” Nonteand Company, Ine, New York, 1947, tea of the ris wag V2 reine ed ppt, 1D. Van text, the xlrength of wimp nie94 RADIATION OF SOUND Chap. 4 traveled a distance r — (b/2) sin 6, and the sound pressure will be pir) = YEAS pep tnmenan i ‘The wave from source 2 will have traveled a distance r + (b/2) sin 8, 50. that Vids patra) = ele Oa HD, (4.50) ‘The sum of pi + ps, assuming r>> b, gives ViAy pir) cig HAA (ileb ah fe gion) (46) Multiplication of the numerator and the denominator of Eq. (4.6) by ia, 42. Two simple (point) sources vibrating in phase located a distance b apart and at distance r and angle 8 with respect to the point of measurement A. exp (jrb sin 6/2) ~ exp (~Jjeb sin 6/A) and replacement of the expo- nentials by sines, yields — VIAL joie staeeny Sit. [(2xb/d) sin 6] ae “sin [(#b/d) sin @] ed) ‘The equation for the magnitude of the rms sound pressure |p| is ip) = 24s | in [C2xb/) sin 6) en 7 | Bsin (Gab /A) sin 8] ‘The portion of this equation within the straight lines yiels the directivity pattern Referring to Fig. 4.2, we see that if bis very amall compared with a wavelength, the two sources essentially coalesce and the pressure at a Part Xi] DIRECTIVITY PATTERNS 9 distance r at any angle @ is double that for one source acting alone, Th: directivity pattern will be that of Fig. 4.1 As b gets larger, however, the pressures arriving from the two source will be different in phase and the directivity pattern will not be a circle 0/4: 01-09 08, 22: 01-3008 Tae ior dea2 —pIe3008 ba 01-2008 @ a Fic. 43. Directivity patterna for the two simple in-phase sources of Fig. 4.2. Sym: metry of the directivity patterns occurs about the xis Passing through the two sources. Hence, only a single plane ie necessary to deeeribe the directivity character. istis at any particular frequeney. The boxes give the directivity index at. 9 = 0" One angle of zero directivity index is also indicated, (The directivity index is ‘cussed in Part XI.) In other words, the sources will radiate sound in some directions better than in others. Asa specific example, let b = /2. For @ = 0 or 180° it is clear that the pressure atriving at a point 4 will be double that from either source, However, for 6 = 490° the time of travel between the96 RADIATION OF SOUND pt pletely cancels the radiata from the other. Hee the presure at all points along the +90° axis is zero. Remember, we have limited our dis- tumion tors Directivity patterns expressed in devibels relative to the presuire at = 0, are given in Fig, 43 for the two inephave sources with & = 2/4 \/25¥5 3/3; and 20, very important observation can be made from the directivity p tomns for thissimpl typo of radiator that applies to all types of rai niger the exlont of Uno radiator (here, the ge ill be tho prinipal lube along the @'= 0 axis at any give quency and the greater the number of side lobes. As we shall see in the next paragraph, iis posible tosup- pres tho side lobes, thats Lo sey, those other than the prineipal lobes atOand 180%, by simply increasing the number of elements. Linear Array of Simple Sources.* ‘The goometite situation for this type of radiating array is shown in _ Big 14. The rns sound pressure ico vibrating input Mets produced at a point a by m identi- Tauern gpa SE eee ee a straight Tine, the sources diy and with the extent tance d= (x ~ Hb small compared with the dist nity | in [(n 2) sin livin (Gxb/A) sin 49) As a special ease, very large and th before, us assume that the number of points becomes separation b becomes very small. ‘Then, as : d= (n= 1b & nb (4.10) a sin [xd a aay P= pe where ps is the ide of the rms sound pressure at a distance r from the array at an angle @ = 0. As before, it is assumed that the extent of the array d is small compared with the distance r Plots of Eq. (1.9) for w= 4 and d= A/M, 4/2, A, 3n/' shown in Fig. 4.5. lar plots for n —+ & and b> 0, that is 4.5, and 4.6 for a given ratio of array length Lo wavelength is in the suppression of the “side lobes.” Part XI DIRECTIVITY PATTERNS 7 ‘That is, sound is radiated well in the @ = 0° and @ = 180° directions for all three arrays. However, as the array becomes longer and the number of elements becomes greater, the radiation becomes les in other directions ‘than at @ = 0° and @ = 180°. 4/4: 01-08 08 ana v2:01-1508 a Ste ERG KRY [iS $e) Si @-22 01-6008 4-2) 01-6608 @ a iretivity patterns for a linear array of fou simple i-phase sourees evenly spaced overs length The boxes give the directivity index at @'= 0% One angle Of zero dretivity index is also indested by the azom. Doublet Sound Source. A doublet, sound source is pair of simple sound sources, separated a very small distance b apart and vibrating in opposing phase. The geometric situation is shown in Fig. 4.7. The distance r to the point A is assumed to be large compared with the separation b between the two sources Tt ean be clearly seen that the sound pressure at, @ = 90° and @ = 270° will be zero, because the contribution at those points will be equal from98 RADIATION OF SOUND (Chap. 4 the two sources and 180° out of phase. ‘The pressures at @ = 0° and 9 = 180° will depend upon the ratio of 6 to the wavelength ». For example, if b = 2, we shall have zero sound pressure at those angles just as we did for b = 4/2 in the case of two in-phase sources, In the present case, we have a maximum pressure at @ = 0° and @ = 180° for b = 2/2. yA: 01-0308 N21 08 Toe od @-a/2 Ol-8108 4-2 Di-6208 @ @ Fro. 46, Directivity patterns for a linear line array radiating uniformly along ite length d. ‘The boxes give the directivity index at @'= 0". One angle of sero direc tivity index is also indieated by the arrow. ‘The usual case of interest, however, is the one for ben (4.12) In this case, the complex rms pressure py at a point A can be shown to equal’ * Kinsler and Prey, op. cit, pp. 280-285, Part X} DIRECTIVITY PATTERNS 99 212 (44-2) anvem 1) ‘msstrength in cubic meters per second of each simple source. ratio of the complex rms sound pressure py produced by the oublet to the complex rms sound pressure p, produced by a simple source is found by dividing Eq. (4.13) by Bq. (4.8). This division yields pelty@) = Ba 4 si) ost aay ‘When the square of the distance r from the acoustic doublet is large Fra. 4.7. Doublet sound souree, This type of source consists of two simple (point) sources vibrating 180° out of phase. They are located a distance b apart and are at fan angle # and a distance r with respect 10 the point of measurement A. The lower half of the graph shows by the aren ofthe circles the magnitude of the sound preagure ‘sa function of angle #. ‘The upper half of the graph showa the variation ofthe radial ‘and tangential componenta ofthe particle velocity as a function of angle @ compared with 2/36 (k'+* >> 1), Eq. (4.18) reduces to pe = OUD cos em (4.15) For this case the pressure varies with @ as shown in Figs. 4.7 and 4.8. It changes inversely with distance r in exactly the same manner as for the simple source. Near the acoustic doublet, for r? < \*/36, Eq. (4.18) reduces to pr = 29UB cos pen 10)100 WADIA YN OF SOUND. Jechap. Vor th eos @ as ws in Big 18, bt iLchangesinversely with Use square of the distanver, We arestill assume ing that r > 6. Near-field and Far-field. ‘The difference between nearsfield and far-field behaviors of sources must always he borne in mind, When the diveetivity pattern of a loudspeaker or some other sound source is presented in technical publication, itis always understood that the data were taken 2 a distance r sulliiently large so that the sound prosure was decreasing. nearly with distance along a radial line connecting with the source, a8 was the ease for Eq. (4.15). ‘This is the far-field ease. For this to be true, two conditions usually have to he met. First, the extent 6 of the the pressure also varies wi ‘DI=88 DB W/ decty 20 0 (9/0) (a) () Fis, 48. Dinectivity pattern for & doublet sound sonree. (a) Sound-pressure ratio Pipes. &. (b) 2 logis p/pe vs. 8. The boxes ive the dirvetivity index xt 9 = 0 ne angle of ser0 directivity index ie aleo indicated by the arrow, radiating array must be small compared with r, and 7? must he large compared with 2/36. In acousties the size factor indicated is usually taken to be larger than 3 to 10, One more item is of interest. in connection with the acoustic doublet, ‘The particle velocity is composed of two compouents, one radially directed, and the other perpendicular to that direction. At @ = 0 and 180° the particle velocity is directed radially entirely (soe Rig. 4.7). At 4 = 90 and 270° the particle velocity is entirely perpendicular to the raglial line. In between, the radial component varies ws the eos @ and the perpendicular component as the sin 0. ‘An interesting fact is that at @ = 90° and 270° a doublet sound source appears to propagate a transversely polarized sound wave, ‘To dem strate this, take two unbaflled small loudspeakers into an anechoie chamber. “Unbaffled loudspeakers (trunsducers) are cyuivalent to doublets herause the pressure increases on one side of the diaphragm Part X} DIRECTIVITY PATTERNS 101 whenever iG decreases on the other Hold the two transducers about 0.5 m apart with both diaphragms facing the floor (wot facing each other). Tot one transiuver radiate a low-frequency sound and the other act as & microphone connected to the input of an audio amplifier. As we see from Fig. 4.7, no sound pressure will he produced at the diaphragm of the microphone, but there will be transverse particle velocity. A particle is always the result of a pressure gradient in the direction of the velocity. ‘Therefore, the diaphragm of the microphone will be eaused to move when the two transducers are held as described above. When fone of the transducers is rotated through 90° about the axis joining the units, the diaphragm of the microphone will not move because the pressure gradient will be in the plane of the diaphragm, Hence, the sound wave appears to be plane polarized. ‘You have now learned the elementary principles governing the diree- tional characteristies of sound sources. We shall be able to use these Fia, 49, Rigid cireular piston in. rigid baffle. The point of measurement A ie located at distance ran angle @ with respect ta the center of the piston principles in understanding the measured or calculated behavior of some ‘of the more complicated sound sources found in acoustics, 43. Plane Piston Sources. Rigid Circular Piston in Infinite Bate Many radiating sources can be represented by the simple concept of vibrating piston located in an infinitely large rigid wall. ‘The piston ir assumed to be rigid so that all parts of its surface vibrate in phase and ite velocity amplitude is independent of the mechanical or acoustic loading on its radiating surface. ‘The rigid wall surrounding the piston is usually called a baffle, which, by definition, is a shielding structure or partition used to inerease the effective length of the external transmission path between the front and back of the radiating surface. ‘The geometry of the problem is shown in Fig. 4.9. We wish to know the sound pressure ata point A located at a distance r and an angle @ from the center of the piston. To do this, we divide the surface of the piston into a number of small eloments, each of which isa simple source vibrating in phase with all the other elements. ‘The pressure at 4 is, then, the sum in magnitude and phase of the pressures from these elementary elements,102 RADIATION OF SOUND (Chap. 4 ‘This summation appears in many texts' and, for the case of r large compared with the radius of the piston a, leads to the equation V2 ifestuera® [@ (ka sin 6) r asin 108 Jan Ja(_) = Bessel function of the first order for cylindrical coordinates* Fro, 4.10, Directivity patterns for a rigid cieular piston in an inGnite bale ws function of ka = 2ra/hy mhere a is the radius of the piston. The boxes give the directivity index at @= O°. One angle of zero directivity index sls indieated. The DI never hecomes less than 8 db because the piston radiates only into half-space, The portion of Eq. (4.17) within the square brackets yields the directivity pattern and is plotted in decibels as a function of @ in Fig. 4.10 for six values of ka = 2ra/a, that is, for six values of the ratio of the cireumference of the piston to the wavelength. ‘When the circumference of the piston (2na) is less than one-half wave- * Morse, op. ct, pp. 826-846. A table of Besocl functions is given age 444 Part XI DIREGTIVETY PATTERNS 103, length, thats, ka < 0.5, the piston behaves essentially likea point souree. When ka hecomes greater than 8, the piston is highly directional. We see rom Fig. 4.24 that an ordinary loudspeaker aiso becomes quite directive at higher frequencies in much the same manner as does the vibrating piston, Rigid Circular Piston in Bnd of a Long Tube? In many instances, sound is radiated from a diaphragm whose rear side is shielded from the front side by a box or a tube. If the box does not extend appreciably beyond the edges of te diaphragm, its performance may be estimated by comparison with that of a rigid piston placed in the end of a long tube ‘Phe geometrical situation is shown in Fig. 4.11. ‘The pressure at point A is aguin found by suunming the pressures from a number of small elements on the surface of the piston, each acting as a simple source. ‘The A Long estar ston ‘of slong tube, ‘The point of measurement "espeet tothe eenter of the piston. solution of this problem is complex, however, becwuse radiation ean take place in all directions and the sound must diffract around the edge of the tube to get to the left-hand part of space (Fig. 4.11). Hence, a theory that includes the effects of diffraction must be used in solving the problem analytically. ‘The results from such a theory are shown in Fig, 4.12 for six values of ka. It is assumed here also that the distance r is large compared with a, so that the directivity pattern applies to the far-field. Rigid Circular Piston without Baffe. To complete the eases wherein pistons are commonly used, we present the results of theoretical studies on the directivity pattern of a rigid piston of radius a without any baffle, radiating into free space. ‘These results are shown graphically in Fig 4.13 for four values of ka. It is interesting to note the resemblance hetween these curves and those for an acoustic doublet. In fact, to 8 lirst approximation, an unbaflted thin piston is simply a doublet, because an axial movement in one direction compresses the air on one side of it and causes a rarefaction of the air on the other side. Levine and J. Schwinger, On the Radiation ofS le Pine, Phy. fen, T8: 481-4106 (Feb, 1, 1948) "PUM. Wiene, On the Kelation between the Sound Fields Raviated and Diffracted hay Phan Obstacles, J. dewnat, Soe. Amer., 98: 697-700 (1951), from an Unflanged Cire:RADIATLON OF SOUND Part X} DIRECTIVITY PATTERNS 105 Square or Rectangular Piston Sources.» natn SQUARE PISTON IN INFI- NITE marruE. This type of radiating souree is not very common in acoustics. Tt suffices to say here that the directivity pattern for such a piston in a plane perpendicular to the piston and parallel to a side is, identical to that fora linear array of simple sources as given by Eq. (4.11). ‘The pattern in a plane that is parallel to either of the two diagonals differs very little from that in a plane parallel to a side. i PISTON IN INFINITE BAFFLE. The directivity pat- radiating source with dimensions d, and ds are given ,] sin ((rds/A) sin Oa] (ed2/2) sin By (418) * Olson, op. it, pp. 39-40,106 RADIATION OF SOUND (Chap. + where # = angle between the normal to the surface of the piston and the projection of the line joining the middle of the surface and the observation point on the plane normal to the surface and parallel to ds 4, = same as 8, with ds substituted for dy Note that the directivity pattern is equal to the product of the dir fivity pattems for two line arrays at right angles to each other [s Bq. (4.11)]. 44. Curved Sources."® Tn the preceding paragraphs of this part we have dealt, with the radiation of sound from straight-line and plane- surface arrays. These types of arrays are closely resembled by open- ended organ pipes, direct-radiator loudspeakers, simple horns, and other devives. One characteristic common to all these arrays, except the doublet array, is that they become more directional as the ratio of their length to the wavelength becomes greater. ‘This is usually an undesirable trait for a loudspeaker to exhibit beeause it meaus that the spectrum athe F10.4.14. Parabolic megaphone suitable fur use by a cheesleader ina football stadium, of music or speech ax reproduced will vary from one position to another around the loudspeaker. ‘To overcome, in part, this increase in directivity with inereasing frequeney, curved surfaces are commonly employed as sound radiators Thi aces can be made up of a number of small loudspeakers or small homs or as a megaphone with a eurved front. Curvod-line Source (Parabolic Megaphone). An example of a eurved- line source is the parabolic megaphone of Fig. 4.14. ‘The megaphone ‘opening is thin enough (1 in.) to be roughly equivalent to a simple line source for frequencies below 4000 cps. ‘The horn is parabolie because the sectional area is proportional to Uhe distance from the apex. ‘The sound pressure at a point A a distance r from the apex is found by summing the pressures in amplitude and phase originating from an assumed curved line coimposed of simple sources. When the distance ris "Ibi, pp 40, Part XI DIRECTIVITY PATTERNS 107 Fro, 4.15. Directivity patterns for the parabolic mogaphone of Fig 4.14 in the plane containing the ate of the opening (6) Fie, 4.16, Multicclular horas with curved radiating fronts, (@) 3X 5 = 16 cells 0) 2X4 = Bealls, large compared with the radius 2 of the horn, we obtain the directivity patterns shown in Fig, 4.15. ‘This simple case illustrates basic principles applicable to all curved radiating sources. At low frequencies the curved source is largely nondirectional. As the frequency is increased, the souree becomes more108 RADIATION OF SOUND (Chap. 4 directional, achieving its minimum angle of spread when the chord of the curved source is approximately equal to one wavelength. At high frequencies, the directivity pattern becomes broader, reaching its maximus. width when it becomes equal to the width of the are expressed in degre for example, 90° in Fig. 4.14, Curved-surface Sources (Multicellular Horns). ‘The most common curved-surface sources found in acoustics at the present time are multicellular homs. Two typical examples of such curved-surface horus are shown in Fig. 4.16, ‘The beam width of the directivity pattern was defined as the angular distance between the two points on either side of the principal axis where 160 pean} i i e Rarer oes & E 8 OT gg tg FR Frequency in cycles per second Fic. 4.17, Beam widths of multicellular horns constructed as shown in the inscst und as sketehed in Fig 10, the sound pressure level isdown 6 dbf from its valueat@ = 0. ‘The beam widths of the directivity patterns for two, three, four, and five cell widths of multicellular horns were measured on commercial units and are shown n Fig. 4.17. ‘These data are useful in the design of sound systems, It should be noted that the minimum beam width occurs when the are the multicellular horn about equals X. Also, at high frequencies the beam reaches a width of about n - 25° — 15° for the size of cell shown in Fig. 4.17. ‘The theoretical directivity indexes for these maximum widths of beams are found from the nomogram in Fig. 4.22, which is based on the sketch of Fig. 4.21 {No standard value has boon chosen for dhe nuanber of devibels down from the 4 = 0 value of the sonnal pressire level in determining beam width, Values nf 3s 12d 10 dy are often encountered in the literature Pact XH DIRECTIVITY INDEX 109 pare XI Directir ily Inder and Directivity Factor Charts of the directivity patterns of sound sources are sufficient in many cases, such as when the source is located outdoors at a distance from reftecting surfaces. Indoors, it is necessary in addition to know something about the total power radiated in order to calculate the reinforcing effect of the reverberation in the room on the output of the sound source. A number is calculated at each frequency that: tells the degree of directivity without the necessity for showing the entire diree- tivity pattern. This number is the directivity factor or, when expressed in decibels, the directivity index. 4.5. Directivity Factor [Q(/)]. ‘The directivity factor is the ratio of the intensityt on a designated axis of a sound radiator at astated distance r to the intensity that would be produced at the same position by a point, source if it were radiating the same total acoustic power as the radiator. Free space is assumed for the measurements. Usually, the designated axis is taken as the axis of maximum radiation, in which case Q({) always: exceeds unity. In some cases, the directivity fuctor is desired for other axes where Q(/) may assume any value equal to or greater than zero. 4.6. Directivity Index (DI(f)}, ‘The directivity index is 10 times the logarithm to the buse 10 of the divectivity factor. DI(A) = 10 log QC) (4.19) 4.7. Calculation of Q(/) and DI(f). ‘The intensity J ata point removed distance r from the acoustical center of a source of sound located in free space is determined by first measuring the effective sound pressure p and Jetting / = |pl/a. If the source is a point source so that I is not a function of @and is located in free space, the total acoustic power radiated W, = tert If the source is not @ point source, the total acoustic power radiated is determined by summing the intensities over the surface of a sphere of radius r. ‘That is, the total radiated power is ar er) where the coordinate of any point in space is given by the angles @ and and the radius r (see Fig. 4.18) and p%(6,9,7) equals the mean-square sound pressure at the point designated by 6, ¢, and r. {See the definition for intensity om page 11, The intensity equals the sound pres sure squared, divided by aye fora plane wave in Free space o10 RADIATION OF SOUND Chap. 4 Usually an analytical expression for p(0,6) does not exist. In practice, therefore, data are taken at the centers of a number of areas, approximately equal in magnitude, on the surface of a sphere of radius r surrounding the source. As an example, we show in Fig, 4.19a a spherical Fic. 4.18. Coordinate system defining the angle @ and ¢ and the length r of a line connecting a point A to the center of a sphere. ‘The area of the incremental surface aS =r? sin Ode ds. erton view () i 2 KN 9 Ly . View trom top of sphere Pian view (a) (e) io, 4.19, (a) Division of a spherical surface into 20 equal areas of identical shape, (@) and (0) Division of hemisphere into 8 parte of equal aree but unequal shape. surface divided into 20 equal parts of the same shape. The measured intensities on each of these paris may be called I,, Is, I, ete. ‘The total power radiated W is w + TSu0 (4.21) Part XI] DIRECTIVITY INDEX m where Si, Sx. . . Sie are the areas of the 20 parts of the epherical surface. If, as in Fig. 4.19a, the surface is divided into 20 equal parts, thea 81 = & = S Ste. For less critical cases, it is possible to divide the spherical surface into 16 parts of equal area but of different shape, as we can see from Fig. 4.19 and c. By definition, the directivity factor Q(f) is Part a — oe WY Front sin oaoas where [pad is the magnitude of the mean-square sound pressure on the designated axis of the sound source at a certain distance r (see Fig, 4.23, 0° axis, as an example) For the special case where, for any particular value of @, the sound pressure produced by the sound source is independent of the value of 4, that is to say, there is an axis of symmetry, Eq. (4.22) simplifies to on (4.22) spat 2 [96 sina ‘The magnitude signs are left. of for convenience. Many sources, such as loudspeakers, are fairly symmetrical about the principal axes so that Eq. (4.23) is valid. In this case, data are generally. taken at a number of points with the angles @ in a horizontal plane around the source so that. a = (4.23) AN = - ae SNe sin 28 (424) where Ad = separation in degroes of the successive points around the sound source at which measurement of p(@,) was made (see 4.23 as an example). 180°/A9 = number of measurements that were made in passing from a point directly in front of the souree to one directly behind the source (0 to 180"). The sound source is assumed to be symmetrical so that the variation between 360 and 180° is the same as that between 0 and 180°, If the source is mounted in an infinite baffle, measurement. is possible only in a hemisphere. Hence, the value of n in Eq. (4.23) varies from 1 to 90°/a8. If the source in an infinite baffle is nondirectional in the hemisphere, which is usually the case for ka < 0.5, then the directivity factor Q = Q = 2, that is, DI = 3 db If the directivity pattern is not quite symmetrical, then the factor of 4ne RADIATION OF SOUND (Chop. 4 in the numerator of Eq. (4.23) becomes 8 and the value of m varies from 1 to 360°/A8. This, in effect, averages the two sides of the directivity pattern. ‘The directivity index at @ = 0° for each directivity pattern shown in Part X and in this part is written alongside each directivity pattern. » 00 0 i e as 8 i S gu 20 bu : $19) 10S ee re z ints tae ea ES BiE------- 5 5 7 a f gh pete tan(uofe so 4.20, Diecttey indexes or the elt sun (2) one ae ty of pt in an ‘Minite place bales (2) pst inthe end of long tube, and @) a piston i ee Space without any bale ‘The reference axis is the principal axis at @ = Oin every case. An angle @ at which the directivity index equals 0 db is also marked on these graphs, Hence, the directivity index at any other angle @ can be found by subtracting the decibel value for that axis from the decibel value at the axis where DI = 0. Fra, 4.21. Radiation into a slid cone of space defined by the angles a and 0. For easy reference, the directivity indexes for a piston in (1) an infinite plane baffle, (2) along tube, and (3) free space are plotted as a function of ka in Fig. 4.20. ‘Many horn loudspeakers at high frequencies (above 1500 ops) radiate sound uniformly into a solid rectangular cone of space as shown in Fig. 4.21. These horns are of the type discussed in Par. 4.4 (page 106). The directivity indexes in the frequency range above 1500 cps as a function of Part XI} DIRECTIVITY INDEX us a and 8 may be estimated with the aid of the nomogram given in Fig. 4.22.01 Detailed calculations are shown in Table 4.1 for a box-enclosed loud speaker having the directivity pattern at a frequency of 1500 eps shown Bieta index DB 30: oe 50g 60. 70°. 105 ds 135 wot £ ‘0° 0 Fra. 4.22. Nomogram for determining the directivity indexes of a source of eound fadiating uniformly into a solid cone of apace of the type shown by Fig. 4.21. [After ‘Molloy, Calculation of the Divectivity Index Jor Various Types of Radiators, J. Acoust Soe. Amer, 20: 387408 (1948).] in Fig. 428, The left (L) and right (2) sides of the directivity characteristies are not alike, so that the averaging process of the previous paragraph is used. = C.T. Motlay, Caleulation of the Directvity Index for Various Types of Radiators, J. Acoust, Soe, Amer, 20: 987-405 (1048).us RADIATION OF SOUND (Chap. 4 Fic. 4.23, Measured directivity patterns for a typical 12in. direct-radiator loud speaker in a 27- by 20- by I2in, rectangular box. "The squares give the directivity index at # = 0", One angle of zero directivity index is nls indicated. sL | iE 1 or - 7 _ | Rae ] : olen 1 oobi. 1 Legal att the ah Freqeney in ys per scond Fic. 424, Dinetivity indexes for O° axe ofthe directivity pattern of Fig. 4.23 computed an though the souree were symmmeriiealsboot the O" ans. 'Phe dats ppl to & {apical 12m, diet-aintor loudspeaker mounted in» 27-by20-by 12, rectangular Part XI DIRECTIVITY INDEX 15 After a directivity factor has been calculated at each frequency, a plot of directivity index DI({) in decibels is made with the aid of Eq. (4.19) For the loudspeaker with the directivity patterns of Fig. 4.23, the directivity index as a function of frequency is shown in Fig. 4.24, TABLE 4.1, Calculation of Directivity Index DIG) Directivity 355 | 5 | 0.087 1.00 | 0.95 | 0.09 343 | 15 | 0.250 1.00 | 0.79 | 0.96 a5 | 25 | 0.423 om | 050 | 030 325 | 35 | 0.574 0.45 | 0.25 | 0.26 ais | 45 | 0.707 oz | o10 | 0.19 aos | 55 | 0.819 14 | oot | on as | 65 | 0.900, o.o7 | 0.01 | 0.06 285 | 75 | 0.968, 0.03 | o.o1 | 0.03 275 | 5 | 0.996 0.01 | oor | 0.01 265 | 95 | 0.990 o.01 | oor | oor 255 | 105 | 0.960 oor | oor | o.08 245 | 15 | 0.906 oor | 0.01 | 0.01 23s | sas | oso o.o1 | oor | 0.01 0.01 0.01 0.00 0.00 0.00 Be X 573" a) DICH = 10 log 104 = 10.2 db {ALY = 1500 eps for a commercially available loudspeaker having the directivity, patterns shown in Fig. 4.23. ‘The quantity a@ = 10° = »/18 radians,CHAPTER ACOUSTIC COMPONENTS pant XIT Radiation Impedances rece eee The fields of radio and television have advanced rapidly beeause of the bility of electrical components with well-known physical properties that are simple to assemble into a completed mechanism. With such components (resistors, capacitors, and inductors) the advanced research engineer and the high-school student alike are able to experiment with new circuits. Such complicated devices us electrie-wave filters often can he designed hy selecting from among readily available parts until a desired performance characteristic is achieved: -a feat that otherwise might require a lengthy mathematical aualysis, No such satisfactory situation exists in the field of acoustics. Acous- tieal elements have not been available commercially. Advanced text~ books have often side-stepped the theoretical treatment. of their per. formance, Even those texts which deal primarily with acoustic devices sive limited information on how to predict the performance of eavities, holes, tubes, screens, slots, and diaphragms—the elements of acoustical circuits. ‘This text does not. pretend to advance the science of aroustieal= circuit theory to anything approaching a state of completion. Much basic research remains to be done. It does attempt to interpret the {atest theories in such a way that the reader ean construct and understand the performance of the usual types of acoustie devices, Loudspeakers, microphones, and acoustic filters are the most common devices composed of mechanical and acoustical elements. One obvious acoustical element.is the air into which the sound is radiated, Others are ir cavities, tubes, slots, and porous sereens hoth behind and in front of actively vibrating diaphragms. ‘These various elements have acoustic impedances associated with them, which can, in some frequeney ranges, he represented ax simple lumped elements. ‘In other frequeney ranges, ‘listributed elements, analogous to electric lines, must be used in explain. ing the performance wf the devices. U6 Part XIN) RADIATION IMPEDANCES wt The frst acoustical element that we shall deal with is the radiation impedance of the air itself. Radiation impedance is a quantitative statement of the manner in which the medium reacts against the motion of a vibrating surface. Sound is produced by vibrating surfaces such as the diaphragm of a loudspeaker. In addition to the energy required to move the vibrating surface itself, energy is radiated into the air by the diaphragm, Part of this radiated energy is useful and represents the power output of the loudspeaker. The remainder is stored (reactive) energy that is retumed to the generator. Consequently, the acoustic impedance has a real part, accounting for the radiated power, and an imaginary part, accounting for the reactive power. a ao #410 © @ Fic, 5.1. Exact radiation impedances and mobilities for all values of ka for a sphere with “a surface that vibrates radially, (a) Mechanicalimpedence analogy (0) acoustie-impedance anvlogy; (¢) mechanieal-mobility analogy? (@) scoustie mobility aanslogy. The quantity aia the radins of the sphe ‘The four simplest types of vibrating surface treated here are (1) a pulsating sphere, (2) a plane circular piston mounted in an infinite surface (baffle), (3) a plane circular piston in the end of a long tube, and (4) a plane circular piston without baffle. We have already derived the radiation impedance for a pulsating sphere. ‘The mathematical solution of the radiation from a circular piston mounted in an infinite baffle appears in many advanced texts so that only the results will be presented here. ‘More complicated problems are to solve analytically for the radiation impedances and directivity patterns of a long tube and a vibrating piston without baffle. Those solutions are now available, and the results are given in this part. Most other types of vibrating surfaces are exceed- ingly difficult in mathematical treatment, and the results will not be resented here, 1PM. Morse, “Vibration aod Sound,” 2d ed., pp. 326 :46, MeGraw-Ifill Book Company, Ine,, New York, 1948,ir ACOUSTIC COMPONENTS. (Chop. 5 Pore rela yee tote Guat and ie 2 eee een eee eee i eaaeN cine releieeee eae eluate aces siiiccomparel wits he once eon son lernerlaeag ay thet Aeneas ceer aeeurary Sacco ucemsa aera ates Ee oni cane eee gaits ene nae aes Seca eteeeiegts igucrger —— uiigntieutval Fie bt The de sion nina te mobi and impelavee are g 2 Plane Chesac Piston in Yntite Beas "tho mechan! impede Zu Wa + iN = rate where Za = = rallius of piston in density of gus in kilograms per eubie meter, speed of sound in meters p mechanical resistance in newton-seeonds per meter. ‘The naan 9 indieates that the resistive component is a funetion of frequency. Xue = mechanical reactance in newton-seconds per meter, k= w/e = 2s/X = wave number. Ju, Ky = bwo types of Bessel fanetion be second, Tw ven by the series. Kinder ar ALR. Frey, Wiley & Sons, Hue, New York, 1050, "Mors, op rity pp. 392, 383. Morse gives in (ka) that eqaly K ken) 2a #G,N, Watson, "Theory of Bowl Lomlon, 1922 lamentabs of Acoustien,” pp. 187-195, Jobe ible VIIL on page 447 a function Puetions," Cambridge University Pres Pert XU) ne UW) (62) KW) (63) where W = 2ka, 0.0005 = 0.0003 0.0002 coo! stu shu shu aor “oo” 1 03," 19 30" 10 alu Fia. 53. Real and imaginary parts of the (Zu /ratpr) of the air oad on one side of « plan Infinite fat baile. Frequeney is plotted on a normalized seale, where kn = 2efa/e = 2ea/d. Nota also that the ordinate is equal to Zera%/oye, where Z is the beoustie impedance. Graphs of the real and imaginary parts of Za _ Wut waipe ~~ ray 64) ‘are shown in Fig. 5.3 as a function of ka. The German §t indicates that the quantity varies with frequency. Similar graphs of the real and imaginary parts of the mechanical mobility, aunty = vatpye(tu + jz) = ratpae (att a tea) 65)120 ACOUSTIC COMPONENTS [Chap. 5 are shown in Fig. 5.4. ‘The mechanical mobility is in meters per newton- second, ie., mks mechanical mohms. ‘The data of Fig, 5.3 are used in dealing with impedance analogies and the data of Fig, 5.4 in dealing with mobility analogies We see from Fig. 5.3 that, for ka < 0.5, the reactance varies as the first power of frequency while the resistance varies as the second power of fre quency. At high frequencies, for ka > 5, the reactance becomes small 200) 100} 1g] 50 330} os og 02 oy 05 |= 09 x2} Webi analogy 00 ty ot Oo "093° 01 03 19 30° 10 ‘ha Fic, $4. Real end imaginary parts of the normalised mechanical mobility (rato) Of the air load upon one side of a plane piston of endius @ mounted in an infinite flat balfle. ‘Frequency ig plotted on & normalized stale, where ba ~ 2sfa/e = 2ra/>. [Note also that the ordinate is equal (o z4py-/ra, where 24 is the acoustic. mobility. compared with the resistance, and the resistance approaches a constant value, ‘The mobility, on the other hand, is better behaved. ‘The responsiveness is constant for ka < 0.5, and it is also constant for ka > 5 although its value is larger: Approximate Analogous Circuits. ‘The behavior just noted suggests that, except for the wiggles in the curves for ka between 1 and 5, the impedance and the mobility for a piston in an infinite baffle can be approximated over the whole frequency range by the analogous circuits Part X11] RADIATION IMPEDANCES 12a of Fig. 5.5. ‘hose circuits give the mechanical and acousticimpedances ‘and mobilities, where Rus = x0%p9 mks mechanical ohms (newton-see/) 68) Ru = Rus + Rw 128a%pec/9e = 4.53a%pee mks mechanical ohms (5.7) Reus = 1.386a"p¢ mks mechanical ohms 68) Cus = 0.6/apee* m/newton 69) Man = Sa'pa/3 = 2.67a%p9 ke (6.10) ras = I/na'pe ~ 0:318/a%pee mks mechanical mohms 6.1) rus = 0.7210"p9 mks mechanical mohms (12) Rea = pec/ra® = 0318 pu¢/a? mks acoustic ohms (613) Ra = Raat Ras = 128pe/ 0.158.e/a? mks acoustic ohms (5.14) Rey = 0.1404p.e/0% mks acoustic ohms (6.15) Cay = 5.940" / pyc? m*/newton (5.16) May = 8po/30'a = 0.27 p0/a kg/m* (5.17) Tax = 7a*/pec mks acoustic mohms (6.18) Pas = 7.128"/pye mks acoustie- mohms 6.19) ot y ao Im ten am Mh Mn Za Ie iM, Rue Raa () @) te) @ Fic, 5.5. Approximate radiation impedances nnd mobilities for a piston inn jofi ic halfe of for a piston in the rad of along tube for all values of fa. (e) Mechanieal- Impndanee analogy, (0) aeousti-iinpedance analogy ; (2) mechanienl-smobility analogy {Ud} acousticemobility analogy. |All constants are dimensionless and were chosen to give the best average fit to the functions of Figs. 5.3 and 5.4 Low- and High-frequency Approximations. At low and at high frequencies these circuits may be approximated by the simpler cireuits given in the last column of Table 5.1 It is apparent that when ka < 0.5, that is, when the circumference of the piston 2ra is less than one-half wavelength 4/2, the impedance load122 Aco STIG COMPONENTS [Chap. 5 presented by the air on the vibrating piston is that of a mass shunted by a very large resistance. In other words J? = (Jt, + 2)* is large compared with w*M,*. In fact, this loading mass may be imagined to be a layer of air equal in area to the area of the piston and equal in thickness to about 0.85 times the radius, because Piston of adus @ (ra*)(0.85a)p9 = 2.67a"ps = Mus tong wte At high frequencies, ka > 5, the air Fic. 56. Piston vibrating in the load behaves exactly as though it were ‘end of long tube connected to one end of a tube of the same diameter as the piston, with the other end of the tube perfectly absorbing. As we saw in Eq, (2.60), the input mechanical resistance for such a tube is xa"pee. Hence, intuitively one might expect that at high frequencies the vibrating rigid piston beams the sound outward in lines perpendicular to the face of the piston. This 207 + a el rT o2}+ | 003 fe a IE oor 0.005, 0003} 0.002 0001] 0.0005} 00003} : 00002} —| 000: V1 s Iu rh oo “00a” or 03,” 10 30 ,, 0 Fic. $:7, Real and imaginary parts of the normalized mechanival impedance Za roo) of the air lou spon one ide of « plane piston of rai # monsted sn thi end Of a Tong tube, Frequrncy is plotted on ssl eal, where k= 2xfa/e = Braid Note also thr anlinate is equal to Zena/p, where Zain the acoustic ‘impedance, Part XI) RADIATION IMPEDANCES 123, inactually the case for thenear-feld. At‘a distance, however, the far-field radiation spreads, as we learned in the preceding chapter. 5.3. Plane Circular Piston in End of Long Tube.’ ‘The mechanical impedance (newton-seconds per meter) of the air load on one side of a plane piston mounted in the end of a long tube (see Fig. 5.6) and vibrating sinusoidally is given by a complicated mathematical expression that we shall not reproduce here. 209 100} xo EY ES 1] 5 F330] B29 Pe os a | oa 005 003 ce] oo oo “ors* o1 "03," 10 30° 10 Fio, 58. Real and imaginary parts of the normalized meshanieal mobility (wolycey) of the aic load upon one side of a plane piston of radius a mounted in the end of long tube. "Frequency i plotied on a normalized seal, where ka = 2efa/e ~ 2ya/h Note leo that the ordinate is equal to zapu/xa?, where ca isthe acoustic mobility Graphs of the real and imaginary parts of the normalized mechanical impedance 2 /xa"pi as a function of ka for a piston so mounted are shown, in Fig. 5.7. Similar graphs of the real and imaginary parts of the normalized mechanical mobility are shown in Fig, 58. The data of Fig. 5.7 are used in dealing with impedance analogies and those of Fig. 5.8 in dealing with mobility analogies. To a fair approximation, the radiation impedance for a piston in the end of a long tube may be represented over the entire frequency range by 1, Levine and J. Schwinger, On the Radiation of Sound from an Unflanged Cite lar Pipe, Phy, Re, TB: 385 06 (148)125 oq ‘so[uossuaup are sug}EU08 [Ty own 9 noys g'¢ pu zg sHLy Jo Si2NYo yoexD ay) 20 6g “Mla JO g-< BY puE ED > Dy 144 suOIde: oy) UL yuuproudde oxy (og PUB G9 Ha9MIIg HoHBLE aN iofojeuw Joy sas quowsy9 $0.18 9196) SUL | | : jones | ne waoajy = Om 1 Ssaoauodst z w< ot 2 : seesoro = ir | Soceioo = seoieet = 8H | ye ‘sey 3 y Caapeenne = pefewgy = 82 | aoforiey = 4 eoussteuode wt 2 i > ° a nay = nous o 2 dosp =n dap =» amen 2 eae ear ca aanereeel = tw | 19 < Oy g : a epesoro = Ye | Seeeetoo= "HN 281 = IE ‘oH, Sy ; / 88600 = agog't = Fa x ‘sourrasor ung 5 OMB RIB! | crsimocuoo = re | orteimuseo = 6 1% (sun SIS a % | go> oo oy = oy sso sno30ysay uae core sourpody anenooy ananoae aged z Joqny, 3007 Jo pay Ur UOIsIG sUeId B 70 OpIg IND ayy 40} AEGON puL souvpadw] UONMIPYY Z's TTAVL & posn 09 prnoys ys pus gg sSka Jo msEYD yonxD 99320 gg Bly Jo sas ayvuLOUddy ayy ‘0% PUY gO UooaIog UOHFOL O43 uy “seoquotruoup aiv ust TTY" < Of puwe'g > Dy aseyN UoHBe as WT SyNosT smoMowUe 30] soe YuOTIOP ANH og) BTL + a afb m= aaa = Os 1 sounds 9< oy 5 sore = a0 se19% = FI yy 0g e vepvoro = | omusteg = me ode 8 z | 0 > % 5 = oy = a ‘oy = a Woy =f aay g ere eae cae 90H = |__ — = $ [pepe = oy ma a Hy Bees my ty soumsI & | 9< 71 ° eS apoio =H sgt = "97 sqo19't = 841 yy ‘ey & wrmesyo = meee = asoes's = My ay ‘sous gongs 2 areesro = 7% | oreueey =m | 2/ynnLET & 5 25 Souris BUI & o> neg qos aunt rmtnan oBoqwey ous | uP lop =f sowrpedary anenoay ananoae agiands a Jourea yogoy oY Gow suNIg & Jo epg oUO 79} SNNGON pur suvpedwy ONPRN “Ts ATAVL iat126 ACOUSTIC COMPONENTS (Chop. 5 Part X11} RADIATION IMPEDANCES a: 2 | | E om ™ TE E s Bt f ss | br g i | & r q ke [| 4 4 i | i Hap HES ope oe Sh “HHA = Be Elias] 3/8 003] s|2| $2 3/48 Sls agave Ha 4 vt | att e/2 S ae Z aS "ioe tls Root 4 : A a alk 005 a oa i —|3 oma | & | 3 Tpedineantoay 3 ; | 3 oot LJ : ly nn xy iles| dz ales! als a aay efur| 22 |77} Fle aaa tah elas LJ Z)aa oo 0.0002}—}—+ a oa s ele é ge g al fou fou ful 4 $ | hy oor “ocs* of 03," 19 "39° 10 ard im 4a: Fra. 59, Real and imaginary parts of the normalized mechanical impedance (Ze/ & q . | y] eh a%pc) ofthe air load upon both sides oft plane eitular dink of radius aa tre space. Slgla.| 2. dle 2/28 Frequency is plotted on a normalised ttle, where a = 2efo/e = 2ea/s, Note He | 82} alia as the ordinate is equal to Zava"/e, where Za the normalized acoustic impedance, Bales) 8 4 (ee ales the same analogous ciouite used for the piston in an infnite baffle and Pile ae ee 2/82 shown in Fig. 5.5, where the elements now are a3 I Bs Rvs = wa?pye mks mechanical ohms (6.20) : | Sls Ru = Ruz + Rus = 4x(0.6133)'a'pe é = alae 4.73a%pe mks mechanical ohms (5.21) < sf 2) a8 Ruy = 0.604na%pc = 1.58a*oe mks mechanical ohms (6.22) 4 } ¢ 2 Ries 0.55 /apac? m/newton (623) Bl Gos a/28 0.6133natp = 1.9270 ke (528) Z| a 2 8 alte = \/ra'pec = 0.318/a"pye mks mechanical mohms 25) i as 3/3 gles 0.683/a%o¢ mks mechanical mohins (5.26) z aya ra Raz ™ pet/ra? = 0.318pqc/a" mks acoustic ohms (6.27) 2 3] E|38 Ram Rar+ Ray = (4)(0.6133)*p00/xa* wf2 £] flat = 0.479p4¢/a* mks acoustic ohms (5.28) Sea ad ehles 0.504p<¢/ra* = 0.1604p0¢/a* mks acoustic ohms (6.29) vgs ag ng t 5.d4a¥/pect m*/newton 630) ae" 8 al 8 0.1952p./0 ke/m* 631)128 ACOUSTIC COMPONENTS Chap. 5 Fax = na/poe mks acoustic mohms (5.32) 741 = 6.230"/pye mks acoustic mohms (6.33) In the frequeney ranges where ka < 0.5 and ka > 5, analogous eireuits of the type shown in Table 5.2 may be used. 6A. Plane Circular Free Disk.* A disk in free space without, surrounding structure is a suitable model, at low frequencies, for a direct radiator loudspeaker without a baflle of any sort. A simple equivalent vireuit, approximately valid for all frequencies like those shown in Fig. 5.5, cannot be drawn for this ease. At very low frequencies, however, it is possible to represent the impedance by a series eombination of a mass aud a froqueney-dependent resistance just as was done for the pistons in bales Graphs of the real and imaginary parts of the normalized mechanical- impedance load an both sides of the diaphraym, Zu/ra*py, 25 a function of ka for the free disk, are shown in Fig. 5.9. ‘The data of Vig, 5.9 are used in dealing with impedance analogies. ‘The complex mobility ean be obtained by taking the reciprocal of the eomplex impedance In the frequency ranges where ka < 0.5 and ka > 5, analogous circuits of the type shown in Table 5.3 may be ased pawr XLT Acoustic Elements 5.5. Acoustic Compliances. Closed Tube. In Ex. (2.50) we showed that a length of tube, rigidly closed on one end, with a radius in meters jer than 0.05 /4/J and less than 10/f hil an input acoustic impedance ba = RF cot ht (5.34) coustic impedance in mks acoustie ohms pevilie acoustic impedanee in wks rays a= rains of Lube in meters V = length of tube in meters For values of & that are not too large, the colungent may be replaced by the first two terms of its equivalent-series form pee ree 45 cot tt = fy — F - G&D (6.35) © FM, Wiener, On the Hola hotweon the Sawn Biekds adinted and Diffected hy Plone Obatseles, f. Aenat Se ster, 282 67 8 (51), Past NU ACOUSTLG ELEMENTS 129 Equation (6.34) becomes Z Joy tin Em. 5.36) Be Fv oat tH Sat where V = I'rat = volume of air in the tube. ‘The acoustic impedance Z, is.a series combination of an acoustic mass (I'ge/3ra") and an acoustic compliance (V /oee?). Equation (5.36) is valid within 5 per cent for ’s up to about X/7-+ If the impedance of the cavity is represented by the series combination of an acoustic mass and an acoustic compliance as is shown by Eq. (5.36), 8 lumped-cireuit representation is permissible out to dimensions of 2/8. In this ease the shape of the eavity is important because the magnitude of the inductance involves the ratio 1/xa*, that is, length divided by cross-sectional area. Such a series combination of mass and compliance is shown in Fig. 5.114. Closed Volume, If the second term of Eq. (5:36) has a magnitude that is small compared with that of the first term, 5 per cent perhaps that of the first term, we may neglect it. and (637) where a-2-% (638) is an acoustic compliance with the units of m'/newton, ‘The 5 per cent restriction given above means that [' should be less than (A/16), where is the wavelength in meters. Such a compliance is shown in Fig. 5.110. For quite small pistons, provided the largest dimension of the cavity does not exceed about one-eighth wavelength, the acoustic impedance presented to the piston is La 1 537 Sav ISPD G37) For the second term to be negligible within 5 per cent («7/V) should be greater than about 200. For intermediate sized pistons, the discussion on page 217 should be studied. ‘An acoustic compliance obtained by compressing air in ¢ closed volume is a two-terminal device as shown in Fig, 5.11b, but one terminal must ‘always be at “ground potential.” ‘That is to say, one terminal is the outside of the enclosure housing the cavity, and it is usually at atmospheric pressure. With such an arrangement for an acoustic compliance (see for examples) it is never possible to insert a compliance etsveen acoustic resistors oF acoustic masses as, for example, in the upper branches of Fig. 3.32 {In Ea, (5.42) we define the acoustic mass ofa short opea-cnded tube as equaling Ma = pu'/nat. Hence, loth approximation, the mass fr a cloard tube ionethind that for an open tube130 ACOUSTIC COMPONENTS: {Chap. 5 Stiffness-controlled Diaphragm. ‘To obtain a “series” type of acoustic compliance a diaphragm or stretched membrane must be used. Dia phragms and membranes resonate at various frequencies so that the fre- ‘quency range where they act. 25 compliances is restricted to that region well below the lowest frequency of resonance. A combination of series acoustic compliance, resistance, and mass is shown in Fig. 5.10. o—___] eontted whch ile com ‘taphragm pression occurs (@) @) Fin, 5.10, Beample of a inpiragm, ries acoustic compliance obtained with stifines-controlled o & ra at Cet (a) (0) Me ym $ Teh te o—____. © (a Foo. 5.11 Approximate analogous ciute for a short tube of oped dimer (a) and (b) Cievits used when px/Us ix very large (closed end). (e) and (d) Circuits Used when ps/Us is very small (open end). Cireuita (a) and (¢) yield the impedance within about 5 percent fora tube length Uthat ia lege than 9/8, Cireuits (0) and (a) yield the impedance within about § per eent for" < 3/16 Example 6.1. ‘The old-fashioned jug of Pig. 6.12 is ‘8 musical instrument. You are asked to analyze its performance acoustically. If ‘he inside dimensions of the jug are diameter = B in. and sireavity height = 10 in., ive the analogous eicuit, the clement sizes and the acoustic impedance for the air. cavity portion of the jug at 50, 100, and 300 eps. Assume T= 23°C and Py = 10" ewton/m*. (wors: The neck portion will be discussed later in this part.) Solution. The speed of sound at 23°C ie about 1133 ft/sec, Hence, used in w country dance band Past XIN) ACOUSTIC ELEMENTS. 13 ‘The length (ofthe jug is 0835 ft, Hence, oa dee Me osaa te ‘At 50 cpa whore /% = 363, tho eavity portion ofthe jug may he reprosnted yan Sountiecomplinee ¥ 82 x10 Pet Tac ow SxS ee = 5.89% 10-4 mt/newton ts 384 X 10¢ mks sewsstio ohms e100 ep whre/. = the ny pon he Me EemciysMwceatmeaaeee 4 pt oss x . Ma = sat ~ ie(O.10r6)* ~ 9 K8/m ca =i x iho mtnewun 425088 ~ agg) ~~ 10 AC 300 eps, where 17 the eavily portion of from Eq. (6.88, cared teeny | care eae ee ~~ o1016)7 °°" Hd earn 1/465, the acoustic impedance of Jug taunt be solved for directly | 5.6. Acoustic Mass (Inertance). A tube open at both ends and with rigid walls behaves as an acoustic massif it is short enough so that the air in it moves as a whole without appreciable compression. In setting. up the boundary condition, the assumption is made that the sound pressure at the open end opposite the source is nearly zero. ‘This assumption would be true if it were not for the radiation impedance of the open end, which acts very much like a piston radiating into open air. However, this radiation impedance is small for a tube of small diameter and acts ‘only to increase the apparent length of the tube slightly. ‘Therefore, the radiation impedance will be added as a correction factor later. Tube of Medium Diameter. In order to be able to neglect viscous lasses inside the tube, the radius of the tube a in meters must not be too small Also, in order to be able to neglect transverse resonances in the tube, the radius must not be too large. The equations which follow are valid for a radius in meters greater than about 0.05/-/f and less than about 10/f. Solution of the one-dimensional wave equation of Part II, with the boundary condition p = 0 at z = 1, that is, at the open end, yields tn = 5 tan Bt (639)132 ACOUSTIC COMPONENTS [Chap 5 where Zs = acoustic impedance in mks acoustic ohms of the open tube of Iength U ensity of the gas in kilograms per cubie meter speed of sound in meters per second radius of tube in meters vwave number = w/c in reciprocal meters length of tube in meters measured from the open end to the plane where Z, is being determined For small values of kv’, the tangent may be replaced by the first terms of its equivalent series form. v tan te a 4 AD 4 2 tone (6.40) Raquation (6.39) becomes 1) a jot 4 pl 2aQt) = ja 8 4 5 Se 4) {An equivalent circuit for Eq, (5:41) valid within about 5 per cent for the Frequency range where {"< /8 i given in Tig. S.11e.. fn that analogoun cenit, = (642) and Cy ingiven hy Eq, (638). The quantity M, isan acoustic mars with the units of kilograms per meter, For vanes of < 2/16, the second term of Fa, (641) wil be less than 5 per cont as large as the first, so that pal? 2, = jue, juM a (6.43) ‘The equivalent circuit for this impedance is given in Fig. 5.11d. End Corrections. BND CORRECTIONS IP THE END OF TURE. TERMINATES 1N INFINITE BAFFLE—FLANGED TUDE. Most acoustie masses terminate at cither one end or the other in open air or in a larger eavity. ‘This means that correetions must be added to the length I above if the value of Mis tobe correct when I’ is not large. The correction is the impedance given in Table 5.1 by the cireuits for ka < 0.5. For the case that a is less than 1/25, the circuits of Table 5.1 reduce simply to one element of magnitude May (see Bq. (5.17)]. Hence the end correction £” nevessary for each Daliled end of a vibrating “plug” of air in a tube is equal to y= Marat _ 02To. vat pe a pe meters (Any Past XU ACOUSTIC ELEMENTS 133 Hence, (5.42) becomes My = MELON ot erent 45) where the quantity 1 is by definition equal to the effective length of the tube in meters and is the sum of the actual length I plus the end correc: tion(s) £” (or 20”), ‘The numeral 2 in the parentheses will be used only if there are two free ends to the vibrating plug of air. If the tube is not round, we may replace a by W/S/r, where S is the cross sectional area of the tube. END CORRECTIONS IF END OF TURE 18 PREE (UNFLANGED TunE)® (See Fig, 5.6). In this ease, the correction at each free end of a tube is the impedance given in Table 5.2 by the circuits for ka <0.5. If a is less than 2/25, the circuit of Table 5.2 reduces simply to one element. of magnitude Mur as given by Eq. (5.31), Hence, the end correction 1” necessary for each free end of vibrating plug of air in a tube is equal to 70" oie m 6.40) me and Eq. (5.42) becomes a, = CONT _ pd wa? xa kg/m* 6.47) If the tube is not round we may replace a by ¥/S7x, where S is the cross-sectional arca of the tube. Cases of elements combining both acoustic mass and acoustic resistance will be discussed later in this pat, Example 6.2. ‘The jug of Example 5.1 has a neck with a diameter of 1 in. and a length of 15 in, (ee Fig, §.12). Atwhat frequency will th jug resonate’? Solution. irs, lot us assume that the frequency of resonance ix so Jow that the length of the neck Mis small compared with 2/16, Thea, because the ait in itis not conatraind, it will hewn acoustic mass, = 19 aoa | = 185 keynes h the neck of the jug i Uhe sume a that entering the air ‘wo clemwratarein seriow and will eyonate wt 0 XS 6.1, Acoustic Resistances. Any device in which the flow of gas oceurs in phase with and directly proportional to the applied pressure may be134 ACOUSTIC COMPONENTS chap. 5 represented as a pure acoustic resistance. In other words, there is no stored (reactive) energy associated with the fow. Pour principal forms of acoustic resistance are commonly employed in acoustic devices: fine meshed screens made of metal or cloth, small-bore tubes, narrow slits, and porous acoustical material. Screens are often used in acoustic transducers because of their low cast, ease of selection and control in manufacture, satisfactory stability, and relative freedom from inductive reactance. ‘Slits are often used where an adjustable resistance is desired. ‘This is accomplished by changing the width of the slit. Tubes have the disadvantage that unless their diameter is very small, which in turn results in a high resistance, there is usually appreciable inductive reactance associated with them. However, if a combination of resistance and induetanee is desired, they are useful Such combinations will be treated later in this part. Fibrous or porous Sja ——_=__, (a) Region in which viscous losses occur Fia. 5.13. Sketch showing the diminution of the amplitude of vibration of ait partiles in & sound wave near a surface. The letters (a) show the regions in which viscous Tosses occu, acoustic materials, porous ceramics, and sintered metals are often tsed in industrial applications and are mixtures of mast and resistance. Tn alll four forms of acoustic resistance, the frictional effects producing the resistance oceur in the same manner. In Fig. 5.13, we see the opposite sides 1 and 2 of a slit, or tube, or of one mesh of sereen, An alternating pressure difference (ps — py) causes @ motion of the air molecules in the space between the sides I and 2. At. 1 and 2 the air particles in contact with the sides must remain at rest. Halfway between the sides, the maximum amplitude of motion will occur. Frictional losses occur in a gas whenever adjacent layers of molecules move over each other with different velocities. Hence frictional losses occur in the example of Fig. 5.13 near each of the walls as marked by the letter (a). In any tube, slit, mesh, or interstice the losses become appreciable when the regions in which adjacent layers differ in velocity extend over the entire space. Screens. The specific acoustic resistances of a variety of sereen sizes commonly manufactured in the United States are shown in Table 5.4 Part XU Acoustic MENTS, 135 ‘The acoustic resistance is obtained by dividing the values of Z, in this table by the area of sereen being considered, The acoustic resistances of screens are generally determined by test and not by ealeulations, TABLE 5.4. Specific Acoustic Resistance of Single Layers of Screens Mee | Approx. diam of | Re, rayls, | Zt mks rays, Tacarsn | Wire, em | dyne-sec/em*| newton-see/m* 30 0.033, 5.87 50 0.022 5.88 100 o.011s 8.10 120 0.0092 | 13.5 200 0.0057 24.6 Tube of Very Small Diameter’ (Radius a(meters) < 0.002/s/F]. The acoustic impedance of a tube of very small diameter, neglecting the end corrections, is 24088 + 74 Mw — nowton-sec/m* (6.48) Ma = 2% keg/mé 6.49) where » = viscosity coefficient. For air = 1.86 X 10-' newton-seo/m* at 20°C and 0.76 m Hg. This quantity varies with temperature, that is, « T°, where 7 is in degrees Kelvin, 1 = Iength of tube in meters. a = radius of tube in meters. M, = acoustic mass of air in tube in kilograms per meter‘, po = density of gas in kilograms per cubie meter, ‘The mechanical impedance of a very small tube is found by multiplying Eq, (5.48) by (ra")*, which yields Bu = Seal + 36M (6.50) where Mu = pyra'l = mass of air in the tube in kilograms. Narrow Slit (U(meters) < 0,008/-/7]. The acoustic impedance of a very narrow slit, neglecting end corrections, is 12nl 4 Gpole Pw 1) Sut newton-see/m* 51) > Lord Rayleigh, “Theory of Sound,” 24 ed, Vol. If, pp. 323-328, Macmillan & Co., Ltd., London, 1929. Also see LL. Beranek, "Acoustic Measurements,” pp, 187-185, John Wiley & Sons, Inc., New York, 1949, 1. B. Crandall, “Vibrating Syatems and Sound,” D. Van Nostrand Company, Ine, Now York, 1926,136 ACOUSTIC COMPONENTS lc ol where [ = Jength of slit in meters in direction in which the sound wave is traveling (see Fig. 5.14) w = width of slit in meters as viewed from the direction from whieh the wave is coming (see Fig. 5.14) 1 = thickness of slit in meters (see Fig. 5.14) The mechanical impedance of a very narrow slit is given by multiplying, iq. (BBL) by tot 7 Biscton ot where Muy = palit = mass of air in sahdre the slit in kilograms bil Brample 63. An aewati rsstans Geto prea creer nee (ol Jeena Siew decal ran ing clement in xn earphone. Select & srereen and te diameter of hole necessary bo achieve Solution. As the resistance in needed for an earphon I we select « 20-mnesh seroen (wee Table 8.4), the seo 2G rayls. For | acoustic ohn, an area is octet of should be quite oval resinlanen for unit area is 1 . Sm hg = 0.400 ‘The diameter dof the hole requived for this agen i 5.8. Cavity with Holes on Opposite Sides Mixed Mass-Compliance Element. A special case of an element that is frequently encountered in Pre, 5.15, Keample ol wss-romplianre element made front a envity with i's on appetites acoustival devices and that has often led to eonfasion in analysis is that shown in Fig. 5.15. Emagine this to be a doughnut-shapied element, cach Hf which has a hole of radius @ bored in it, When a flow of air with volume velocity U, enters opening 1, all the air partirles in the vicinity of Part X 1 ACOUSTIC ELEMENTS 137 the opening will move with a volume velocity U,. Part of this velocity goes to ‘compress the arin the eyindrcal space 3, and part oft appears asa movement of air that isnot appreciably compressed, It was pointed out earlicr that a portion ofa gas that compresses without appreciable motion of the particles iso be treated as an acoustic compliance. By inspection of Fig. 5.15 we ace that the portion enclosed approxic mately by the dotted lines moves without appreciable compression and, hence, is an acoustic mass. That lying outside the dotted lines is an acoustic compliance. ‘The analogous cicuit for this acoustic devie is fiven in Fig. 5.16, ‘The volume velocity U; entering opening 1 divides into two parts, one to compress the air (U;} and the other (U/;) to leave opening = 2. By udiciousestimation wearriveat values for My. If the length | of the cylinder is fairly long and the volume Ca 3s large, Ma is merely the end correo- 0.01 /fand a < 10/f]. ‘The acoustic impedance for a tube with a radius @ (in meters) that is less than 0.002/+/f was given by Bags. (6.48) and (6.40). Here we shall give the acoustic impedance for a tube whose radius (in meters) is greater than 0.01/+/f but still less than 10//. For a tube shove radius tes between 0002/7 and 001//7 interpolation must be used, ‘The acoustic impedance of the inter mediatesized Lube is equal to My Me ty Us Za = Rat joMs (6.53) where Ba = Jew vion [+ 2)] mks acoustic ohms (59 My = SEO g/m (6.55) a = radius of tube in meters, x ~ density of air in kilograms per eubie meter. kinematic coefficient of viscosity. For air at 20°C and 0.76 m Hg, = 1.56 X 10" m*/see. This quantity varies about as 7"7/Ps, where 7 is in degrees Kelvin and Py is atmospheric pressure, actual length of the tube. t + U. Inpand, Scattering and Absorption by Acoustic Resonators,” doctoral disserte- tion, Massachusetts Institute of Technology, 1950, and J. Acouat. Sac. Amer, 26: Las 145 (1955),138 Acous 1c COMPONENTS [Chap 5 P'" = end correction for the tube. It is given by Bq. (5.44) if the tube is flanged or Eq. (5.46) if the tube is unfianged. ‘The numbers (2) in parentheses in Eqs. (5.54) and (5.55) must be used if both ends of ‘the tube are being considered. If only one end is being considered, replace the number (2) with the number 1 = angular frequency in radians per second. 5.10. Perforated Sheet--Mixed Mass-Resistance Element? (a(meters) > 0.01/\/f and a < 10/f]. Many times, in acousties, perforated shects are used as mixed acoustical elements. "We shall assume a perforated sheet with the dimensions shown in Fig. 5.17 and holes whose centers are spaced more than one diameter apart. For this case the acoustic impedance for each area 6%, that is, each hole, is given by Za = Rat joMa where a= sho vin [! sai “| mks acoustic ohms (5.56) wen Bfesine(i-d)] ae ee ery asta alae eee rite clele tole Ges het te mb Lean "imran ment S|e|e resonant circuit, we are often inter- qT] ested in the ratio of the angular fre- i @|e quency of resonance wy to the angu- : HL Er tmalviddhs gadiae nee Is a gg. os) mee the baler iain ma nd ath macs Ph cd ea meet the eet ee @ weMa foe + 1.7al1 — (a/b)] = Bee NE + alt — Tea'/i)) (5.58) The Qu is independent of the number of holes in the perforated sheet. Pact XIII} ACOUSTIC ELEMENTS. 139 ‘We repeat that: these formulas are limited to eases where the centers of the holes are spaced more than a diameter apart. 5.11, Acoustic Transformers. As for the other acoustical elements, there is no configuration of materials that will act as a “lumped” trans former over a wide frequency range. Also, what may appear to be an acoustic transformer when impedances are written as mechanical impedances may not appear to be one when written as acoustic impedances, and vice versa. As an example of this situation, let us investigate the case of simple discontinuity in a pipe carrying an acoustie wave. Junction of Two Pipes of Different Areas. 4 junction of two pipes of different areas is equivalent to a discontinuity in the area of a single pipe (Gee Fig. 5.18a). If we assume that the diameter of the larger pipe is less poe nit fs 7 (a) (b) (e) Fro. 5.18. (a) Simple discontinuity between two pipes. (@) Acoustic-impedance transducer representation of (a); beeause the Lnsformation ratio ia unity, no trane= former is required. (@) Mechanieal-impedance transducer representation of (a) than 0/16, then we may write the following two equations relating the pressure and volume velocities at the junction Pap (6.59) =U: (5.60) Equation (5.59) says that the sound pressure on both sides of the junction is the same. Equation (5.60) says that the volume of air leaving one pipe in an interval of time equals that entering the other pipe in the same interval of time. ‘The transformation ratio for acoustic impedances is unity so that no transformer is needed, For the case of a circuit using lumped mechanical elements the discon tinuity appears to be a transformer with a turns ratio of S1:S; because, irom Eq. (5.59), -f S78 (5.61) and, from Eq. (5.60), Ss = uaSs (5.62) where f; and f; are the forces on the two sides of the junction and wand wz are the average particle velocities over the areas $, and S;. We have (5.63)and tn B(YE(Yemn A transformer is needed in this ease and is drawn as shown in Fig, 5.186. It must be noted that a reflected wave will be sent back toward the source by the simple discontinuity. We saw in Part IV that, in order that there be no reflected wave, the specific acoustic impedance in the second tube (ps/us) must equal that in the first tube (pi/u:). This is possible only if S: = So, that is, if there is no discontinuity. ‘To find the magnitude and phase of the reflected wave in the first tube resulting from the discontinuity, we shall use material from Part 1V. Assume that the discontinuity exists at x= 0. The specific acoustic impedance in the first tube is Bm = Bt (6.66) If the second tube is infinitely long, the specific acoustic impedance for it at the junetion will he ee sn [sce Eq. (2.60)]. ‘The impedance Zs at the junction is, from Eqs. (5.59 and (5.64), Zn Ss From Eqs. (5.67) and (5.68), Si 2a, = Spe (6.69) Using Eqs. (237) and (2.58), setting z = 0, we may solve for the rms reflected wave p- in terms of the incident wave py Pimms 6.70) t : ta = 55 ~ p) (6.71) a 67) Pp 623) ‘The sound pressure py of the transmitted wave in the second tube at the Part XIII] acoustic ENTS 11 junction point must equal the sound pressure in the first tube at that point, Pr=P +P (6.74) so that Bee gle (6.75) If 8; equals 8, there is no reflected wave p. and then py = pr. Note also that if S; becomes vanishingly small, this case corresponds to the ease of a rigid termination at the junction, For this ease, P= Mm (6.76) and ry + p= 2 6.7) ‘This equation illustrates the often-mentioned ease of pressure doubling. ‘hat is to say, when a plane sound wave reflects from a plane rigid surface, the sound pressure at the surface is double that of the incident wave. ‘Two Pipes of Different Areas Joined by an Exponential Connector.» An exponential connector may be used to join two pipes of different areas. Such a connector (see Fig. 5.19) acts as a simple discontinuity when its ev mi (6) i a amend a GR (a) te) Fig, $19. (@) Exponential connector between te pipes (b) high-frequency repre: sentation of (@) ‘ising neousticimpetanre trimsideee; (2) highrequency. repre Sentition of (a) using nicchanieat-impedance trance. length is short compared with a wavelength and as a transformer for acoustic impedances when its length is greater than a half wavelength, AF the sevond tube is infinitely long, then at 2 = 1 (see Vig. 3.19), PF = pw (5.78) If the cross-sectional area of the exponential connector is given by S@) = Se (6.79) "LF, Olon, “Hlements of Acoustical Kaginvering,” 2 ed, yp. IDLH, D. Va nk, 17 Nostrand, Campoany, Ine, New142 ACOUSTIC COMPONENTS (Chap. s and the length of the connector is J, then the specific acoustic impedance atx =Ois a Pe ppg S08 (0+ 8) + J sin (01) Joy, PM con (Gl = 8) +f in (Hl) Con where b= 34 VUat/e!) =m? in meters 1m = flare constant in meters [see Eq. (5.79)] @ = tan™* (m/28) = speed of sound in meters per second 1 = length of the exponential connector in meters pa = density of air in kilograms per cubic meter At low frequencies (b imaginary and 4/1 large) Pi _ Pe ee B-fot tu= Za 81) at high frequencies (b real and 1/A > 1) Pipe . ta = Sz, 3) as = Ban 6.83) At intermediate frequencies the transformer introduces a phase shift, and the transformation ratio varies between the limits set by the two equations above. The transformation ratio for acoustic impedance at high frequen: seen from Hq. (5.83) to be V/Sz/5; (see Fig. 5.19). ‘That is to say, t= (yen et For mechanical impedance at high frequencies the transformation ratio is seen from Bq. (5.83) to be V/S\/S; (see Fig. 5.19). That is to say, an = (ff) Example 64. Ttis desired to reaonate a bansrelex loudspeaker box to 100 eps by Grilling number of holes i it.The hox has volume of 1 (and a wall thickness of 34in. Determine the size and numberof holes needed, assuming a Qa = 4nd a ratio of hole diameter to on-center spacing of 0.5 Saluion. From Iq. (6.58) we se that, approximately, a aves (100395) V3 ‘a = (010.0005) V2 Vox8 The diameter of the hole in inches is 26/0,0254 = 040708 in = 0.000893 m Part XIII} ACOUSTIC ELEMENTS 43 1 at 7 GBT 7 GR ICHaaRD 7 7880 mks acoustic oheme ‘The desired acoustic mass of the holes is oe 7 0 If there aren holes, the acoustic mass for each hole equala Ms = (12.55) kg /tat From Eq. (6.57), 118, (12.55) = Zegigegay (0.0125 + (0.00152)0.78) = —1.18)0,0137) 255) X 1} + 510 holes Example 6.6. Design a single-section T constant low-pass wave filter with « ‘cutof frequency of 100 cps and a design impedance of 10*mks neoustic ohm Solution. single T section of this type is shown in Figs. 5.15 and 5.16 except that the acoustic masses for one section have the values 4/2, because each Af has to serve as the acoustic mass both for the section in question and the adjacent section, By definition, the cutoff frequency is he ‘The design impedance is equal to (is Vee From these two equations we ean solve for Ca and Mas = sccig 7 8 x1 Ma = B18 X 10-€ X 108 = 5.18 kg/m* From Par. 58 and Fa. (6.55), with equal to zero, we get the size of the hole in the device of Fig. 6.15, won 0.20.18) _ 545 02m ‘The diameter ofthe hole it 0.4m, The volume of the cavity ie V = CyPy = 14 x 318 x 10 108 0445 mt ‘The elements for the Tsection are thereby determined.CHAPTER 6 MICROPHONES part XIV General Characteristics of Microphones Microphones are electroacoustic transducers for converting acoustic energy into electric energy. They serve two principal purposes. First, they ate used for converting music or speech into electric signals which are transmitted or processed in some manner and then reproduced Second, they serve as measuring instruments, converting acoustic signals into electric currents which actu- ‘Small acta ate indicating meters. In some pola cmd cavity spplications like the telephone, high electrical output, low cost, and durability are greater consid- re erations than fidelity of reproduc- vastreen—“] H—Metanea!,, tion. In other applications, small Geiestiche size and high fidelity are of grester fia. 64 Sketch ate prammeactunted importance than high sensitivity cops copseting oft ri enor and low cost. In mesmurement in one de of which there is « flexible applications we may be intereste diaphragm connected to tranaducing in determining the sound pressure or the particle velocity. Tn some applications the mierophone must operate without appreciable changeit characteristics regardless of major changes in temperature and barometric pressure For these different applications, a variety of microphones have been developed. For purposes of discussion in this part they are divided into three broad classes in each of which there are a number of alternative constructions. ‘The classes are 1. Pressure microphones 2. Pressure-gradient microphones. 3. Combinations of (1) and (2). In this part we shall describe the distinguishing characteristics of these we Pat NIV) CHARACTERISTICS OF MICROPHONES 14s three types. In the next two parts we shall discuss in detail several examples of cach type involving electromagnetic, electrostatic, and piezoelectric types of transduction, G.1. Pressure Microphones. \ pressure microphone is one that responds to changes in sound pressure. A common example of a pressure microphone is one with @ diaphragm, the back side of which is terminated ene™]] in closed cavity (sce Fig, 6.1). A Loy in tiny hole through the wall of the eav- ene ity keeps the average pressure inside of theeavity at atmospheric pressure. However, rapid changes in pressure, such as those produced by a sound wave, cause the diaphragm to move backward and forward, If. pressure microphone is placed ® in a small cavity in which the pres sure is vatied, as shown in Fig, 6.2, the output voltage will be the same regardless of what. position the microphone occupies in the eavity. On the other hand, if a pressure microphone is placed at successive points 1, 2, 8, and 4 of Fig. 6.3a, it will respond differently at each of these points for reasons that ean be seen from Fra. 6.2, Sketeh of «pressure chamber barpastan Cav connect = pi bz os os. ‘acoustic mass, ae oe Sivoo fa (a) _& faerie gon ‘rs fe fos Te fs () Pia, 63, Sketch of an arrangement in which a pressure gradient is produced, Fig. 6.38. The pressure drops ps, Ps, Pa, and px are different from each other by an amount Ap, if the spacings Ax are alike If a pressure microphone is placed in a plune sound wave of constant intensity 1 (watts flowing through a unit area in the plane of the wave front), the force acting to move the diaphragm will he independent of fre quency because p = ~/Ipec {see Eq. (2.88).146 MICHOPHONES (Chap. 6 6.2. Pressure-gradient Microphones. A pressure-gradient micro phone is one that. responds to a difference in pressitre at two closely spaced points. A common example of this type of microphone has & diaphragm, both sides of which arv exposed to the sound wave. Such a at construction is shown in Fig. 64 If a prossure-gradient- micro- moat ost Dressure-gra vo. EY phone is plied in the cavity of Le Fig. 6.2a, there will be no net force ing on the diaphragm and its output will be sero, ‘This happens because there is no pressure sworn] \ wean crnain Raton in the eaty, Tone roa, ‘ewetstuine trast, if a prosiuregrasont mivto- ZEA, Mosgdwdoge phone is placed at the successive tient Positions J to of Big. 63a, it will hie produce an output voltage propor ected tional to the pressure gradient, 4p/At. In other words, if Ax is the same between successive points, the microphone output will be inde= pendent of whichever of the four positions it aveupies in Fig, 6.3. Ifa very small pressure-geadient microphone is placed in a plane sound wave luaveling in Ue x direction, he complex rms toree J, acting to move Pua, 6.4, Sketch of a pressure microphone consisting ‘of Higphragi, both aides exposed, toa Uanaueing element Principal ass ot microphone ‘Drecton of rave the sound wave Fic, 6.5. Prosure-grudient mierophone with principal axis loratedl at an angle 6 with respert to the dieetion of travel of the sound wave. the inp i fo= 8% ateoss wn mre p = ms mind prose he tag 0 tye the not de aphragm mas with de dnon A = chive dodance beeen th oss ofthe diaphragm Mees) ‘The equation for a plane traveling sound wave has already been given (Eq, (2.57); it is p= pe 62 Pert XIV] CHARACTERISTICS OF MICROPHONES, uz where k = w/e Pe = rms pressure at If we assume that the introduction of the microphone into the sound field does not affect. the pressure gradient, we may substitute Eq. (6.2) into Eq, (6.1) and get em (6.3) ‘The magnitude of the force at any point 2 is Ufo) = Plo Al cos @ It should be remembered [see Eq. (2.4)] that in the steady state the pressure gradient is proportional to joe times the component of particle velocity in the direction the gradient is being taken. The force fo is, therefore, proportional to the particle velocity at any given frequency. Reference to Fig. 6.5 is sufficient to convince one that when @ = 90°, the force acting on the diaphragm will be zero, because conditions of symmetry require that the pressure be the same on both sides of the diaphragm. From Eq. (6.4) we also see that the effective force acting on the diaphragm is proportional to frequency and to the rms sound pressure. In spherical coordinates, for a microphone whose dimer compared with r, Eq, (6.1) becomes 6.4) fo= SP atcoso (6.5) ‘The equation for a spherical wave is found from Eq. (2.62). Ae ge fee: 66) Substituting (6.6) in (6.5) gives fo = ALI eaes(at cos 6) (62) This vielde Wola = (6.8) However, we see from Eq. (2.00) that in a plane wave the rms velocity is, related to the rms pressure by 69) 6.10)148 EROPHON Chap. where |uj is the tims partivle velocity in the direction of travel of the sound wave. Hence, Has, (6.1) and (6.8) become Mo In other words, the effective (rms) foree fn acting on the diaphragm of pressure-gradient microphone is dircetly. proportional to the effective particle velocity in the direction of propagation of the wave, to the frequeney, to the density of the air, to the size and area of the diaphragm, and to the angle it makes with the di wave. ‘Thisstatement is true for any type of wave front, plan eytindrical, or other, provided the microphone is so small that its presence does not disturb the sound wave. At any given frequeney, the response of the microphone is proportional to the cos @, which yields the direetiity pattern shown in Fig. 6.6a. ‘This ulopeS Al cos @ an) Beccton (a) () Fie, 6.6, Directivity characteristic of the prossuregrailiont microphone of Big. 64 shape of plot is commonly referred to as a “figure 8" pattern. ‘The same Pattern, plotted in decibels relative to the forve at: @ = 0, is given in (b) It is interesting to observe that the pattern is the same as that for an acoustic doublet or for an unflanged diaphragm at. low frequencies (see Figs. 4.8 and 4.13) ‘The frequency response of a pressure-gradient (particle-velocity) microphone, when placed in a spherical wave, is a funetion of the curva ture of the wave front, ‘That is to say, from Eq. (6.10) we see that for values of A'r* (kr equals wr/e) large compared with 1 the particle velocity is linearly related to the sound pressure. A large value of kr ‘means that either the frequency is high or t of curvature of the front is large. However, for values of 43! small compared with t, which means hat the radius of curvature is small oF the frequency is low, or hoth, the particle velucity is proportional to |pl/(er). As a result whew a perion talking or singing moves near to a pressure-gradient micros phone so that r ix small his voiee seems to become more “boomy” oF radi Part XIV] CHAMACTRIISTICS OF MICRO “heassy” been frequeney JONES. M9 the output of the microphone increases with decreasing 6.3. Combination Pressure and Pressure-gradient Microphones. A combination pressure and pressure-gradient microphone is one that responds to both the pressure and the pressuregradientinawave. A common example of such # micro phone is one having a cavity at the back side of the diaphragm that, hhas an opening to the outside air containing an acoustic resistance (vee Fig, 6.7a). ‘The analogous circuit for this Ueviee is shown in Fig. 6.76, If we let Pim por (6.12) pe mt OE atcane on(i-s Tet us say that Us is the rms vol- ‘ume velocity of the diaphragm; U's is the rms volume velocity of the air passing through the resistance; Po is the net rms pressure acting to movethe diaphragm; and Zayis, thediaphragm impedasice, ‘Then stews) cm =) Diaphragm: xe Mechanica: A teousie ce Cnet to Stance Fy SMonadocing Serpent which i eaey produces an Ettpu vatoge (a) (oy Fie. 6.7, (@) Sketch of @ combination pressure and pressuregradient microphone consisting of right enclosure in ‘ne side of whieh is « movable diaphragm fonneeted to a tranadieing element. and Jn-another side of which is an opening With nn acoustic “resistance Ra. (0) ‘Acoustiorimpedancs circuit for (0). wwe can write the following equations from Fig. 6.7b: ty (e» + ~ jatla Jot te Jota" ‘The pressure difference across the diaphragm is Pom Unkan = ‘Substitution of (6.13) in (6.15) yields Lav (m + Bil > ia ¥ Taoh all +0e(Ret gd) Aree 0001 | tice eect camping ph Nesp assem Camping 1 e) F1o. 6.18, (a) Schomatie representation of a microphone with a slotted back plate, () Crossvectionel drawing of the W.E. G40-AA capacitor microphone. The slotted back plate serves both a the recond terminal of the condenser and az a means for damping the principal resonant mode of the diaphragm. (c) Cross-sectional drawing fof the Altec 21-C eapacitar microphone. The cap with holes init serves both for protection and as am acoustic network at high frequencies. (From Beranck, “Acoustic Measurement," John Waley & Sons, Inc, New York, 1949, and courlety of Allee Lanwing Corporation) Pact XV] PRESSURE MICROPHONES 159 tion in construction of the 610-AA from the 21-Cis that the back plate in the former has several annular slots cut in it. These slots form an acoustic resistance that serves to damp the diaphragm at resonance One manner in which the microphones are operated is shown in Fig. 6.17b ‘The resistance Rs is very large. The direct voltage E is several hundred volts and acts to polarize the microphone. Electromechanical Relations. Electrically, the electrostatic microphone. is a eapacitor with a capacitance that varies with time so that the total charge Q() is QW = G+ a0 = CaOIE + eC] (6.20) where qois the quiescent charge in coulombs, q() is the incremental charge in coulombs, Cx(0)is the capacitance in farads, His the quiescent polarizing voltage in volts, and e() is the incremental voltage in volts. The capacitance Cy(0) in farads is equal to (see Fig. 6.172) So Sfi4@ ae — 2) x +c, [14+ 29] where Cy is the capacitance in farads for x(t) = 0 and Ca(t) is the incremental capacitance in farads, « is a factor of proportionality that for air equals 8.85 X 10°, $ is the effective area of one of the plates in square meters, ze is the quiescent separation in meters, and 2(0) is the incremental separation in meters. It is assumed in writing the right- hand term of (6.21) that the square of the maximum value of x(¢) is small compared with x0! If we similarly assumed that [e(@)]2.. < B*, then (6.20) and (6.21) yield Call) = Cay + Corl) e+ 010 = Cob + Cook [2 + 20] 622) so that a = Cale + E20] 6.28) The total stored potential energy W(t) at any instant is equal to the sum of the stored electrical and mechanical energies, 14Q(0)*/Ca(®) plus 342(0)?/Cus, where Cus is the mechanical compliance of the moving plate jn meters per newton. ‘That is, 1 (get aC, 12H" 2Cy 2 Cua 1 get + 2na(0] Fen [1+ 22] =} tes 2001-2] 4522 @2n wos160 MICROPHONES, Chap. 6 ‘The force at any instant in newtons acting to move the plate is, from the equation for work, dW” = fdr, aw fet sy =O (625) so that, by diferentition of Ba. (6:24), So IO) + — shia tan + 24601 + 2e ot fate . Hence, because B= g4/Ca, 2 _ Ed so = 72 - Fat (627) Rearranging Ba, (6.28) gives ay « — 20. (628) In the steady state facu 6.29) where g, §, x, and u are now taken to be complex rms quantities; so Eqs. (6.27) and (6.28) become 6.30) (a1) with ¢ and f also being complex rms quantities. Analogous Circuits. Equations (6.30) and (6.31) may be represented by either of the networks shown in Fig. 6.19, where Coy = must ECs!) _ Caste Oro + Cust ECust) E*CwsCro + 20 Cx ~ FA F168) Ces Cus = CECT FI (633) Note in particular that Crusis the mechanical compliance measured with the electric current i = 0; Cre is the electric capacitance measured with the force f equal to zero; Cy is the electric capacitance measured with the velocity 1 equal to zero; and Cy, is the mechanical compliance measured with the voltage ¢ = 0. These circuits were first shown as Fig. 3.37, Part XV} PRE 161 SURE MICKOPHONE and the element sizes were given in Eqs. (3.34) to (3.37). In practice, the ircuit of Fig. 6.196 is ordinarily used for electrostatic microphones. ‘When one of the microphones shown in Fig. 6.18 is radiating sound into air, the force built up at the face of the microphone when a voltage is applied to electrical terminals (3-4 of Fig, 6.19) is very small. Hence, : Fenn ‘ A a ae fe te ; (a) a ae " >t —o3 ft Cues fo 7 fe 2 @ lectromechanical analogous cireuts for electrosta phones (snpedance avalon) when an electrie-impedance bridge is used to measure the espacitance of the microphone, the eapacitance obtained is approximately equal to Cr. By Thévenin’s theorem, the eapacitor microphone in a free field can be represented by Fig, 6.20. |‘The quantity ¢9 is the rms open-cireuit voltage produced at the terminals of the microphone by the sound wave and equals {from Eq. (6.31) and tt Fig. 6.19} oS = cow MH 2 SeSPB sy) wee * Fro. 6.20, Thévenin’s where the force fs, acting on the microphone with circuit of » capacitor the diaphragm blocked so that u = 0, is equal to tcrorhonef she ype the blocked pressure ps times the area of the dia- Siusted in free space. phragm 8. “Acoustical Relations. The microphones of Fig. 6.18 have diaphragms swith the property of mass Mu in addition to the mechanical compliance Cys assumed so far. For the 640-AA microphone, the internal acoustical cireuit consists of an air space directly behind the diaphragm with an acoustic compliance, Cai; a back plate and slots with acoustic resistance ‘and mass, Rx and Mas; and an air cavity around and behind the plate with an acoustic stiffness, Cas. ‘The radiation impedance looking outward from the front side of the diaphragm is jwM a4, where Mua is found from Fy. (5.31). The complete acoustical and mechanical circuit in the162 MICKOPHONES (Chap. 6 impedanee-type analogy is seen in Fig, 6.21. In this circuit pe equals the ms pressure at the diaphragm when it is restrained from moving, Mao, = Muro/S* = acoustic mass of the diaphragm, S = effective area of the diaphragm, and Up = Sup = rms volume velocity of the diaphragm, All units are of the mks system. When Fig. 6.19 is combined with Fig. 6.21, the complete circuit for the electrostatic microphone shown in Fig. 6.22 is obtained. Fig. 6.21. Acoustical circuit of a capacitor microphone including the radition mass and the acoustical elements behind the disphragm (impedance analogy), Performance. The performance of the capacitor microphone shown in Fig. 6.18, viz., the Western Hlectrie type 640-AA, can best be understood. by reference to Figs. 6.23 and 6.24, which are derived from Fig. 6.22. At low frequencies the circuit is essentially that of Fig, 6.23a. From this cirouit, the open-circuit voltage ¢, is equal to (6.35) At low frequencies, therefore, eis independent of frequency. ‘This is the frequency region shown as (a) in Fig, 6.24. 12% Y> 08 eH Bh fe Ca Ms Bag Fis; 6.22. Complete electroncoustieal eireut of a eapacitor microphone (impedance analogy), In the vicinity of the first. major resonance, the circuit becomes that of Fig. 6.236. At resonance, the volume velocity through the compliance Cas is limited only by the magnitude of the acoustic resistance Ras. In general, this resistance is chosen to be large enough so that the resonance peak is only about 2 db (26 per cent) higher than the response at lower frequencies. The response near resonance is shown at (8) in Fig. 6.24. Part XVI PRESSURE MICROPHONE: 16s Blas a te Io ce TM SE unc (Ota Mas a Sw Pal Cas Near frst AS a Ge La mee) fe" ae Ma Mao Mas sas i es ‘owe sceod Sia > aby Fic, 6.23. Capacitor microphone. Simplified. circuits for six frequency” regions {inpedance auuiney). ‘The exams ponte producal by te round wave of the diaphragm of the microphone with the microphone belt motionless ix'po, aad the opensicult vollage ice Above the resonance frequeney, the circuit becomes that of Fig. 6.23, The volume velocity is controlled entirely by the mass reactance, Hence, _ — ee (6.36) © BO aa Mao + MasleoS (6.30) In this frequency region the response decreases at the rate of 12 db per octave {ave region (c) of Fig, 6.24 h i mates, and the response. At a high frequency, the parallel circuit antiresonates, and the drops very low. This condition can be seen by reference to Fig. 6.23¢. Just above the point of antiresonance, a resonance occurs as shown in (¢).164 Mickopnos The magnitude of the resonance peak ix here limited by the radiation resistance Kas, Which is no longer negligible compared with joM 44 [see (©) of Fig. 6.24). Finally, above this resonance frequency, the open ireuit voltage again drops off at the rate of 12 db per ovtave [see (J)]. 6.6, Piezoelectric Microphones. Piczoclectrie microphones cmploy crystals or dielectrics which, when acted on hy suitable forces, produce electrical potentials linearly related to the deformation of the substance. g fa) Of i 7 | Eb So Dl he E oo ale ar a \ 19 00 TF Fonoon Frequency in eyes per second Fic. 6.24. Ope voltage response characteristic of capacitor mirvuphone fof the type shown in Fig. 618, "Norinally ewes taken a 1 vole, \ Y 2X°eat shear pate Fic, 6.25. Typical form of a large Rochelle salt erystal. The coordinate axes anid the way in which an X-eut shear plate i¢ ext from theerystal ar inieated (Conrteny of Brush Blectronics Company.) Piezoelectric substances have been used extensively in microphones Decause of their low cost and ruggedness. Four common types of piezoelectric substances are diseussed in this chapter, viz, Rochelle salt, ammonium dihydrogen phosphate (ADP) and lithium sulfate erystals,f and barium titanate ceramic plates 1 ADP erystals are sold unde the trade name Lif th trade name PN, and i n sulfate erystats ter Part XV} PRESS URE MICROPHONES 165, Crystal Microphones. Crystal microphones are used primarily in public-address systems, sound-level meters, and hearing ‘aids. ‘They have sutisfuctory frequency response for these applications and are high in sensitivity and low in cost. A diaphragm type ean be purchased at low cost to cover the frequeney range of 20 to 8000 eps sith maximum variation of 6 db from the average. Sound levels as low as 20 db and as high as 160 db re 0.0002 microbar may be measured. ‘The electrical impedance is that of a pure capacitance, which for some crystals varies with temperature, as we shall see later. “The mechanical impedance for microphones using a diaphragm is not usually high enough | _.y hs Zea ‘shea pte Fic. 6.26. Typical form of large Pra, 6.27. Typical form of a large lithium ammonium ‘dihydrogen phosphate ulfave (Lil) erystal. (Courtesy of Brush Blec- {ADP or PN) crystal. (Courtesy of tronics Company.) Brush Blectronie Company.) so that the microphone may be used in a closed cavity without seriously changing the sound pressure therein. ‘Typical forms of the whole erystals of Rochelle salt, ADP, and lithium sulfate are shown in Figs. 6.25 to 6.27. Transducing elements are obtained by cutting slabs of material from these whole erystals. Usually these slabs are thin and either square or rectangular in shape. If the X axis of the crystal is perpendicular to the flat face of the slab, the crystal is said to be X-cut (see Fig. 6.25). ‘Two other common cuts are the ¥ cut and the Z cut. Reference to Table 6.1 indicates that each crystal material can best. be used only in certain cuts. SHEAR PLATES. If two edges of an X-cut Rochelle salt crystal are Parallel to the Y and Z axes, a shear plate is obtained (see Fig. 6.28). When a shear plate is used as a transducer, a metal foil is cemented to each side of it, as shown in Fig, 6.28, The two foils and the crystal itself form an electrical eapacitor of the solid dielectric type. When the166 MICROPHONES (Chop. 6 Part XV} PRESSURE MICROPHONES 167 crystal is deformed as shown in Fig. 6.28b, the resulting shearing effect in the crystal causes a charge to appear on the eapacitor, thereby producing | UE es eset Eee lstic Suber 4 potential differenco between the two plates, witha certain polarity, come a Deformation of the crystal in the opposite direction, as shown in Fig i : 6.28e, reverses the polarity. Similarly, if a potential is applied between | tne . = \. . the two faces of the slab, shearing stresses are produced in the plane i Norham | ADE ' Gone peo of the plate and a deformation takes place like that of (b) or (c) depending — - on the polarity. masiomeee (tomes, [Trewees [Roam] toga ogee Microphones generally use either X or ¥ cuts of Rochelle salt, and : Sade | Somer | = ot Z cuts of ADP. ender tn EXPANDER PLAS, Another important form of erystal transducer, Common ote] xe ater: Ys | anes xnar viz, the 45° expander plate, is obtained if cutis taken from a shear plate eee re eee | ee ere Towieimnry cut] “Oven” [Anat nem [iewtian0t |Lacwtinn | tas tan Valens roi fee Fi 637 [iow chm | Woof ore daw 10 j Semprnt tw Jaco | 4 1 i pte | econ frase asm free corny | stove 200 mie | Som oe st | eve iat ‘ rats | foreneret Me | oe ° © Cy we | | : Fio, 628. Distortion of «shear plate eur Fie. 0.29. (a) Method of eutting | az | wa | wn [umayman | oom vwhen a potential is applied between two fols 45" expander bar fromashesr pte, “ii ‘ghee emented toils faces.” (a) Placement of fol. (@) Sketch showing expansion of me ‘her pee (0) Distortion of shear plate with one polarity sides and compression of length af | sre | set mate ©) Distortion of shear plate with opposite uch bar whens potential i breed os | polarity. (Coutny of Brush Blactronir Com applied between the "two. faces pany) (Courtesy of Brush Blectronice Com: | na on | on in the manner shown in Fig, 6.29. From Fig. 6.28, it is apparent that. | Mivehnany | ™ = when a voltage is applied to a shear plate, one diagonal lengthens and the (greeted | other shortens. Henee, if a crystal is made by cutting a piece from a ' Dientctoone | shear plate in the manner shown in Fig. 6.29a, potentials will be gener- sare wa | as “0 ated by squeezing or extending the crystal along its length, as shown in Msimum tra + fro asx se Fig. 6.29, ‘ec ie | sectarns | ‘novecaat Most piezoelectric crystals of commercial interest, including Rochelle ontemion) salt and ADP, yield no significant output when subjected to hydrostatic — pressure. This happens because the algebraic sum of the potentials "He deta aratin developed by deformations along the three axes simultaneously is zero.168 MICROPHON (Chup. 6 However, there is a sizable response of the lithium sulfate erystal to hydrostatie pressure ‘Phe same eleetromechanical relationships exist fora erystal microphone as for an electrostatic microphone (see the preceding. sect part), oz. where e = open-circuit voltage in volts produced by a deformation & of the erystal in meters, f = force acting to deform the erystal produced bby an electrical change 9, and 7 is » coupling coefficient defined after Bq. (6.276). As was described in Part, VIII (pages 71 to 75), such a ral Me (b) Fra, 630. Blectromechanical equivalent cirenita. for piezoelectric microphones impedance analogy). (a) Circuit, for one face of piezdeletric clement blocked, em. cemented to a rigid surfuce, (8) Circuit for both face fee to move. device may be represented by the equivalent circuit of Fig. 6.300, provided one of the two faces of the crystal across which the force is being produced is held stationary. If both faces are permitted to move, the equivalent cireuit must be modified as shown in Fig, 6:306. Jn these circuits, Cy is the electrical capacitance, in farads, measured at low frequencies with the crystal in a vacuum (or, for all practical Durposes, in ait). My is the effective mechanical mass, in kilograms; Nw is the electromechanical transformation ratio in volts per newton (or eters per eouloml); and Cy is the mechanical compliance, in meters per howton, measured at low frequencies with the electrical terminals ‘open-rircuited. For Rochelle salt X cut erystals, the electrical eapacitance varies as a function of temperature. ‘The factor of proportianality y for unmounted and unrestrained crystals is shown in Fig, 6.31. ‘Two discontinuities in Pact XV} PRESSURE MICROPHON 169 this curve occur at 18°C and +22°C and are known as the Curie points." At these points, the capacitance varies radically with temperature and is influenced considerably by the manner in which the crystal is mounted, For example, a mounted crystal of the Bimorpht type has a value of Cy at 22°C, which isa little over three times that at 10°C. This anomalous behavior is usually described as a violent variation of the free dielectric constant. An obvious way to avoid variations in the terminal voltage corresponding to variations in capacitance is to ‘operate the microphone into an open circuit, 100210" sono" * x04 10x10" suo" i a a Tempeatuein °C Fio. 6.1, Factor of proportionality in the expression for elcetrical capacitance Co pl /2t of an unmounted X-eit Rochelle salt expander bar as a function of temperature. ‘The discontinuities at —18°C and +22°C are known as the points.” (Courtesy of Brush Blectronics Company.) ‘The electromechanical transformation ratio Nw for this type of crystal in effect does not vary with temperature. For an unmounted uncoated crystal the variation amounts to about 0.05 db per degree Fahrenheit. For a coated crystal mounted in a microphone this variation may be as little as 0.02 db per degree Fahrenheit. Again, this holds for a erystal working into an open circuit. A discussion of the detailed performance of piezocleteic dileetries and their resemblance to fertomsgnetic substances appears in a text now in press: T. F. Hueter and IH. Bolt, *Sonies,” John Wiley & Sons, Ine, New York 4 Bimorph is @ registered trade-mark of Brush Biectronies Company, Cleveland, Ohio,170 MICROPHONES [chep. 6 ‘The element sizes (Fig, 6.30) for a Rochelle salt X-cut length expander bar are Nw = 0093/0, Cu = 814 X 10-"I/wt, Mu = 715!ut, and Ce = nlw/2t. The dimensions t, w, and tin meters are as shown in Fig, 6.32, and pis given in Fig, 6.31. "The element sizes for an ADP Z-cut expander bar are Nw = 0.185/'v, Cy = ATA X 10-Yl/we, Muy = 737lut, and Cy = 128 X 10-lw/t. The dimensions l, w, and £ in meters are as shown in Fig. 6.324. ‘The clement sizes for a Y-cut LH thickness expander bar (the elec~ trical field is parallel to the mechanical deformation of the crystal, sce Fig. 6.826) are Nw = 0.175t/lw, Cw = 16.3 X 10-"t/to, Nw = 882l04, and Cx = 91 X 10-#hw/t. The dimensions 1, w, and ¢ are in meters. ( A... “ver the end Length expander bar Pi sene rk (0) —_Thammsegne Fic, 632. Dimensions, method of mounting, and pointe of appliration of force for (@) length expander bar; (8) thickness expander bar. ‘The element sizes, in mks acoustic-impedance units, for an LH block ‘operated in the hydrostatic mode are Nz = 0.1484, C4 = 32 X 10-*lut, and Cx = 91 X 10-Mlw/t. The value of M- has little significance because of ‘the various resonances that can be excited. The dimensions J, to, and t in meters are as shown in Fig. 6.82b except that the crystal is free of support. mimonesis. An important disadvantage of the expander-bar type of structure in some applications is its large mechanical impedance. For uuse in liquids, such an impedance is desirable, but in air, because of the large mechano-acoustic impedance mismatch, very small output voltages aire obtained for normal sound pressures. ‘To reduce this mechanical impedance appreciably without lowering the output voltage, two plates or bars may be combined to produce a Bimorph, In a sense, a Bimnorph isa mechanical transformer operating on & principle resembling that, of a Part XV} PRESSURE MICROPHONE: i bimetallic strip. ‘The flat faces of two erystals are cemented together in such a way that when a potential is applied to them one expands and the other contracts. Conversely, a force acting perpendicular to the face of the “‘bimetallie” strip causes a sizable compressive foree in one plate and a tensile force in the other. Examples of a bender Bimorph and a torque Bimorph are shown in Figs. 6.33 and 6.34, respectively. The former makes use of two expander bars and the latter of two shear plates. Generally, the bender Bimorph e = iA rT @ 1 | o ieee come, t — areca | clamped (6) Force! 7 ae i i = re be Ls © @ Pic. 638. Bender Bimorph. (a) Two Fic. 6.34. Torque Bimorph. (a), (b expandercbar erystals ready to be ee- and (¢) Distortion of shear plates when Iented together to form a bender potentials of opposite polarity are Bimorph. (0) Application of potentials Applied. scram their Tacos. (8) Two of “opposite polarity to. the erystals shear plates are cemented together to ‘causes the upper to lengthen snd to form a torque Bimorph. When either contract in width and the other to force for clectricnl potentials are shorten and to expand in width. (2) applied, the Bimorph assumes the shape When these two erystals are cemented shown.” (Couttay ef Brush Hlecivotce together and either 4 force for electrical Company.) potentials are applied, the crystal as- Ssumnes. the shape shoven, (Courlesy of Brush Blectrones Company.) is supported at both ends and the force applied in the middle, although it can be clamped at one end and the force applied at the other. A torque Bimorph is generally held at three corners and the force applied at the fourth. With these types of structures, a force applied as shown in Figs. 6.38c and 6.34d is equivalent to a greater force acting on the end of a single expander bar or along one end of a single shear plate. Electrodes may be applied to a Bimorph in two ways to form either a series or a parallel arrangement of the two plates. For a series connection, the clectrical terminals are the two outer foils. For a parallel connection, the foil between the crystals forms one terminal, and the two ‘outside foils connected together form the other172 MICKOPHON (Chap. 6 Acsories Bimorph has one-quarter the capacitance and Lice the output voltage of a parallel Bimorph, Refer to the circuit of Fig. 6.30a (the circuit of Fig. 6.301 cannot be used for Bimorphs mounted as described above). For square-torque Bimorphs made from X-cut Rochelle salt shear plates as shown in Fig, 6.35at connected electrically in paratel, the element sizes at 30°C are Ny = 0.143/t, Cu = 287 X 1O-ME/E, Me = 2620, and Cy = 8X LIBE/LX 10% ‘The factor Bis given in Fig, 6.36 and equals 1.0.t30°C. ‘The dimensions Land farein meters. When the plates are connected clectrivally in series, im 7 oe = ae ole plates 7 < _ 3 4 ad — ah 7 fa) Diving tA Fra. 6.35, 2 35 40 “Terperatre in Factor of proportionality 9 ‘mounting, dimensions, and points of in the expression for clectrieal wayne Application of foree for (a) squnee tances Cx of an X-eut. Moree salt torque Binorph and (0) rectangilae tare torque Iinorph mounted as bender Bimnorph. shown in Fig. 86a. (Courtay of Hash Hlerionies Compan.) Method of Fra. 6.36, the value of Ww is doubled and that of Cis reduced to one-fourth its arallel vale. Por ADP square-torque Bimorphst (vee Vig. 6 in parattel, the cloment sizes are Nw = 0: My = 2280, and Cp = 5OTE/t X 10-*. For Rockelle salt bender Bimorphs (cee Vig. 6.35b) connected electrically in parallel, the element sizes are Nw = 0.07354/ut, Cw = 179 X 10 U/l, Ma = A07ot, and Ce = 118 X10-%w/t. ‘The dimensions 1, w, and ( fare in meters. When the plates are connected electrically in series, the value of Ny is doubled and that of Cr is reduced to one-fourth its parallel value. ‘The cireuit of Fig. 6:30a must be used For ADP bender Himorphs (see Fig. 6.350) connected electrically in 5a) connected electrically 50/t, Cu = 145 X 10 ME YE, 4 Element sizes for revtangulas-torque Bimorphs uf Rochelle salt or ADP are avai able from the Bush Hlectronien Company, Cleveland, Ohi Part XV] PRESSURE MICKOPHONES 173 parallel, the element sizes are Nay = O.1S231/ut, Cu = 19 X 10-0 /r’, Mac = 435let, anid Ce = 5.09 X 10-tw/t ‘The nonlinear distortion produced by ADP and LU erystals is very small, Rochelle salt is an exception, however. In the temperature range from 18°C to +24°C, it exhibits hysteresis effects in the relation between the applied foree and the voltage produced across a small load resistance. ‘This hysteresis effect arises in the electrical capacitance Cr anil is of negligible importance if the microphone operates in a near-open- cineuit condition, such as into the input of a eathode-follower vacu tube stage, In microphones working at ordinary sound levels, the 100 ja Maximum low frequncy ext! in celet 000 0 Tempeatuein °C Fie. 637, Maximum eapectel low-frequency cutoff in eyeles for ADP:(PN) erystals ratn function of ambient temperate in centigrade degrees. ‘Tho cutoft frequency lofined a that free Wve eeaetanee and the resistance (eae wear tice the equal. (Courtesy of Brush Blectronies Company) thre Tae) be hysteresis effect in Rochelle salt is negligible even when the erystal is loaded All erystals have a bulk re vityt which appears as a shunt. resistance across the lectrical terminals of the erystal. ‘This faetoris not of practical importance at ordinary temperatures exeept in ADP crystals. Vor ADP crystals, the bulk resistivity increases rapidly with tempera ture, as ean be seen from Wig, 6.87. ‘The ordinate of this graph is the cutolt frequency, cual lo the frequency at which the reactance and resistance measured actoss two faces of a erystal plate are equal. 1 Bale resistivity ete esis in sb toon fu fara nea174 MICROPHONES. Chap. 6 The self-noise produced by an electrostatic microphone is that produced by the d-e resistance of the crystal or dielectric shunted by the capacitance Cx. In addition, the load resistance will contribute to the noise in that it is in parallel with the internal resistance. ‘The spectrum level of the self-noise drops off at:6 db per octave as soon as the capacitive reactance becomes less than the total shunt resistance (external and internal). A crystal that has been damaged by heat or humidity is likely to have a high internal noise level. ‘Temperature not only affects the electrical capacitance of Rochelle salt crystals and the resistivity of the ADP crystals but also may permanently damage the crystals. Data showing the approximate maximum safe temperature, temperatures for appreciable leakage, and the temperatures for complete destruction are given in Table 6.1 Crystals are affected by humidity, as can be seen from Table 6.1. ‘This is particularly true of the Rochelle salt type. Rochelle salt: chemically is sodium potassium tartrate with four molecules of water of erystallization. If the humidity is too low (less than about 30 per cent), the crystal gradually dehydrates and becomes a powder. If the humidity is too high (above 84 per cent), the erystal gradually dissolves. Neither result is reversible, Protective coatings of recont vintage give nearly complete protection from the influence of humidity extremes. To test a crystal for normalcy, measure its resistivity and electrical capacitance, being careful not to apply so much voltage to the crystal that, it will fracture. If the capacitance is lower than normal, the crystal has dehydrated. If the resistivity is too low and the eapacitance is too high, tive crystal has partially or entirely dissolved. If the resistivity is low and the shunt capacitance has not changed, surface leakage is taking place. Barium Titanate Ceramic Microphones. A. recent. and useful form of piezoelectric material is a ceramie made from barium titanate. ‘The ceramic is rendered piezoelectrie by permanently polarizing it with a high electrostatic potential of about 40,000 to 60,000 volis/in. for a period of several minutes, Lower voltage is satisfactory if polarization is cartied out at elevated temperatures. Pure barium titanate has a Curie point like that of Rochelle salt, except that it occurs at a temperature of 120°C, well above the normal operating range for a microphone. Dielectric anomalies also exist near 12°C. Barium titanate microphones may be used interchangeably with crystal microphones, except that. their sensitivity is about 6 to 18 db below that of Rochelle salt or ADP types. ‘Temperature changes do not affect the dielectric constant appreciably ‘The advantages of harium titanate over the crystal materials described above are that it has a higher dielectric constant than that for Rochelle sall, and one that changes very little with temperature in the range Part XV) PRESSURE MICROPHONES 175 between +10 and +80°C, The mechanical compliance Cy and the transducer ratio Nw are lower than for Rochelle salt. Two length- expander units which are made from Rochelle salt and barium titanate, respectively, and which are designed to have the same capacitance and mechanical compliance at 15°C will differ in sensitivity by 12 to 16 db, the barium titanate unit being the less sensitive of the two. Barium titanate ceramics exhibit a slow decrease in dielectric constant. and clectromechanical response with age. ‘The decrease of each quantity is believed to be less than 10 per cent, that is, 1.0 db, during the first year after manufacture and only a few per cent per year thereafter. The free unmounted capacity Cy as a function of temperature of barium titanate ceramic unit relative to the free capacity for a tempera ture of 25°C is shown in Fig. 6.38, 20 . ols. 4 1 i i a a a ET) Temperature in° Fro. 638, Barium titanate ceramic. Relative free capacitance Ce as a function of temperature, (Courtesy of Brush Blecironice Company.) The exact nature of this curve depends upon the manner in which the dielectric is rendered piezoclectre, Referring to the circuits of Fig. 6.30, for barium titanate length- expander bar like that shown in Fig. 6.324, the element sizes are Vy = 0.0051 /w, Cu = 10.2 X 10-"1/ut, Mor = 2260hwt, and Ce (at 25°C) 1.5 X 10-lw/t, ‘The dimensions J, w, and { are in meters. For a barium titanate thickness-expander bar like that in Fig. 6.82b, the element sizes are Nyy = 0.0127t/lw, Cu = 8.13 X 10-¥t/lu, My = 2260Iut, and Cr (at 25°C) = 1.5 X 10- w/t. ‘The dimensions 1, w, and ¢ are in meters, Referring to the circuit of Fig. 6.30a, the element sizes for a barium fitanate bender Bimorph like that shown in Fig. 6.35b, with the plates connected electrically in series, are Vw = 0.00681/uY, Cw = 40.6 X J0-"1/wt, My = 1530h0t, and Cx (at 25°C) = 1,28 10~"hw/t ‘The barium titanate transducer also responds to hydrostatic pressures ‘such as would be obtained if a ceramic plate were suspended freely in air orin a liquid. ‘The circuit of Fig. 6.30a is likewise used here except that170 MICROPHONES Chap. 6 Insulating pte Tun meta dapvagm itt contact wth 45" expander bars ‘(AOP or Rocnele st) ler Dt foe ough skcoh ny P. Masa of Mase Fal.) Pesoratea ‘fran ro Sintered Diophagm ret te Bimeroh Fe. 6.10, Diaphragm type af crystal mierophione aig lange sintered tal ph susie npg. seven nd tensile Me cote Part XY PRESSURE MICROPHONES. elements, Th cease the transducer is called a volume expander, and the element si Wa = 0.00221 volt/nevston/m® (or m'/roulomb), Cx = 12.2 x 1 rm*/newton, and Ce = 1.5 X 10-Mw/t farads, ‘The dimensions J, are in meters. Uses of Piezoelectric Transducers. Ways in which piezoclectrie ducers are incorporated into microphones are shoxcn in Figs. 6.39 t In the first of these figures, a microphone 1 is sketched using ADP or Rochelle salt ‘Ei crystals forming 45°-cut expander bars or barium titanate Jength-expander bars. A sound pressure, causes a force to be exerted on the end of the piezoelectric plates, and a voltage is generated because of changes in their lengths. The sides of the plates are free to expand into the air space. In Fig. 6.10 a diaphragm is connected to a torque Bimorph element by a short rod. The Bimorph is mounted securely Castor ‘on three corners so that the force is up- ">" plied in the manner shown in Fig. 6.354. ‘The sintered metal plate in [cunt of the méUming Pod \e diaphragm is an acoustic resistance that Fra, 6.42. Sketch of = damps the resonance of the mechanical statically actuated erystal a honing a a Another form of mounting Bimorph ‘Seite Geass § plates is shown in Fig. 6.41. Here two convey, the sound square bender Bimorph elements are held Unllormly, 2 al suriaces part by two rubber blocks. ‘The entire sink et-pites cone ‘assembly isencased in a thin wax-impreg- parellel, 14 in. on a side wi nated paper jacket, thereby forming two Bate OOK In. thick, on “diaphragms” with an airspace between. gr farad and. the ope ‘An inerease in sound pressure causes a evtpt vag about distortion of the plates as shown in Fig. Yl nerdeme/em’ (Ce 641, This distortion is of such a nature. Brak Bl ee 5 to produce a potential difference across the cleetrodes fastened Piezoelectric plates ‘A fourth form of mounting suitable for use with piezoelectric plat respond to hydrostatic changes in pressure is shown in Fig, 6.42. block of four lithium sulfate plates is shown immersed in a castor 0 ‘The outer flexible housing serves both as a diaphragm and asa retai the castor oil. A pressure over the outside surface is transferred t the housing and the castor oil bath to the element inside, acoustival elements are substituted for mecha178 MICROPHON Chap. 6 part XVI Gradient and Combination Microphones 6.7. Pressure-gradient Microphones. General Features. ‘The ribbon microphone has approximately the same sensitivity and impedance as a moving-coil microphone when used with a suitable impedance-matching transformer, Because of its figure 8 directivity patter it is extensively used in broadcast and public-address applications to eliminate unwanted sounds that are situated in space, relative to the microphone, about 90° from those sounds which are wanted. It is also used by singers to Top clamp. (insulteay Twi teas BG)ar Section — Pole pieces through poe piece a Bottom clamp asta Twisted lead Po, 648, Sketch of the ribbon and magnetic structure for a velocity microphone (Aer Oluon, Elements of Acoustical Engineering, 24 el, p. 238, D. Van Nostrand Company, Ine, 1947.) introduce a “throaty” or ‘“bassy’” quality into their voice. A disadvan tage of the ribbon microphone is that, unless elaborate wind sereening is resorted to, ibis often very noisy when used outdoors, Construction, A typical form of pressure-gradient microphone is that, represented by Fig. 6.43. It consists of a ribbon with a very low resonant frequency hung in a slot in a baile. A magnetic field transverses the slot so that a movement of the ribbon causes a potential difference to appear across its ends. In this way, the moving conductor also serves as the diaphragm, In modern design, the ribbon element might be 1 in. Jong, 446 in. wide, and 0.0001 in. thick with a clearance of 0.003 in. at each side, From Bq. (6.11) we see that the pressure difference acting to move the diaphragm is Pn = wpe Al cos 8 (6.38)180 MICROPHONES Chap. 6 Substitution of (6.38) in (6.89) yields ty) «(B00 At 2Mua + Mu lel = S cos 6 (6.40) The open-circuit. voltage is directly proportional to the component of the particle velocity perpendicular to the plane of the ribbon. In a. well-designed ribbon microphone, this relation holds true over the fre- ‘quency range from 50 to 10,000 eps. The lower resonance frequency is usually about 15 to 25 eps. ‘The effects of diffraction begin at frequencies cof about 2000 eps but are counterbalanced by appropriate shaping of the magnetic pole pieces, 6.8. Combination Pressure and Pressure-gradient Microphones. Electrical Combination of Pressure and Pressure-gradient Transducers. One possible way of producing a directivity pattern that has a single ‘maximum (Go-called unidirectional characteristic) is to combine electrically the outputs of a pressure and @ pressure-gradient microphone fp th al hte Fic, 6.45. Simplifed electromechanical analogous citeuit for a ribbon microphone (amobility analogy). The two units must be located as near to each other in space as possible 50 that the resulting directional characteristic will be substantially independent of frequency. Microphones with unidirectional, or cardioid, characteristics are used primarily in broadcast or publie-address applications where it is desired to suppress unwanted sounds that are situated, with respect to the microphone, about 180° from wanted sounds. In respect to impedance and sensitivity this type of cardioid microphone is similar to ribbon or to a moving-coil microphone when suitable impedance-matching transformers are used The equation for the magnitude of the open-cizeuit output voltage of a pressure microphone in the frequency range where its response is “fit” ia @= Ap a1) ‘The equation for the open-circuit output voltage of a magnetic or ribbon type of pressure-geadient microphone in the same frequency range is = Cpcose (642) Pact XVI GRADIENT MICROPHONES 181 B-co phs of the expression ® = 20 log {1 + B cos #)/(1 + B)] a a funetion 0,05, 10,50, and =. Fie, 6.46, of Ofor B Adding (6.41) and (6.42) and letting C/A = B gives e = Ap(l + B cos 0) (6.43) B will be a real positive number only if ¢ and ¢, have the same phase. ‘The directional characteristic for a microphone obeying Eq, (6.43) will depend on the value of For B = 0, the microphone is nondiree-182 MICROPHONE: (Chap. 6 tional type; for B = 1, the microphone is a cardioid type; for B = «, the rierophone is a figure 8 type. In Fig. 6.46 directional characteristics for five values of B are shown, ‘The voltage ¢, is a function of kr, as we discussed in Par. 6.3, so that the voltage « as given by Eq. (6.43) will vary as a function of frequency for small values of ar/e, where ris the distance between the microphone and a small source of sound. Here, as is the case for a pressure-gradient microphone, a ““bassy ” quality is imparted to a person’s voice if he stands very near the microphone. Acoustical Combination of Pressure and Pressure-gradient Microphones. One example of an acoustical design responding to both pressure andl pressure gradient in a sound wave was described earlier in Par. 63 (pp. 149 to 150). ‘The directional patterns for this type of design are the same as those shown for Fig. 6.46. In order that this type of microphone have a flat response as a function of frequency for Pm constant (ée., constant sound pressure at all frequencies in the sound wave), a transducer must be chosen whose output, voltage for a constant differential force acting on the diaphragm is inversely proportional to the quantity @ defined in Eq. (6.19), i.e, 1 Zan — j\(Ra + Zan) /wCaRa) cee py = [Beales Basel] gay ‘As an example, let us take the case of a microphone for which Zan >> Ra and 1/oC4R,>> 1. In this case the response of the transducer must be proportional to ho Tal * SCR, ~ Ste (6445) where B is given by Eq. (6.17). Restated, the transducer must have an output voltage for a constant net force acting on the diaphragm that is inversely proportional to fre ‘quency, if a flat frequeney response is desired. This is the case for a moving-coil or ribbon transducer above the natural resonance frequency of the diaphragm. CHAPTER 7 DIRECT-RADIATOR LOUDSPEAKERS parr XVII Basie Theory of Direct-radiator Loudspeakers ‘TA. Introduction. A loudspeaker is an electromagnetic transducer for converting electrical signals into sounds. There are two principal types of loudspeakers: those in which the vibrating surface (called the diaphragm) radiates sound directly into the air, and those in which a horn is interposed between the diaphragm and the air. The direct-radiator type is used in most home radio receiving sets, in phonographs, and in small public-address ayatema. ‘The horn type is used in high-fidelity reproducing systems, in large sound systems in theaters and auditoriums, and in music and outdoor-announcing systems. ‘The principal advantages of the direct-radiator type are (1) small size, (2) low cost, and (3) a satisfactory response over & comparatively wide frequency range. ‘The principal disadvantages are (1) low efficiency, (2) narrow directivity pattern at high frequencies, and (8) frequently, irregular response curve at high frequencies. For use in home radio receiving sets where little acoustic power is necessary and where the listeners are generally not very critical, the advantages far outweigh the disadvantages. In theater and outdoor sound systems where large amounts of acoustic power are necessary and where space is not impor- ‘ant, the more efficient. horn-type loudspeaker is generally used. ‘All the types of transduction discussed in the previous chapter on Microphones might be used for loudspeakers. In this text, however, we shall limit ourselves to moving-coil loudspeakers, the type commonly used im radios and home musie systems. 72, Construction.! A. cross-sectional sketch of a typical direct. radiator loudspeaker is shown in Fig. 7.1. The diaphragm is a cone, + For supplemental reading, the student will Bnd the following publications vali able: Ti F. Olson, “Elements of Acoustical Engineering,” 24 ed,, Chap. VI, D. Van Nostrand Company, Ine, New York, 1947; M. 8. Corsington, Amplitudes and Phase Mensurements om Loudspeaker Cones, Prac, 1B, 88: 1021-1026 (1951); Transient “Testing of Loudspeakers, Avulin Fngineering, 4 9-18 (August, 1980), vas184 DIRE V-RADIATOR LOUDSPEAKERS (Chap. 7 generally made of paper or aluminum, which is supported at the outer edge and near the voice coil so that itis free to move only in an axial direction. Current through the voice coil ereates a magnetomotive force which interacts with the air-gop flux of the permanent magnet and causes A translatory movement of the voice coil and, hence, of the cone to which it is attached. Usually the cone is sufficiently sti at low frequencies to move as a whole. At high frequencies, however, vibrations from the renter travel outward toward the edge in the form of waves. ‘The results of 1. Infinitely large bate 2. Fee edge suspension ot 3 Fens center suspension 4, Voie ca 5 Electrical connections 6 Hotes fora rate 7. Gone (aphregm) Permanent magnet 5. Open ne supporting oto Fia. 71, Cromnctionel sketch of dieeetradiator loudspeaker assumed to be snownted in an infinite bale of these traveling waves and of resonances in the cone itself are to produce irregularities in the frequency-response curve at the higher frequencies and to influence the relative amounts of sound radiated in different, directions In Fig. 7.1, the loudspeaker is shown mounted in a flat baffle assumed to be of infinite extent. By definition, a baffle is any: means for acous tically isolating the front side of the diaphragm from the rear side. For purposes of analysis, the loudspeaker diaphragm may be considered at low frequencies to be a piston of radius a moving with uniform velocity over its entire surface, This is a fair approximation at frequencies for which the distance b on Fig. 7.1 is less than about one-tenth wavelength Part XV] DIRECT-RADIATOR LOUDSPEAKERS 185 7.3. Blectro-mechano-acoustical Circuit. Before drawing a circuit diagram for @ loudspeaker, we must identify the various elements involved. ‘The voiee coil has inductance and resistance, which we shall call L and Rs, respectively. The diaphragm and the wire on the voice coil have a total mass Myo. ‘The diaphragm is mounted on flexible suspensions at the center and at the edge. The total effect of these suspensions may be represented by a mechanical compliance Gus and a mechan ical resistance Rus = 1/rws, where rus is the mechanical responsiveness. (a) Fic, 72. (a) Mechanical circuit of direet-adiator louapeaker; (0) electromechanienl snaiogous circuit of the mobility type; () electrieal eiruit showing motional electrical impedance; (2) analogous circuit of the mobility type with electrical quantities referred to the mechanical side. ‘The air cavity and the holes at the rear of the center portion of the diaphragm form an acoustic network which, in most loudspeakers, can be neglected in analysis because they have no approciable influence on the performance of the loudspeaker. However, both the rear and the front side of the main part of the diaphragm radiate sound into the open air. ‘A radiation impedance is assigned to cach side and is designated a5 Zoun = Sens, where 2un is the radiation mobility. We observe that one side of each flexible suspension is at zero velocity. For the mechanical resistance this also must be true because it is con186 DIRECT-RADIATOR LOUDSPEAKERS —[Chap.7 tained in the suspensions. We already know from eater chapters that, one side of the mass and one side of the radiation mobility must be con~ sidered as having zero velocity. Similarly, we note that the other sides of the masses, the compliance, the responsiveness, and the radiation mobilities all have the same velocity, viz, that of the voice coil From inspection we are able to draw a mechanical circuit and the clectromechanical analogous circuit using the mobility analogy. ‘These are shown in Fig. 7.2a and b, respectively. ‘The symbols have the following meanings: €4 = open-circuit voltage of the generator (audio amplifier) in volts. generator resistance in electrical ohms inductance of voice coil in henrys, measured with the voice- coil movement blocked, i, for 1, = 0. Re = resistance of voice coil in electrical ohms, measured in the same manner as B = steady air-gap flux density in webers per square meter. 1 = length of wire in meters on the voice-coil winding. electric current in amperes through the voice-coil winding. force in newtons generated by interaction between the alternating and steady mmf, that ig, f. = Bli 1% = voice-coil velocity in meters per second, that is, u. = e/Bl, where ¢ is the so-called counter emf. radius of diaphragm in meters. mass of the diaphragm and the voice coil in kilograms. Cus = total mechanical compliance of the suspensions in meters per newton. Tus = 1/Rus = mechanical responsiveness of the suspension in meters per newton-second (mks mechanical mohims#). Rus = mechanical resistance of the suspensions in newton-seconds per meter (mks mechanical obms). tun = 1/Zun = twe + jown = mechanical radiation mobility in mks mechanical mohms from one side of the diaphragm (see Fig. 5.4). ‘The German t indicates that twe varies with frequency. Zun = Rive + iXue = mechanical radiation impedance in newton- seconds per meter (mks mechanical ohms) from one side of a piston of radius a mounted in an infinite batile (see Fig, 5:3). ‘The German 3 indicates that Rux varies with frequency, R, fe ‘The circuit of Fig. 7.2) with the mechanical side brought through the ransformer to the electrical side is shown in Fig. 7.2c. ‘The mechanical nobility 21 = t/fis zero if the diaphragm is blocked so that there is no {A molim isa mobility om. See Pat. 33 for discussion. T-HADIATOR LOUDSPEAKERS 187 Port XVII} DIRE motion (ue = 0) but has a value different from zero whenever there is motion, For this reason the quantity BY, is usually called the motional electrical impedance, When the electrical side is brought over to the mechanical side, we have the cireuit of Fig, 7.2d ‘The circuit of Fig. 7.2d will be easier to solve if its form is modified First we recognize the equivalence of the two circuits shown in Fig. 7.30 (@) () ia, 73, ‘The electrical cireuit (referred to the mechanical sie) is shown here in two ‘equivalent forms. The circuits are of the mobility type. peaustc Electcal Mechaniat ‘aaigion ON aii eh a) an 0) 1. 7, (a) Low-frequency analogous cieuit of the impedance type with electrical ‘untls sete to methane fide Zug ie given by Pg, 8. "The une f Tepresente the total force acting in the equivalent eicuit to produce the voice- Sua, and b, Next we substitute Fig. 7.96 for its equivalent in Fig. 7.24. ‘Then we take the dual of Fig. 7.2d to obtain Fig. 74a. ‘The performance of a direct-radiator loudspeaker is directly related to the diaphragm velocity. Having solved for it, we may compute the coustie power radiated and the sound pressure produced at any given distance from the loudspeaker in the far-field Voice-coit Velocity at Medium and Low Frequencies. ‘The voiee-coil velocity uw, neglecting 'L* compared with (Ry + Rs)*, is found from188 PCT-RADIATOR LOUDSPEAKERS —(Ghep 7 Fig, 740, el = RF Ria FIR) a rore _ Bt 2 Ru = Rothe + Rus + Ru (7.2) Xu =oMy = eM + Xx = 3h 3) Voice-coit Velocity at Low Frequencies. At low frequencies, assuming n addition that Xia" >> Rus’, we have from Fig. 7.40 that, 1 Kader wea (Maco + 2M) — ay here My; = 2.670%» = mass in kilograms contributed by the air load mn one side of the piston for the frequency range in which ka < 0.5. The juantity ka equals the ratio of the circumference of the diaphragm to the vavelength. ‘The voice-coil velocity is found from Eq. (7.1), using Eqs. (7.2) and 7.4) for Ry and Xw, respectively. TA. Power Output. The aeoustic power radiated in watts from both he rear and the front sides of the loudspeaker is W = [u29Ren) 5) Fence, asuming wD" & (Ry + Re)’, Deg Ben W = OF Ra + Xa) ) 7.5. Sound Pressure Produced at Distance r. Low Frequencies. In Shap. 4 we showed that a piston whose diameter is less than one-third ravelength (ka < 1,0) is essentially nondirectionsl at low frequencies, Fence, we ean approximate it by a hemisphere whose rms volume velocity quals U, = Sot, where Sp is the projected area of the loudspeaker cone. ay the projected area, we mean a? of Fig. 7.1 From Eq. (4.3) we see that the magnitude of the rms pressure at a joint in free space a distance r from either side of the loudspeaker in an nfinite baile is WW Ip@| = a7) t is assumed in writing this equation that the distance r is great enough o that it is situated in the “far-field.” Hence, the pressure at ris eBlSofoa Ol Re Ra VR Xa (78) Part XVII] DIRECT-RADIATOR LOUDSPEAKERS 189 Equation (7.8) is also readily derived from Eq. (7.6) by observing from ‘Table 5.1 (page 124) that, at low frequencies, e'Sotoe San = Sere 9) and W = dart = (7.10) where I is the intensity at distance r in watts per square meter Medium Frequencies. At medium frequencies, where the radiation from the diaphragm becomes directional but yet where the diaphragm vibrates as one unit, /-c., a8 a rigid piston, the pressure produced at a distance r depends ou the power radiated and the directivity factor Q. ‘The directivity factor Q was defined in Chap. 4 as the ratio of the intensity on a designated axis of a sound radiator to the intensity that would be produced at the same position by & point source radiating the same acoustic power. From Eq. (7.10) we see for a point source radiating to both sides of an infinite baffle that We ae? Ini = qa For a directional source in an infinite baffle such as we are considering here, Wikowe tar? Pol = (7.12) where W; = acoustic power in watts radiated from one side of the Toudspeaker. Q = directivity factor for one side of a piston in an infinite plane bafite, Values of @ are found from Fig. 4.20. Note that W, equals 17/2 and, at low frequencies where there is no directionality, Q= 2, so that Eq, (7.12) reduces to Eq (7.11) at low frequencies. "The sound pressure is found by substituting Eq, (7.6) divided by 2 into (7.12), giving, Bl GoeWns Br Vu (Re + Re) VRaP oF Xa 1.6. Frequency-response Curves. A frequeney-response curve of loudspeaker is defined as the variation in sound pressure or acoustic power as a function of frequency, with some quantity such as voltage o electrical power held constant. Inspection of Eqs. (7.6), (7.8), and (7.13) shows that the quantity Ru? + Xu? in the denominator, ee in the numerator, and the directivity factor Q are terms that vary with free ivr) (7.13)190 DIRECT-RADIATOR LOUDSPEAKERS — [Chap 7 quency, The variation in sound pressure or diaphragm velocity due to the variation of the denominator is exactly the same as the variation of electric current as a function of frequency in a series RLC electrical resonant circuit. A plot of this variation as a funetion of normalized fre- ‘quency is called a universal resonance eurve. At low frequencies, the quantity Stwx varies with the square of the frequeney. So, if Q is constant, the numerator of (7.13) varies in direct proportion to frequency. If Q is not constant but increases with frequency, the variation is more rapid. In other words, neglecting the directivity (Ge, let @ = constant), the curve of sound pressure as a function of frequency is identical in shape to a universal electrical resonance curve multiplied by frequency. ‘When an electrical universal resonance curve is expressed in decibels, below the resonance frequeney it has a slope of +6 db per octave of fre~ quency. Above the resonance frequency it has a slope of ~6 db per octave, In the case of a sound-pressure or an acoustic-power vs. frequency-response curve, the slopes are further increased by the linear frequency term, so that below the resonance frequency the slope is +12 db per octave and above it the slope is 0 db (Rat). ‘Whether one considers sound pressure or power radiated, the frequency response curve when expressed in decibels has the same shape, provided the directivity factor is constant. We shall now study one of the quantities, the power radiated, as a function of the quantity °/(Itu? + Xw) 7.1. Maximum Power Available Efficiency (PAE). Often, the response of a loudspeaker is stated in terms ofits mazimum power available eficiency, which is 100 times the ratio of the acoustie power radiated to the maximum power that the electrical generator can supply. The maximum power is available from the electrical generator when the load resistance ‘equals the generator resistance. Wem te Can) Medium and Low Frequencies. ‘The mazimum power available efficiency (PAE) at medium and low frequencies |w%L* K (Ry + Re)'] is Eq. (7.8) divided by the maximum power available. aw SOOB TR ue 5 PAR = Wy, X10 = Ge pmmaters 18) where Ry and Xy are given by Eqs. (7.2) and (7.8). At the lower frequencies we may replace Rua by Eq. (7.9) and Xy by Bq. (74) Let us divide the frequency region into five parts and treat each part separately by simplifying the circuit of Fig. 74a to eorrespond to that part alone. Reference is made to Figs. 7.5 and 7.6 for the breakdown Part XVII] DIRECT-RADIATOR LOUDSPEAKERS, 191 In region A, where the loudspeaker is stiffness-controlled, the power output increases as the fourth power of frequency or 12 db per octave. Tn region B, at the resonance frequency ws, the power output is determined by the total resistance in the circuit because X passes through % i Gj a all (2) Verto requenes fg Te BI aii Saat Us Pan (b) Ax peincpa resonance requeney wy a aeat Geko BP Rent (G) Newnd resonance reeoney Mio we exBl sa 2poera (2) High tequences Fio, 7.5, Simplified forms of the circuit of Fig. 74a valid over limited frequenc: ranges In region C, above the first resonance frequency, the power output (anv the sound pressure) approaches a constant value, provided the circui ‘impedance approaches being a pure mass reactance. That is to say ‘tier increases with the square of the frequency, and Xw* also increases a the square of the frequeney, and so the frequency variation cancels oui For small values of amplifier resistance R, the total mechanical resist ance Ru becomes quite large in some loudspeakers so that the resonane is more than critically damped, Reference to Eq. (7.8) shows that j192 DINECT-RADIATOR LOUDSPEAKERS [Chap.7 Ra? >> Xt, the sound pressure increases linearly with frequeney J. This condition is shown in Fig. 7.6 by the dashed line. High Prequencies, Referring back to Fig. 74a, we see that there is a possibility of a second resonance taking place involving L/B%l and the masses Muy + Xux/a. ‘The voice-coil velocity at this resonance ean be determined from the circuit of Fig. 7.5d. ‘The resonance frequeney will occur when olinte war ag 7 one + ae 7.16) ‘We must note, however, that if (It, + »)*is large compared with L's, the reactance of the eapacitance L/B* and resistance B°/(Ie, + Rs) z a g 10 Frequency (logaritimic scale) Fic, 76. Graph of the power available eiiency in decibels of « hypothetical direet- radiator loudspeaker in an infinite bafle. It is assumed that the diaphragm acts Tike a rigid piston over the entire frequency range. ‘The power is the total radiated from both sides of the diaphragm. Zero decibels ie the reference power available effcieney level. The solid curve is for a loudspeaker with « Qr of about 2. The dashed curve is for « Qr equal to wbout 0.5, n parallel becomes that of a negative inductance equal to ~B°PL/ (Ry + Re)*. In this ease, no resonance ean occur. pas tlution of Fig. 7.54 epic othe pe ofthe region marked Din Fig. 7.0 [At frequencies above the second resonance frequency, the radiation resistance on each side of the diaphragm becomes approximately equal to ‘ra'pec, where a is the effective radius of the loudspeaker. Also, a’Mwo ‘becomes large compared with the resistance in the circuit, and w'Z? becomes large compared with (R, + Rr). The voice-coil velocity is determined from Fig. 7.5¢. ‘The power available efficiency is, 800R,B*pyceat a *M sa aay) ‘This region is marked Hin Fig. 7.6. Here, the power output decreases: by 12 db for each doubling of frequeney Pat XV] DIRECT-RADIATOR LOUDSPEAKE! 193 ‘Phe response curve given in Fig. 7.6 is for a typical loudspeaker used for the reproduction of music in the home. For this application, the mass of the cone is made as light as possible and the compliance of the suspension as high as possible consistent with mechanical stability. For Special applications, Cvs can be small so that the resonance frequency is high. Also, it is common in practice to make Rw so large that the yelocity u. 38 nearly constant as a function of frequency through regions B and C. In this case, the sound pressure inereases linearly with frequency, and there is no flat region C. 18. Reference Efficiency. It. is convenient. to define a reference efficiency which permits one to plot the shape of the frequeney-response curve without showing the actual acoustic power that is being radiated ft the time. The reference power available efficiency (both sides of the diaphragm) is defined as, 800R,B*Stws PABw = C.F Ra)'st(Muvo + Me™ a or, with the help of Ea. (7.9), 800K, BPPSoton PABY = scp a ay) If the loudspeaker is Jess than critically damped, Eq. (7.19) gives the actual response in frequency region C, which lies above the first resonance frequency. Even for loudspeakers that are highly damped so that there is no flat region C, Eq, (7.19) forms a convenient reference to which the rest of the curve is compared, Expressed as a ratio, the PAE response at medium and low frequencies where the radiation is nondirectional {see Bq. (7.15)] is PAB _ o'(Muv +2Mu)* PABS "Ra + Xt Gm) At the resonance frequency ws, where Xw = 0, PAB _ wr'(Myo + 2Mu _ gs 20 PAES Rut where Qr is analogous to the Q of electrical circuits. Equations (7.20) and (7.21) may be expressed in decibels by taking 10 logie of both sides of the equations, In Chap. 8 of this book, on Loudspeaker Enclosures, design charts are presented from which it is possible to determine, without laborious com- putation, the sound pressure from a direct-radiator loudspeaker as a function of frequency including the directivity characteristics. Methods for determining the constants of loudspeakers and of box and bass-reflex enclosures are also presented, If the reader is interested only ir learning194 DIRECT-RADIATOR LOUDSPEAKERS wap. how to choose a baffle for a loudspeaker, he may proceed directly to Chap. 8. ‘The next part deals with the factors in design that determine the over-all response and efficiency of the loudspeaker. 7.9. Examples of Loudspeaker Calculations Example 74. Given the reference power avalable efficiency of Ba, (719) fora loudspeaker in un infinite ball, determine the reference sound prewure equvelent to the reference ower avilable ficiency aaniming thatthe directivity actor @ or ‘adintion tone aie) equals. Solum. The sound presse at distance resuming no direstivty, io rated to the ncountie power radisted to one wide a fllome ee a. 7.12) po Veal = Vee where J = intensity at distance r W, = W/2 total acoustic oer adie fom one side of the diaphragm ‘The equivalent reference sound prea it a= NaS eeBlSo, IR, FR) Mun + BA) Example 7.2. As an example of the power available efficiency to be expected from 8 direct-radiator loudspeaker of conventional desiga mounted 1a an infinite bale tnd radiating from both sides of the baffle, let us calculate the reference power available effcieney (PA) from Bq. (7.18) for the cate of a commercial loudspeaker with an advertised diameter of 12 in. Also, let us ealeulate the ratio of the PAE to the reference power available efficiency at the frst resonance frequency. ‘Typical values of the constants are B = 10,000 gauss = 1.0 weber/m* 1=9m L= 7X10 henry Rp = Bohms R, = 2ohme of diaphragm = 0.18 m fective ares of disphragm = 0.0531 mt fs density of air = 1.18 eg /m* ‘¢ = speed of sound ~ 344.8 m/sec Stun ~ 167o%atpa/e (see Table 6.1) = 1.58 X 10-%st newton-see ma May = 2.67p.2% (o00 Table 5.1) = 0,00604 ke, Mao = 0011 ke Rus = 0.5 mks mechanieal obm Cua = 179 X10 Solution. Prom Bq, (7.18), we obtain 800 2X 18 X 9X 1.58 x 10-4 TOF X (0025)* ton from one side of the loudspeaker only, divide this Ggure by 2. Only 1.6 percent of the available electrical power in region C of Fig. 7.0 is radiated ‘to one side of the dixphragm. This illustrates the statement made at the boginning ofthe chapter that the efficiency of this typeof loudspeaker i usually low. PAB = 232% For ra Part XVI} DESIGN FACTORS 19 "The upper resonance frequency, if such exists, i determined f mn Bg, 7.16), 58, ie + SUB ] For our example, the (Ite +) is so large compared with L4s* that the shunt resistance and enpacitanco act like a series resistance and a negative induetanee, ane no second resonance oveurs inthe frequency range for which & lamped-elementeireut hl, ‘The boundary between regions C and. D of Fig. 7.5 oecars when ka lies approxi mately between I and 2, For our exumple ka = 1 corresponds toa frequency af ceeeaa Dea” FXO Obviously, a smaller diaphragm of lighter weight would result in region Cextending tom higher Trequeney. However, a reduction in the mass Mp occasioned by 8 smaller diaphragm will cause an increase in the first resonance frequency With 3 resulting loss in bass response. A further disadvantage of smaller diaphragm ic that, fora given sound pressure, a grester voie-col velocity tues needed. A longer air gap and a larger magnet structure must therefore be provided. "The firt resonance frequency equals = sh eps Bo tus where Mar Mary + 28 = 0025 ke eo that 100 aD = 4 fa Toms e Ru = 8p 405-103 4 $9 mks mechanical ohms aM (4200085) or = Sl 120025) 2y rom Ba, (7:21) we se thatthe ratio of the PAE at uso the rference PAB is equal to.Grs, Hone thie vats equals 1.74, and PAB at equals 8. percent (bth de). vant XVIII Design Factors Affecting Direct-radiator Loud- speaker Performance A loudspeaker generally is designed to provide an efficient transfer of electric power into acoustic power and to effect this transfer uniformly over as wide a frequency range as possible. To accomplish this, the voice coll, diaphragm, and amplifier must be properly chosen. ‘The choice of the elements and their effect on efficiency, directivity, and transient, response are discussed here. 7.10. Voice-coil Design. Inspection of Fig. 7.6 reveals that region C isa very important part of the response curve, because the average effie196 DIRECT-RADIATOR LOUDSPEAKERS (Chap. 7 cioney is governed by it. From Eq. (7.19), which is valid for this region, if the indicated approximations hold, we see that the maximum power available efficiency at a given frequeney is proportional to a [Re + RM vo + UMP Now, the resistance Re can be expressed in terms of the mass of the woice-coil winding Muse by writing PAE « (7.22) (7.23) where « resistivity of voice-coil conductor in units of chm-meters, ‘The value of x for different materials is given in Table 7.1 ay = radius of wire in meters. 1 = length of voice-coil winding in meters. Also, Mue = 10,4lp0 (2a) where pu = density of the voice-coil wire in kilograms per cubie meter (see Table 7.1). Combining (7.28) and (7.24), we get spe EF Re = ee (7.28) Substituting (7.24) in (7.22) yields PAE « PRs (7.26) Ret (Ove + Migs + 2M) Mae where Miyy = Muro — Mu. Differentiation of this equation with respect to Me and equating the result to zero gives the value of Marc necessary for maximum power output from a generator of impedance R, provided we assume that the coil length is already predetermined. Hence, Mwe for maximum PAE is found from Mact = (My, + 2M an) a2 Further, substituting (7.25) in (7.27), we get aoc Bl + 2809 oe As an alternate possibility, we assume that the resistance of the coil, Rr, is to be constant. Allow Muc and [to vary. Then determine M we for maximum PAE. Part XVIII) DESIGN FACTORS 197 TABLE 7.1. Resistivity and Density of Various Metals Metal clement [Resistivity ohm-n avity, chem] Density, kg/m" ‘omioum om x10 | 2.70 x10 iuaneny oat oe Samat 0 38 Caimi oars 87 Gaisum owe te Carbon se 235 Can Om i Cromiam oan be Gant oom on Copper torre He Gale oom 03 frm ovo Ba Tren on 1s Lead 0.220 | 11.0 fit om 0. Magresiom o.oWe er Masuanee Oost re Merry Oost ws Mabenum | Oost 12 Nick | ore a3 Pou | Gon ver Siver | ons ws soln Cone os Te os 3 Tiaiam So i Tuersten oss wo ee oso nt From Bq, (7.25), we have p= Bellas 729) om Putting this in (7.22) yields RRM uc PAE S EP Eyfnc Mp Fm 780) where, 08 above, Mi, = Mun —Muc. Maximizing, we get Mure = Migp + Mon (7.31) Finally, let us assume that M we of Eq. (7.30) is a constant and that we wish to let Re and I vary. Then determine Rx for maximum PAE. Maximizing Bq, (7.30), we get Ra = Ry (7.32) Hence, for the optimum value of power available efficiency, we see198 DINECT WADIATOR LOUDSPEAKERS [Chap 7 im mnesnese 007 Clamping eee poe oise* sre" re rae Fio. 77, Detail of the edge of a felted-paper loudspeaker cone from an Sin. loudspeaker. [After Corrington, Amplitude and Phare Measurements on Loudepeaker Cones, Proc. TRB, 89: 1021-1026 (1951), 8 ects reltve to relerance pressure 2.4 : impetince z-4 = & Se 3 ~ist wm 1000 Frequency in eels per second Pio, 7.8. Relative power-available response of un 8-in.-diameter loudspeaker mounted in an infinite baffle. The dashed curve was computed from Figs. 8.12 and 8.13 for Qr=6. [After Corrington, Amplitude and Phase Measurements on Loudspeaker Cones, Proc. IRB, 89: ¥021-1026 (1951). 0,00 | from Ras. (7.28), (781), and (7.82), that Re = R, and Muro = Miyy + 2M an Itis not usual, however, that the voice coil should be this massive, for the reason that a large voiee eoil demands a correspondingly large magnet, structure, Values of voice-coil resistances and masses for typical American loudspeakers are given in Table 8.1 of the next chapter. Part XVID} DESIGN FACTORS 199 TAL. Diaphragm Behavior. ‘The simple theory using the method of equivalent circuits, which we have just derived, is not valid above some frequency between 300 and 1000 eps. In the higher frequency range the cone no longer moves as a single unit, and the diaphragm mass M wo and also the radiation impedance change. ‘These changes may occur with great rapidity as a function of frequency. As a result, no tractable mathematical treatment is available by which the exact performance of a loudspeaker can be predicted in the higher frequency range. A detailed study of one particular loudspeaker is reported here as an example of the behavior of the diaphragm.? The diapbragm is a felted paper cone, about 6.7 in. in effective diameter (see Fig, 7.7), having an cluded angle of 118°, ‘The sound-pressure-level response curve for this loudspeaker measured igo on the principal axis is shown in Fig, & Se 7.8, This particular loudspeaker has, a in addition to its fundamental reson- gy ance, other peaks and dips in the response at points 1 to 8 as indicated on the curve. ‘The major resonance at 90 eps is the principal resonance and has the relative amplitude given by Eq. (721). “ay a Above thatiis the fairly flat region that yg,"V', Nodal pattern of the cone we have called region C. At point , of the loudspeaker whose response which is located at 420 eps, the cone curve, i given in, Fig. 7-8. The breaks up into aresonance of the form faded, and dented ines ada shown by the first sketch in Fig. 7.9. The (+) and (~) signs indicate Here, there are four nodal lies on the eplons moving in oppor restos, cone extending radially, and four re- t+ opposia phases erring gions of maximum movement. Asin- dicated by the plus and minus signs, two regions move outward while two regions move inward. ‘The net effeet is a pumping of air back and forth across the modal lines. The cone is also vibrating as a whole in and out of the page. The net change in the output is an increase of about 5 db relative to that computed. A similar situation exists at point 2 at £48, Conrington, Amplitude and Ph Proc. URE, 380: 12-1026 (1951) 349 es o © 3309 & we Measurements on Loudspeaker Cones,200 DIRECT-RADIATOR LOUDSPEAKERS. 500 eps, except that the number of nodal lines is increased from 4 to 6. At point 3, 650 eps, the vibration becomes more complex. Nodal lines are no longer well defined, and the speaker vibrates in such a way that the inerease in pressure level is about 4.5 db, exclusive of 0.5 db inerease due to directivity For point 4 at 940 eps, a new type of vibration has become quite appar ent. ‘The diaphragm moves in phase everywhere except at the rim. Looking at the rim construction shown in Fig. 7.7 and at the vibration pattern of Fig. 7.9(4), we ean deduce what happens. ‘The eenter part of the cone vibrates at a fuirly small amplitude while the main part of the cone has a larger amplitude. At the 5.78in. diameter the amplitude of vibration is very small. At this point the corrugation has a large radius (0.156 in.). As the cone moves to and fro, the paper tends to roll around this curve, and this excites the 0.094in. corrugation that follows into violent oscillation at its resonance frequeney. The rim resonance is 180° out of phase with respect to the main part of the cone, However, ‘the main part of the cone has a high amplitude produced by the rocking motion around the 5.78-in. diameter, and because of its greater area, only part of its effect in producing a high sound level is canceled out by the rim motion. ‘The net result is a peak in output (see point 4 of Fig. 7.8) ‘At point 5, 1100 eps, a sharp decrease in response is observed. ‘The decrease seems to be the result of a movement of the nodal line toward the apex of the cone, and a reduction of the amplitude of the (+) portion. Here, the effect is a pumping of air back and forth across the nodal line, with a cancellation in output. ‘This vibration is very characteristic, and at the time such motion occurs, the response drops vigorously. ‘As frequency is increased, the loudspeaker breaks up into still different characteristic modes of vibration. As shown in Fig. 7.9, case 6, several nodal lines appear concentric to the rim of the loudspeaker. When these ‘occur, a large increase in output is obtained, as shown at point 6 of Fig 7.8. ‘As frequeney is increased, other such resonances occur, with more nodal lines becoming apparent. ‘These nodal lines are the result of waves traveling from the voice coil out to the edge of the cone and being reflected back again. These outwardly and inwardly traveling waves combine to produce a standing-wave pattern that will radiate a maximum of power at some particular angle with the principal axis of the loudspeaker. In order to reduce standing-wave patterns of the type shown in eases 6,7, and 8 of Fig, 7.9, itis necessary that a termination of proper mechan- ia! impedance be placed at the outer edge of the diaphragm. This termination must be one that absorbs waves traveling outward from the center of the cone so that no wave is reflected back. In practical design, 4 leather supporting edge is frequently employed. Such a leather might bea very soft sheepskin, having @ weight of approximately 3 oz/yd2, and Part XVII} DESIGN FACTORS 201 carefully tanned, A leather supporting edge is also effective in reducing the rim resonance. ‘The resulting effect is to produce a more uniform response in the frequency region between 700 and 1500 eps of Fig. 7.8. 7.2, Divided-cone Driving Unit. We have seen that important deterrents to a flat response at high frequencies are large radius and large mass of the diaphragm. A logical means, therefore, for improving the Cue M a =e 2 (6) Fic, 7.10, Two methods for effectively having two sizes of cones with a single voice coil, (@) Two separate cones joined by a compliant element. (6) One gone with a compliant element molded into it at fraction ofits radius. arizal Fio, 7.11. Mechano-acoustical cirouit for loudspeakers with the cones shown in Fig. 7.0. high-frequency response would be to design the diaphragm so that at the high frequencies only the portion of it near the voice coil will move. Two means for doing this are shown in Fig. 7.10. ‘The performance of these diaphragms is indicated by the equivalent circuit of Fig. 7.11. Here, the quantities Mo, Cs, and Rus are the mechanical constants of the smaller (inner) cone, and Mus, Cus, and Ruz are the mechanical eon- stants of the outer cone. ‘The two are connected mechanically with each other through a compliance Cys. ‘The radiation impedance of the inner cone is Zuas, and that of the outer cone is Zar. The other constants are the same as in Fig. 74. At high frequencies 1/Cys is essentially a short eiteuit, and only the constants of the smaller cone are involved.202 DIREC RADIATOR LOUDSPEAKERS (Chap. 7 In practice the design of Fig, 7.105 is more commonly used because of its lower cost. Another but more expensive means for separating the inner portion of the cone from the outer is to use a leather strip for the compliance Cs A material sometimes used for this is goatskin, having a weight of about 2.5 og/yd#, At the highest frequencies at which the inner cone of the loudspeaker is to operate, the goatskin acts to absorb waves traveling in the cone outward from the voice coil. A smoother response curve is thereby obtained. 7.13. Multiple Driving Units. Another means of accomplishing the equivalent of several sizes of cones is to mount two or more loudspeakers of different diameters near each other. An electrical network, called a crossover network, is used to supply electrical power to one loudspeaker at High-requeney wn beriegery Fic, 742. A “coaxial” or Fro. 7.18. Cross section of a “triaxial,” or “three "two-way," loudspeaker. way” loudspeaker. (Courlaty of Jensen Mfg. Con, Chicago, 1.) low frequencies and to the other, or others, at higher frequencies, A. ifficulty with this arrangement is that if the loudspeakers are mounted side by side, the path that the sound has to travel from each of the loud~ speakers to a listener will be different in different parts of the listening room. Hence, in the vicinity of the crossover frequencies cancellation of the sound will result at some parts of the room, and addition will occur at others, ‘To avoid this effect, the loudspeakers are often mounted concentrically ‘ie,, the smaller loudspeakers are placed in the front of and on the axis of the larger loudspeaker (see Figs. 7.12 and 7.13), In the vicinity of the crossover frequency there is usually some shielding of the radiation from the larger loudspeakers by the smaller ones, with resulting irregularity in the response curve, ‘TAA. Directivity Characteristics. The response curve of Fig. 7.6 and the information of the previous three paragraphs reveal that, above the Part XVII) DESIGN FACTORS 203 frequency where ka = 2 (usually between 800 to 2000 cps), a direct- radiator loudspeaker can be expected to radiate less and less power. ‘The rate at which the radiated power would decrease, if the cone were a rigid piston, is between 6 and 12db for each doubling of frequency. This decrease in power output is not as apparent directly in front of the loud~ speaker as at the sides because of directivity. ‘That is to say, at high frequencies, the cone directs a larger proportion of the power slong the axis than in other directions. Also, the decrease in power is overcome in part by the resonances that occur in the diaphragm, as we have seen from Fig. 78. Directivity Patterns for T'ypical Loudspeakers. ‘Typical directivity pat terns for a 12-in.-diameter direct-radiator loudspeaker, mounted in one of the two largest sides of a closed box having the dimensions 27 by 20 by 12 in, were shown in Fig. 4.23. These data are approximately correct for loudspeakers of other diameters if the frequencies beneath the graphs fare multiplied by the ratio of 12 in. to the diameter of the loudspeaker in inches. Comparison with the directivity patterns for a flat rigid piston in the end of a long tube, as shown in Fig. 4.12, reveals that the directivity patterns for a flat piston are different from those for an actual loudspeaker. ‘This difference results from the cone angle, the apeed of propagation of sound in the cone relative to that in the air, and the resonances in the cone, In this connection, it is interesting to see how the speed varies with frequency in an actual cone. Speed of Propagation of Sound in Cone. Let us define the average speed ‘of propagation of sound in the cone as the distance between the apex and the rim, divided by the number of wavelengths in that distance, multiplied by the frequency in cycles per second. For the particular 8-in. loudspeaker of Figs. 7.7 to 7.9, the phase shift and the average speed of propagation of the sound wave from the apex to the rim of the cone are given in Fig. 7.14. At low frequencies the cone moves in phase so that the speed can be considered infinite. At high frequencies the speed asymptotically approaches that in a flat sheet of the same material, infinite in size Intensity Level on Designated Anis. We have stated already that at high frequencies loudspeaker diaphragm becomes directional. In order to calculate the enhancement of the sound pressure on the axis of the loudspeaker as compared with that indicated by the equations for maximum available power efficiency, it is convenient to use the concepts of directivity factor and of directivity index as defined in Part XI (pages 109 to 115). For example, we might wish to know the intensity (or the sound pressure level) on the axis of the loudspeaker, given the power ‘available response and the directivity factor. This is done as follows: ‘The intensity as a function of frequency on the axis of symmetry of the204 DIRECT-RADIATOR LOUDSPEAKERS (Chap. 7 loudspeaker divided by the electrical power available is equal to the product of (1) the power-aveilable-efficiency response characteristic, (2) the directivity factor, and (8) 1/4xr*, where r is the distance at which the intensity is being measured. In decibels, we have 1Olowse HE = log PAB: + DI — 10logieter? (7.88) where Ia: = Par?/oee = intensity in watts per square meter on the designated axis at a particular frequency Pos = sound pressure level in newtons per square meter measured on the designated axis at a particular frequency PAE, = W,/Ws = ratio of total acoustic power in watts radiated by the front side of the loudspeaker to the maximum electrical power in watts available from the source [see Eq. (7.14)] and where DI is given by Eq, (4.19) and Fig. 4.20. Note that, for the piston in an infinite baffle, the DI at low frequencies is 3 db because the . = eens 1 2| JomEs ‘Baal TTT] 1600 z = z ey pam ; LS to 3 & fe 400 4 lo eee Frequency in cycles pee second Fra. 74. Phaso shift and average wave apced in the cone of an Sin, londspeaker. TAfter Corringion, Amplitude and Phase Measuremente on Loudspeaker Cones, Proc TRE, 99: 1021-1026 (1951).] power is radiated into a hemisphere, and that the last term of Eq. (7.32) is the area of a sphere, in decibels. Expressed in terms of the sound pressure level on the designated axis +76 0.0002 microbar (0.00002 newton/m?), Eq. (7.33) becomes SPL re 0.0002 microbar = 20ogye gE 5 = 10 loge We + 10 logie PAE, + DI ~ 10 loge drt + 1Olog pe + 9db (7.34) Pact XVII) DESIGN FACTORS 205 7.16. Transient Response. The design of a loudspeaker enclosure and the choice of amplifier impedance eventually must be based on subjective judgments as to what constitutes “quality” or perhaps simply on listening “‘satisfaction.” It is believed by many observers that a flat sound-pressure-level response over at least the frequency range between 70 and 7000 eps is found desirable by most listeners. Some observers believe that the response should be flat below 1000 eps but that between 1000 snd 4000 cps it should be about 5 db higher than its below-1000-eps value, Above 4000 eps, the response should return to its low-frequency value. It is also believed by some observers that those loudspeakers ‘which sound best generally reproduce tone burstst well, although this requirement is better substantiated in the literature for the high fre- ‘quencies than for the low. ‘An important factor determining the transient response of the cireuits of Fig. 7.4 is the amount of damping of the motion of the loudspeaker diaphragm that is present. For a given loudspeaker, the damping may be changed (1) by choice of the amplifier impedance R, ot (2) by adjust ‘ment of the resistive component of the impedance of the enclosure for the loudspeaker, or (8) by both. Generally, the dampingis adjusted by choice ‘of amplifier impedance because this is easier to do. The instantaneous velocity u, for both steady-state and snddenly applied sine waves is [sin (ot — 0) + ahem sin (wo + va (7.35) = [Sule = Ru + 5Xw (7.36) where Rw and Xw are given by Eqs. (7.2) and (7.4), 737) ag atts] 2r times the resonance frequency (7.38) ON ACs” EM? From Bq. (7.1) we see that the first fraction on the right side of (7.35) and the first term within the brackets is the steady-state term. The second term within the brackets is the transient term, which dies out at the rate of exp (—Rwl/2If w) Tt is known that the reverberation time in the average living room is A tone burst is a wave-teain pulse that containg a number of waves of « certain Frequency.206 DIRECT-RADIATOR LOUDSPEAKERS (Chap. 7 aout 0 se, which coresponds 0 4 deny conan of 138 som cease tf «trast fara hw rat este ta sl iO ene tees ‘are not disturbed by the “overhang” of the sound. This corresponds to, et asec ahi ia more rai say tan oar n the average living room. Although criteria for acceptable transient dis- th ee em led er lndnpeery omens etna ‘assume that if the decay constant for a loudspeaker is greater than four ‘times this quantity, i.¢., greater than 92 see~', no serious objection will be Ce en othe nent csang wih tone Dus ‘Accordingly, the criterion that is suggested here as representing satisfactory transient performance is. Be ens 30 ‘Equation (7.35) reveals that, the greater Rw/2Mw, the shorter the transient. Equation (7.39) should be construed as setting a lower limit, ‘on the amount of damping that must be introduced into the system. Tt is not known how much damping ought to be introduced beyond this mri aon Tn the next chapter we shall discuss the relation between the criterion ‘of Eq. (7.39) and the response curve with bafile. eae ace spnances (eqs pa 110 8in Fg 78) has associated with it a transient decay time determined from an equation like Eq. (7.35). In order to fulfill the criterion of Eq, (7.39), it is usually Moe atts spar tone and. ena ge that a response curve smoother than that shown in Fig, 7.8 is obtained. ‘With the very best direct-radiator loudspeakers much smoother response we ewe encase ad the poston ented ae estou uaentdiorion an asso sponse cose a et eau fre cmt depen Example 7.8. If tho circular gap in the permanent magnet has a radial length of og .em, a circumference of Bem, acd an axial length of 1.0 em, determine the eneray stored in the sir gap if the fux density is 10,000 geuss. ‘Solution. Volume of sir gap = (0.002)(0.08)(0.01) = 1.6 X 10° m* lux density = 1 weber/m™ rom books on magnetic deviees we find that the energy stored is woe oy = te 10°F weber/ampeturn im). Hence, the = 01636 joule Part XVENI) DESIGN FACTORS 207 Example TA. A 12in, loudspeaker is mounted sn one of the two largest sides of closed box having the dimensions 27 by 20 by 12 in. Determine ant plot the relative power available efficiency and the relative sound pressure Level on the principal ass ‘Soluion. Typical directivity patterns for this loudspeaker are shown in Fig. 4.23 “The directivity index on the principal axis axa function af frequency is shown in Fig. 424, Ttis interesting to note that tho transition frequency from low directivity to high dicectivity i about. $00 pe, Since the effective radius of the radiating cone for this loudapeaker ia about 0.13 m, ka at this transition frequency is Zefa _ 1000r x 0.18 c 7 ans el Las or neatly unity, as would be expected from on provious studies. ‘The transition from fegion C [where we assuined that tM yi? > Riau and wil (Ry + Re}" to region B of Fig. 7.6 also occurs at about ka = I In the frequency region betweon ka = 0.5 and ka = 3, the loudspeaker can be represented by the cirouit of Fig, 74a. Let us assume that itis mounted in an infinite ‘aif and thet one-half the power ie radiated to each side. Also, let us assume that the amplifier impedance is very low. "The power available efciency, from one side ofthe loudspeaer, Ww, $0084 R, Sve PAE = Wg ~ ORF ear + Xn) Ried where Bw ~s- Rar = us + 280) TE ow eas * Me NEY RE ay Xue = (oun + 2Xwa) V+ a @, FR VIF ORE FA a2) If we assume the constants for I tis seen thet, above f = 1000 eps, 10 mple 7.2, we obtain the solid curve of Fig. 7.152 fe power available eficiency drops of Response in desbale srowary relrence 200 oo 2000 Frege ince Send Fin, 718, Gra ofthe lave pow aval ene and the wud rare level mensured on the printpal sui of typial T-n inmeterloudspeater i & -closed-box baffle. The reference level is chosen arbitrarily. am Now, let us determine the sound pressure level an the prineipal axis of the loud speaker, using Hq. (734). The directivity index for a piston in a long tube is found ftom Fig. 420, The rosulte are given by the dashed curve in Fig, 7.18. Obviously, the directivity index is of great valic in maintaining the Frequeney respnnse on the principal axis out to higher frequencies. AL stil higher Frequencies, eone resonances ‘cur, a8 we suid before, and the typical espouse curve of Fig 7.8 i obtainedCHAPTER 8 LOUDSPEAKER ENCLOSURES part XIX Simple Enclosures Loudspeaker enclosures are the subject of more controversy than any other item connected with modem high-fidelity music reproduction, Because the behavior of enclosures has not. been clearly understood, and because no single authoritative reference has existed on the subject, ‘opinions and pseudo theories as to the effects of enclosures on loudspeaker response have been many and conflicting, ‘The problem is complicated further because the design of an enclosure should be undertaken only with full knowledge of the characteristics of the loudspeaker and of the amplifier available, and these data are not ordinarily supplied by the manufacturer. ‘A large part of the difficulty of selecting a loudspeaker and its enclosure arises from the fact that the psychoacoustic factors involved in the reproduction of speech and music are not understood. Listeners will rank- order differently four apparently identical loudspeakers placed in four identical enclosures. It has been remarked that if one selects his own components, builds his own enclosure, and is convinced he has made a wise choice of design, then his own loudspeaker sounds better to him than does anyone else's loudspeaker. In this case, the frequency response of the loudspeaker seems to play only a minor part in forming a person's opinion, In this chapter, we shall discuss only the physics of the problem. The Uesigner should be able to achieve, from this information, any reasonable frequency-response curve that he may desire. Further than that, he wil, have to seck information elsewhere or to decide for himself which shape of frequency-response curve will give greatest pleasure to himsolf and to other listeners, With the information of this chapter, the high-fidelity enthusiast should be able to calculate, if he understands ace cireuit theory, the frequency response curve for his amplifier-loudspeaker-baflic combination. Design 208 Part XIX} SIMPL) CLOSURES 209 graphs are presented to simplify the calculations, and three complete examples are worked out in detail. Unfortunately, the ealeulations are sometimes tedious, but there is no short cut to the answer, ’As we have stated earlier, all calculations are based on the mks system. A conversion table is given in Appendix II that permits ready conversion from English units. The advantage of working with metersand kilograms is that all electrical quantities may be expressed in ordinary watts, volts, ohms, and amperes. It is believed that use of the mks system leads to Jess confusion than use of the cgs system’ where powers are in ergs per second, electrical potentials are volts X 10%, electrical currents are amperes X 10-1, and impedances are ohms x 10% 8.1, Unbaffied Direct-radiator Loudspeaker. A balfte is a structure for shielding the front-side radiation of a loudspeaker diaphragm from the rear-side radiation. ‘The necessity for shielding the front side from the rear A side can be understood if we consider that an unbaffied loudspeaker at low L frequencies is the equivalent of a pair of simple spherical sources of equalstrength located near each other and pulsing out $y ‘of phase (see Fig. 8.1). The rearside of Bio. g1. Douhlat sound xource the diaphragm of the Joudspenker is silent a low freien equivalent to one of these sources, and a2 unbaffed vibrating diaphragm. the front side is equivalent to the other, TMs Print A islovsind e distance © Tf we measure, as a function of fre- the axa the louapeaker quency f, the magnitude of the rms Sound pressure p ata point A, fairly well removed from these two sources, find if we hold the volume velocity of each constant, we find from Eq. (4.15) that, (| = EE cos 8 61) where Uy = rms strength of each simple source in cubie meters per second. b = separation between the simple sources in meters. ps = density of air in kilograms per cubie meter (1.18 kg/m* for ordinary temperature and pressure). ‘r= distance in meters from the sources to the point A. It is assumed that 13> b. 0 = angle shown in Fig, 8.1 ‘¢ = speed of sound in meters per second (344 8m/sec, normally). HL P. Olson, “Elements of Aroustical Engineering," 24 ed., pp. 84-85, Table $3, D. Vin Nostrand Company, Ine, New York, 1947. For a discussion of simple Toudspeaker enclosures, soe pp. N14 154210 LOUDSPEAKER ENCLOSURES, (Chap. 8 In other words, for a constant-volume velocity of the loudspeaker diaphragm, the pressure p measured at a distance r is proportional to the square of the frequency f and to the cosine of the angle @ and is inversely proportional to r. In terms of decibels, the sound pressure p increases at the rate of 12 db for each octave (doubling) in frequency. Tn the case of an actual unbaffied loudspeaker, below the first resonance frequency where the system is stiffness-controlled, the velocity of the diaphragm is not constant but doubles with each doubling of frequency. ‘This is an increase in velocity of 6 db per octave. Hence, the pressure p from a loudspeaker without a baffie increases 12 +6 = 18 db for each. octave increase in frequency. Above the first resonance frequency, where the system is mass-controlled, the velocity of the diaphragm decreases 6 db for each octave in frequency. Hence, in that region, the pressure p increases 12 — 6 = 6 db for each octave increase in frequency. 8.2. Infinite Baffle. In the previous chapter we talked about direct- radiator loudspeakers in infinite bates. Reference to Fig. 7.6 reveals that with an infinite baffe, the response of a direct-radiator loudspeaker is enhanced over that just indicated for no baffle. Tt. was shown that if ‘one is above the first resonance frequency usually the response is fat, with frequency unless the BI product is large (region C) and that if one is below the first resonance frequeney the response decreases at the rate of 12 db pet octave insteal of 18 Ub per octave. Hence, the isolation of the front. side from the back side by an infinite bafile is definitely advantageous. In practice, the equivalent of an infinite baffte is a very large enclosure, well damped by absorbing material. One practical example is to mount. the loudspeaker in one side of a closet filled with clothing, allowing the front side of the loudspeaker to radiate into the adjoining listening room. Design charts covering the performance of a direct-radiator loudspeaker in an infinite baffle are identical to those for a closed-box. We shall present these charts in Par. 8.5. 8.3. Finite-sized Flat Baffle. The discussion above indicated that it is advisable to shield completely one side of the loudspeaker from the other, as by mounting the loudspeaker in a closet. Another possible alternative is to mount the loudspeaker in a flat baffle of finite size, free to stand at one end of the listening room. The performance of a loudspeaker in a free-standing flat baffle leaves much to be desired, however. If the wavelength of a tone being radiated is greater than twice the smallest lateral dimension of the baflle, the loudspeaker will act according to Eq. (8.1). ‘This means that for a finite flat haffle to act approximately like an infinite baflle at 50 cps, its smallest Iateral dimension must be about 3.5 m (11.5 ft). However, even above this frequency, sound waves traveling from behind the loudspeaker reflect (off walls and meet with those from the front and cause alternate cancella- tions and reinforcements of the sound as the two waves come into phase or Part XIX] SIMPLE LOSURES 0 out of phase at particular frequencies in particular parts of the room. ‘This effect can be reduced by locating the loudspeaker off center in the baflle, but it cannot be eliminated because of reflections from the walls of the room behind the loudspeaker, Also, a flat baffle makes the loudspeaker more directional than is desirable because in the plane of the baffle the sound pressure tends to reduce to zero regardless of the baffle size. 84, Open-back Cabinets. An open-back cabinet is simply a box with ‘one side missing and with the loudspeaker mounted in the side opposite the open back. Many home radios are of this type. Such a eabinet performs nearly the same as flat baffle that provides the same path length between the front and hack of the loudspeaker. One additional LS ‘1 @ = L- ote] 4 Fie. 82, Loudspeaker of radius a Fic, 83, Typieal plywood box with mounted in-an unlined bow with praer mounted off eenter in one imensions Lx E 1/2. While ined with & layer of oft ab tha type of box is convenient sorbent acoustical materia ‘analyse the eonstmeton shown ta ig. 83 ta more commonly used. effect, usually undesirable, occurs at the frequency where the depth of the box approaches a quarter wavelength. At this frequency, the box acts ‘as a resonant (ube, and more power is radiated from the rear side of the loudspeaker than at other frequencies. Furthermore, the sound from the rear may combine in phase with that from the front at about this ‘same frequency, and an abnormally large peak in the response may be obtained, 8.5. Closed-box Baffie.'* ‘The most commonly used type of loud speaker baffle is a closed box in one side of which the loudspeaker is mounted. In this type, discussed here in considerable detail, the back side of the loudspeaker is completely isolated from the front. Customary types of closed-box baffles are shown in Figs. 82 and 8.3. ‘The sides are made as rigid as possible using some material like S-ply plywood, 0.75 to 1.0 in, thick and braced to prevent resonance, A slow air leak must be ©. 3, Plach and P. 1. Williams, Loudspeaker Enclosures, Audio Bagincering, $5: 12g. (July, 1961),212 LOUDSPEAKER ENCLOSURES [Chep. 8 SUMMARY OP CLOSED-BOX BAFFLE DESIGN A, To determine the volume of the closed boa: 4 Find the values of fe (without baffe) and Cys from Par, 87 (pages 220 to 250), Approximate values may be determined from Fig, 8.5 and d », Determine Sp from Fig. 85a, and ealeulate Cas = CusSe" «. Decide what percentage shift upward in resonance frequency due to the ‘addition of the box you will tolerate, and, from the lower curve of Fig. 8.1, deter- smine the values of Cax/Cas and, hence, Cas, 4. Having Cae, determine the volume of the box from Eq. (8.7). ©. Shape and line the box according to Pars, & 6 (page 227) and &5 (page 217), 2, To determine the response of the loudspeaker at frequencies below the ri resonance frequency (about 500 pe) 4a. Find the values of Mwo, Rus; Cus, Bl, Rx, and So from Pat. 8.7 (pp. 228 to 282). Approximate values may be obtained from Fig. 85, Table 8.1, and the sen= tence preceding Table 8.1 0, Determine Man = Mun/Sot, Cas = CusSo8, and Ras = Rus/So* 6 Determine Rae, Xan, May, Cas, and Man from Baa, (8.4) to (8. 4. If the flow resistance and volume of the acoustical lining are known, determine Raxfrom Fig RR Otherwise, neglect Rue to a fret approsimation. «Determine the actual (not the rated) output resistance of the power amplife Ry. All the constants fr solving the eireut of Fig. 8.4 ate now available J. Calculate the total resistance Rg, total mase M4, and total compliance Cx from Eqs. (8.19) to 21). Determine wy and Qr from Eqs. (B22) and (8.23). 49, Determine the reference sound pressure at distance r from the loudspeaker by Eq. (8.27). A, Determine the ratios of the driving frequencies at which the desired t0 the resonance frequency wy, that is, w/a. Determine the My/Ry Obtain the frequency response in decibels relative to the reference sound pressure directly from Figs. 8.12 and 8.13, provided in the box s0 that changes in atmospheric pressure do not displace the neutral position of the diaphragm. Analogous Circuit. A closed box reacts on the back side of the loudspeaker diaphragm. This reaction may be represented by an acoustic impedance which at low frequencies is a compliance operating to stiffen the motion of the diaphragm and to raise the resonance frequency. At high frequencies, the reaction of the box, if unlined, is that of a multiresonant, cireuit, This is equivalent to an impedance that varies cyclically with frequency from zero to infinity to zero to infinity, and so on. This varying impedance causes the frequency-response curve to have corresponding peaks and dips. Part XIX} SIMPLE ENCLOSURES. 23 If the box is ined with a sound-absorbing material, these resonances are damped and at high frequencies the rear side of the diaphragm is loaded with an impedance equal to that for the diaphragm in an infinite baffle radiating into free space. At low frequencies, where the diaphragm vibrates as one unit so that it can be treated as a rigid piston, a complete electro-mechano-acoustical circuit can be drawn that describes the behavior of the box-enclosed loudspeaker. ‘This circuit is shown in Big, 8.4 and was developed by procedures given in Part XVI Some interesting facts about loudspeakers are apparent from this ireuit. First, the electrical generator (power amplifier) resistance Ry Frontside of Mechanical part “Gaphagm Beil of oadspeaker radiaten oe atast Fro, $4, Circuit diagram for a diseet-radiator loudspeaker mounted in a closed-box baffie. Thiscireut eval for fequencinsbelow about 4O0eps. The velume velocity of the diaphragm =U: e, = apeacircuit voltage of generator, My = generator resistance; Ry = voice-coll tsistance; B= airgap thix density: f = length of wire on voice-col winding; Sp ~ elletive area of the diaphragm; Map ~ acoustic mass of diaphragen and voiee coil; Cas = total acoustic compliance of the suspensiuns, Has = acoustic resistance in the suspensions; 4x, Nag ~ acoustie-radiation imped: ‘ace from the front side of the diaphragm; as, Nas = acoustic loading impedance of the box on the rear side of the diaphragm, and the voice-coil resistance Rx appear in the denominator of one of the resistances shown, This means that if one desires a highly damped or an overdamped system, it is possible to achieve this by using a power ier with very low output impedance. Second, the circuit is of the simple resonant type so that we can solve for the voice-coil volume velocity (equal to the linear velocity times the effective area of the diaphragm) by the use of universal resonance curves, Our problem becomes, therefore, one of evaluating the circuit elements and then determi the performance by ui . Values of Electrical-circuit Blements. All the elements shown in Fig, 84 are in units that yield acoustic impedances in mks acoustic ohms (newton-seconds per meter‘), which means that all elements are trans: formed to the acoustical side of the circuit. ‘This aecounts for the effec tive area of the diaphragm S,, appearing in the electrical part of the cireuit. ‘The quantities shown are open-circuit voltage in volts of the audio amplifier driving the loudspeakeraut LOUDSPE KEN ENCLOSURES Chap. 8 B = flux density in the air gap in webers per square moter (1 weber/ m? = 10 gauss) T= length of the wire wound on the voiee coil in meters (1m = 39.37 in.) R, = output electrical impedance (assumed resistive) in ohms of the audio amplifier Rr « electrical resistance of the wire on the voice coil in ohms a = effective radius in meters of the diaphragm Sp ~ za? = effective area in square meters of the diaphragm Values of the Mechanical-cireuit Elements. ‘The elements for the mechanical part of the circuit differ here from those of Part XVII in that they are transformed over to the acoustical part of the cireuit so that they yield acoustic impedances in mks acoustic ohms Mav = Mu/So? = acoustic mass of the diaphragm and voice coil in kilograms per meter Myo = mass of the diaphragm and voice coil in kilograms Cus = CusSot = acoustic compliance of the diaphragm suspensions in meters* per newton (1 newton = 10° dynes) Cus = mechanical compliance in meters per newton Ras = Rus/So* = acoustic resistance of the suspensions in mks acoustic ohms Rus = mechanical resistance of the suspensions in mks mechanical ohms As we shall demonstrate in an example shortly, these quantities may readily be measured with a simple setup in the laboratory. It is helpful, however, to have typieal values of loudspeaker constants available for rough computations, and these are shown in Fig. 8.5 and in Table 8.1. The magnitude of the air-gap flux density B varies from 0.6 to 1.4 webers/m? depending on the cost and size of the loudspeaker. TABLE 1, Typical Values off, Rg,and Mac for Various Advertised Diameters ‘of Loudspeakers T Advertised} Nominal) te, | ic, mats of diam, in, | ™Pedanee, | b™ | chins | voice coil, g in he | \ - 5 3.2 27 3.0 | 035-04 es aa | aa | ao | os-o7 wae | aa fad} ao | coats 2 so | 80 | 70 a wie | 160 ; 2 Values of Radiation (Frontside) Impelance. Acoustical elementsalways ive the noweomer to the fielt nf acoustics some difficulty heeause they are Part XIN; SIMPLE ENCLOSURES 215 "0 3% et] se fn ed 4 8 : Bus] 7 aa ” 21 : Poo 7 7 = s|— z = za s 5 i a onetsed dame niches ‘overied Sameer in cet (@ @) | | Cus ie | aha - EA : ai i EF aa 82 - ~_ Sess ees én dame nincnes ‘hovered date inne © (d) io, 880. Relation between efestve diameter of loudspeaker and ie averted Fo. 8.5), Average resonance frequencies of directradietor loudspeakers when ‘mounted in infinite bales ve. the advertised diameters, Fio, 8.8, Average mass of voice coils and diaphragms of loudepeskers as a function cof advertised diameters, Myo is the mass ofthe diapheagea including the mass of the Yoice-coil wire, and Myp is the mase of the diaphragm excluding the mast of the voice-eoil wine to. 8.5d. “Average compliances of suspensions of loudspeakers aa a function of adver- ited diameters. - Note, for example, that Son the ordinate meane3 x 10"'m /newton, not well behaved. ‘That is to say, the resistances vary with frequency, and, when the wavelengths are short, so do the masses, ‘The radiation impedance for the radiation from the front side of the diaphragm is simply a way of indicating schematically that the air has ‘mass, that its inertia must be overcome by the movement of the diaphragm, and that it is able to accept power from the loudspeaker. ‘The magnitude of the front-side radiation impedance depends on whether the box is very large so that it approaches being an infinite baffle or whether216 Loupsp KER ENCLOSURES Chap. & ‘the box has dimensions of less than about 0.6 hy 0.6 by 0.6 m (7.6 £0), in which ease the behavior is quite different. \VERY-LARGE-RIZED NOX (APPROXIMATE INFINITE BAFFLE) Sa radiation resistance for a piston in an infinite baffle in mks acoustic ohms. ‘This resistance is determined from the ordinate of Fig. 5.3 multiplied by 407/So. If the frequency is low so that the eflective cireumference of the diaphragm (2ra) is less than d, that is, ka < 1 (where k = 2x/X) Sax may be computed from Bag = 2 BU & Ooni5 (62) radiation reactance for a piston in an infinite baffie. Determine from the ordinate of Fig. 5.3, multiplied by 407/So. For ka <1, Xay is given by 0.27069. Xan = oar a 2a 880) and (8.30) MRDIUACSIZED BOX (LESS THAN 8 FT") Stan = approximately the radiation impedance for a piston in the end. of along tube, ‘This resistance is determined from the ordinate of Fig. 5.7 multiplied by 407/So. If the frequency is low so that the effective circumference of the diaphragm (2ra) is less than A, Rae may be computed from San = 0.01076" a) X4q = approximately the radiation reactance for a piston in the end of along tube, Determine from the ordinate of Fig. 5.7 multiplied by 407/S». For ka <1, Xan is given by Xun = aM and (0.1952)e0 Mar (8.50) Closed-box (Rear-side) Impedance. ‘The acoustic impedance Zap of @ closed box in which the loudspeaker is mounted is a reactance Xx in series with a resistance Ran. As we shall see below, neither Xz» nor Tas is well behaved for wavelengths shorter than 8 times the smallest dimen- Part XIX] sim CLOSURES. 217 sion of the box.t If the dimension behind the loudspeaker is less than about. A/4, the reactance is negative. If that dimension is greater than A/4, the reactance is usually positive if there is absorbing material in the box so that the loading on the back side of the loudspeaker is approximately that for an infinite baffle. meptun-sizen vox, For those frequencies where the wavelength of sound is greater than eight times the smallest dimension of the box GL <} for the box of Fig. 8.2), the mechanical reactance presented to the rear side of the loudspeaker is a series mass and compliance, 66) where as Cu Po (8.7) is the acoustic compliance of the box in meters per newton and Mas = Bet 68) is the acoustic mass in kilograms of the air load on the rear side of the diaphragm due to the box; and where Vo = volume of box in cubic meters. ‘The volume of the loudspeaker should be subtracted from the actual volume of the box in order to obtain this number. Toa first approximation, the volume of the speaker in meters? equals 0.4 X the fourth power of the advertised diameter in meters, 4 = 1.4 for air for adiabatic compressions. Po = atmospheric pressure in newtons per square meter (about 10° on normal days). xa = V/Scr if the loudspeaker is not circular. B = a constant, given in Fig. 86, which is dependent upon the ratio of the effective area of the loudspeaker diaphragm Sp to the area L? of the side of the box in which it is mounted, ‘As an example, assume that the depth of the box L/2 is 1 ft. ‘Then, since Eq. (8.6) is restricted to the frequency region where 2/8 > L/2, the maximum frequency for it is 140 eps, LARGE-SIZED nox. If the box is large so that its smallest dimension is greater than one-cighth wavelength, and if it is unfined, the mechanical reactance is determined from Fig, 8.7. Impedance of Closed Box with Absorptive Lining. ‘The type of reactance function shown in Fig. 87 is not particularly desirable because of the very + At 1000 eps, a wavelength nt 72°F isabout 13.5; at 500 eps, 27 in.; at 2000 eps, 6.8in.: and so on218 OUDSPEAKER ENCLOSURES, Chop. & a T L Moss oodingfeetor oa} —— UH A oot 008 on on 10 (Saft Fis. 86, Bnd-eorrection factor B for the reactance term of the impedance st the rear ‘ide of the loudspeaker diaphragm mounted in a box of the type shown in Tig. 83. ‘The acoustic weattanee of the box on the diaphragm is given by Xue = oP e/a “Bre/ra. Fors noncicular diaphragm of area So, xa = VSor. &. Eos es Eo 4 2-05] & soto é z Eng OY é 2-20 om” aos oi a 10 5 ‘Square rot of piston area per unit wave length (ySiy/A=/VSn/e) ‘occurs when L/2 = »/2, that is occurs at VSo/A = 0.25 for L4/Sp = 16; at 0.839 for L4/Sp = 9; at 0.5 for L*/Sp = 4; and at WSp/d = 1 for L*/Sp = 1. Part XIX) SIMPLE ENCLOSURES 219 high value that Xx reaches at the first normal mode of vibration (resonance) for the box, which occurs when the depth of the box equals one- half wavelength. A high reactance reduces the power radiated to a very small value. To reduce the magnitude of Xap at the first normal mode of vibration, an acoustical lining is placed in the box. This lining should bbe highly absorptive at the frequency of this mode of vibration and at all higher frequencies. For normal-sized boxes, a satisfactory lining is a 1in.thick layer of bonded mineral wool, bonded Fiberglas, bonded hair felt, Cellufoam (bonded wood fibers), ete. For small cabinets, where the largest dimension is less than 18 in., a }¢-in-thick layer of absorbing. material may be satisfactory, ‘At low frequencies, where the thickness of the lining is less than 0.05, wavelength, the impedance of the box presented to the rear side of the diaphragm equals Ban = Ran + jXae (8.9) where Xs is given in Eqs. (8.6) to (8.8) and Ruy» ——__Js_____, (6.10) wcutent + (14 te + Te) Ray = Ry/35u = one-third of the total flow resistance of a layer of the acoustical material that lines the box divided by the area of the acoustical material Sy, The units are mks acoustic ohms. ‘The flow resistance equals the ratio of the pressure drop across the sample of the material to the linear sir velocity through it. For lightweight materials the flow resistance I is about. 100 mks rayls for each inch of thickness. For dense materials like PF Fiberglas board or rockwool duct liner, the flow resistance may be as high as 2000 mks rayls for each inch of thickness of the material, For example, if the flow resistance per inch of material is 500 mks rayls, the thickness 3 in., and the are: 0.2m?, then Rav = 1500/(3)(0.2) = 2500 mks acoustic ohms Tt is assumed in writing this equation that the material doe: not oceupy more than 10 per cent of the volume of the box. Vy = volume of the box in eubie meters including the volume of the ‘acoustical lining material. Vu = volume of the acoustical lining material in cubie meters. Graphs of Ey. (8.10) are given in Fig. 8.8. Ab ull frequencies where the absorption coefficient of the lining is high (say, greater than 0.8), the impedance of the box presented to the bacl side of the diaphragm will be the same as that presented to piston in at 4 Tubles and graphs of absorption coefficients for common materials are given iy Chap. 10.20 LOUDSPEAKER ENCLOSURES Chap. 8 infinite baffle radiating into free space, so that Zs is found from Fig. 5.3 or Eqs. (8.2) and (83). Acoustical material may also be used to enlarge effectively the volume of enclosed air. Gaseous compressions in a sound wave are normally adiabatic. If the air space is completely filled with a soft, lightweight ani 20" Fic, 88. Graphs for determining the acoustic resistance P, as.n function of Raw with Cao as the parameter. The four charts are for different ‘atiog of the box volume to the volume of the acousticw) materi material such as kapok or Cellufoam (foamed wood fibers), the compressions become isothermal. This means that the speed of sound decreases {rom ¢ = 344.8 m/sec toc = 292 m/sec. Reference to Fig. 8.7 shows that this lowers the reactance at low frequencies just as does an inerease in box dimension L. ‘This also means that in Eq, (8.7) the value of 7 is 1.0 instead of 1-4 Part XIXL SIMPLE ENCLOSULES 221 One special type of closed box with an absorptive lining is shown in Fig. 8.9 Here, the loudspeakeris mounted near the end of a rectangulat box of length L and of cross-sectional area approximately equal to ma’, the effective arca of the loudspeaker diaphragm. A glass fiber wedge whose length is L/2 is used to terminate the box. ‘The specific acoustic resistance @ Ran and reactance Xx (multiplied by So/ece) for such a box with and without i the wedge are given in Fig, 8.10. ee ‘Three things of importance are ob- Scostea mara! served about the impedance: (1) for a °"#! mal xziven volume of box, at low frequencies, isis fber me the reactance Xap is smaller than that Snlmon gee for the box of Fig. 82; (2) at high 825t40}/t! frequencies, the box resonances (normal Fis, 89. Special trp of exbinet for tmovtes of vibration) are damped out so [inimiing the shift of resonance that Ras approaches gxe/So and Xa» ina closed:box bale, approaches zer9; and (3) between these frequency regions [that is, 0.2 184 see, Let us see what this means in terms of Qr. 1 Fquation (8.27) is the same es that derived in Example 7:1, p. 194 except for a factor of 2 in the denominator. This factor of 2 expresses the difference between ‘radiation into full space as compared to radiation into hal space (infinite bale ess).In terms of Qrthe suggested criterion for satisfactory transient response a oalls So Or Re S184 (8.30) As an example, if wo = 2rfo = 2r40 = 250 radians/sec, then Qr should be less than 1.36. This would mean that the peak in the response curve 20 10] 8 ‘Sound-Pressure-Level Response tid at ity on 04 19 4 10 wy Fic, 8:12, Sound-pressureevel response of a direet-radiator loudspeaker at low frequencies. An infinite bam or a closed:box enclosure ie assumed. Zero decibel is referred to the reference sound pressure defined by Eq, (827). Qr isthe came as Qr of Bq. (823) and ws ie found from Bq. (6.22). The graph applies only to the frequeney Tange where the wavelengths are greater than about three times the adver- Used diameter of the diaphragm. must be less than 3 db. Methods for achieving desired Qr's will be discussed as part of the example below. For ease in design of direct-radiator loudspeakers, Eq. (8.28) is given in decibels by Eq, (8.31) below. Equation (8.31) is plotted in Fig. 8.12. It should be remembered that Eqs, (8.31) to (8.33) and Fig. 8.12 are valid for either the small closed box or for the infinite baffle, depending on how ax and M.. are chosen. 20 logss z Be oe 20 logis — rotoge[ ds . (2 = s)] ab (831) F ENCLOSURES. 227 We should observe that, even in the frequency range where the diaphragm diameter is less than one-third wavelength, the value of Qr is not Strictly constant because Qe» increases with the square of the frequency. In using Bq. (8.31) and Fig. 8.12, therefore, Rin Q- probably ought to be calculated as a function of w/se. Usually, however, the value of Ra at os is the only case for which calculation is necessary. Radiated Sound Pressure at Medium and High Frequencies. When the frequency becomes large enough so that the diameter of the diaphragm is greater than one-third wavelength, Eq, (8.28) may no longer be used. If the diaphragm still vibrates as a rigid piston, we may find the ratio of |p| to [pel by taking the ratio of Bq. (8.17) to Bq. (8.27), using Eq. (8.25) for [U,| and assuming that we are above the first sesonance frequency 80 that w/e > on/e. This operation yields ~ 49aador P| = |e (8.32) os | — where on is given in (8.22) and Ma = Man + ot 633) Be Ra Eggs + Rast 2en ean Q = directivity index for a piston in the end of a tube (see Fig. 4.20) Equation (8.32) expressed in decibels becomes, 20 108 | 2 Ip. 10 logue Rta + DI + 10 loge 26 — ror (Fi, , “) (8.35) where DI = 10 logis Q = directivity index for a piston in the end of a long tube (see Fig. 4.20) Oleg! = 38:7 ab at normal rum endo Inspection of Eq. (8.35) shows that the sound-pressure response |p| of the loudspeaker relative to its reference response |pcl is a function of frequency, effective radius of the diaphragm, and the ratio Ra/M- For convenience in design, Eq. (8.35) is plotted in Fig. 8.13 for four values of advertised diameters of loudspeakers having effective radii of 0.0875, 0.11, 0.13, and 0.16 m, respectively. ‘These graphs join on to Fig 8.12 at frequencies where w/ is large compared with o/s. 8.6. Location of Loudspeaker in Box. ‘The results shown in Fig. 8.7 for the reactance of the closed box apply to a loudspeaker mounted in the center of one of the L by I sides. This location of the loudspeaker leaves something to be desired, because waves traveling outward from the dia228 LOUDSPEAKER ENCLOSURES (Chap. 8 phragm reach the outside edges of the box simultaneously and in combina tion set up a strong diffracted wave in the listening space. ‘To reduce the magnitude of the diffracted wave, the loudspeaker should be moved off center by several inches-—preferably in the direction of one corner. 10 —a Tage me ey i _| She + [piss | 15" dia e 12" dia c -29 -25 100 200300500 FO 100 100 e700 0 Frequency in cycles per second Pro. 8.13. Graphs for ealoulating the relative frequeney response in the Frequency region where the loudapeaker is no longer a simple source, but below the frequency of the rim resonance. The parameter i Mf4/Rq. These charts cover the frequency ‘ange just above that covered by Fig. 8.12, ‘The front face of the box of Fig. 8.2 need not be square. Tt is possible to make the ratio of the two front edges vary between 1 and 3 without destroying the validity of the charts, for the same total volume, 8.7. Measurement of Loudspeaker Constants. The constants of a Joudspeaker are not difficult to determine if a-c and d-c voltmeters and a calibrated variable-frequeney oscillator are available. Part XIX] SIMPLE ENCLOSUR 220 During the tests, the direct-radiator loudspeaker is hung without baffle in an anechoie chamber, oF outdoors, or in a large room that is not too reverberant, ‘The various constants are then determined as follows: Measurement of ws. A variable-frequency source of sound with an ovt- put impedance greater than 100 times the nominal impedance of the loudspeaker is connected to the loudspeaker terminals. A voltmeter is then connected across the terminals, and the frequency is varied until a maximum meter reading is obtained (see Fig. 8.14). From Fig. 7.2, we see that maximum electrical loudspeaker impedance corresponds to maximum mechanical mobility, which in turn occurs at the resonance frequency fo OF we Measurement of Rs. The electrical resistance of the voice coil is measured with a d-c bridge. Beat frequency telson Lowdepeator Fic. 8.14. Circuit for determining the resonance frequency of a loudepeaker. ‘The voltage Vis held constant, and the oscillator feequeney is varied until V;ismaximurn Measurement of Q. The Q of the loudspeaker is found by determining ‘the width of the disphragm-velocity curve 1, plotted as a function of frequency for a constant current i (see Fig. 8.14). Because ¢ = Blu, we may determine the shape of the velocity curve by plotting against frequency the quantity e=Vi-Riav—4 ‘The width, af eps, is measured between the points of the curve on either side of the resonance peak where the voltage is 3 db down (0.707) from the peak voltage ‘The value of @ for a constant current é is given by fa o-% (8.36) Measurement of Mu». To find Mun, we must determine the total mass of the diaphragm, including the air load, and subtract from it the air-load mass, A known mass is added to the diaphragm, and a new resonance frequency is determined. ‘The added mass is usually a lead rod or a roll of solder bent into a circle of 3 to 6 in. diameter and taped to thediaphragm 80 that it does not bounce. Lead is used because it is nonmagnetic,(Chap. & If the original resonance frequency was w and the resonance frequency after addition of a mass M’ kg is w,, then coeee Ville F ManyCus we (837) and = (838) Vile + Man FM Cus where Cus = mechanical compliance of the suspension in meters per newton Mary = 2,67a%p = 3.150? = air-load mass in kilograms on the two sides of the diaphragm Simultaneous solution of (8.37) and (8.38) yields Muo+ Mar = (8.39) Determination of Rus. The mechanical resistance Rws of the diaphragm suspension plus the air-load resistance is determined from the data above. From circuit theory we know that, su(M ary + Mans) a where Stve = 0.1886apuat/ct = (8.45 X 10-*)alf* mechanical air-load resistance in mks mechanical ohms for both sides of an unbaffled loudspeaker. This resistance generally can be neglected. ‘Measurement of Cus. ‘The mechanical compliance of the suspension may be determined by adding known masses to the cone and measuring the steady displacement when the axis of the cone is vertical. A depth micrometer is mounted rigidly with respect to the loudspeaker frame to accomplish this. ‘That is to say, Rue + Run = (8.40) displacement (m) Cus = iilded weight (newtons) (at) ‘The added weight is, of course, the added mass in kilograms Af’ times ‘the acceleration due to gravity, ie., Added force (weight) = = Mig = 98M" — newtons (8.42) ‘The compliance Cvs may also be determined from Hq. (8.37) if the masses Myo + Mu, and the angular resonance frequency wy are known, In this ease the depth micrometer is not needed Measurement of Bl. ‘The value of the electromagnetic coupling constant BI is determined by using an added mass and the depth micrometer described above or a bent piece of wire fastened to the frame of the loud: speaker to mark the initial position of the disphragm. When a mass A 231 is added to the diaphragm, a downward displacement occurs. To restore the diaphragm to its original position, a steady foree f, is needed If a direct current é is passed through the voice-coil winding with the proper direction and magnitude, the required force f, will be produced as follows: fom Bi (643) ‘The displacement of the diaphragm was produced by the weight of the added mass (see Eq. (8.42)]. ‘The diaphragm will return to its zero position when fe Solution of Eqs. (8.42) and (8.43) simultaneously gives (44) (8.45) Measurement of So. The effective area of the diaphragm can be determined only by coupling its front side to.a closed box. ‘The volume of air Vs enclosed in the space bounded by the diaphragm and the sides of the box must be determined accurately. Then a slant manometer for measuring air pressure is connected to the airspace. The cone is then displaced a known distance £ meters, the manometer is read, and the incremental pressure p is determined. Then, Po 7 tSe a Yep Pe where Ps is the ambient pressure. The pressures Py and p both must. be measured in the same units, and Vo/& should be determined in square meters, Usually, Sp can be determined accurately enough for most ealculations from Fig. 8.54, that is, Sp = rat 8.8, Measurement of Baffle Constants. The constants of the baffle ‘may be measured after the loudspeaker constants are known. Refer to Fig. 84. ‘The quantities Stax and Xaz are determined from Eqs. (8.4) and (8.5). ‘The electrical and mechanical quantities are measured directly. Measurement of Xs». Using the same procedure as for measuring wo in Par. 8.7, dotermine a new wy and solve for Xs from Fig. 8.4. Measurement of Ras. Using the same procedure as for measuring Q in Par. 8.7, determine a new @ and solve for Ras from e+ Mar) GE A (Ras + Ran) So mt (8.46)232 LOUDSPEAKEM ENCLOSURES (Chap. 8 where Map = Muro/Sat; Maris given by Bq. (8.5), and Rus + Ste, Ras + Ran = “EES ‘Example 81. Direct-radiator Loudspeaker in Infiste Baffle. Determine Qr and urs for the loudspeaker of Pig. 78. ‘Solution. ‘The advertised dinineter of that loudspeaker is 8 in., and from Fig. 8.54 swe find that this corresponds to an efective diaphragm radius of @ = 0.088 m and an Hletive area S» ~ 0.0245 m4, Hence, Fig. 8.12 is valid for this loudspeaker for ‘wavelengths down to 0,6 m, or frequencies as high as about $50 ps. The diaphragm tn this loudspeaker ceases to vibrate asa rigid piston st about 400 eps so that theory find measurements would be expected to agree well only up to shout 200 ens. ‘By superimposing the measured data plotted in Fig. 7.8 on the contours of Fig. 8.12, ‘we find thet Qp = 6, Wealao observe that we = 565 radians/sec. From Fig. 85e me Bind that Myo * 0.0082 kg and = 7.05 keg/m* "The loudspeaker was in an infinite baffle that, from Eq. (83), _ 0.20.38) _ fms May = DOZENS) 2 7.25 kg Hence, Ma = May + 2Mar = 143 kg/m Prom Bq. (8.23) we have fey = ool _ (565)(143 4 Or 6 = 1846 mks acoustic ohms ‘The endiation resistance 2Ran at wy = 565 radiang/se (fo = 90 eps) should equal see Bq. (21 29tar = (2)(0.0215)(90)* = 348 mks acoustic ohms 1: contributed by the suspension at the resonance frequency is Bw Ras = Ra ~ Mae ~ OE RTF From Table 8.1, Re was approximately 3 ohms. ‘The loudspeaker was operated from. fan amplifier with an output resistance of 12 ohms, and Bt was approximately 2.5, ‘webere/m. So ‘The dat Bs rat = 700 mks acoustic ohms a Hoonee, Ras = 1846 ~ 848 — 700 = 208 mike acoustic ohms interesting to observe that Qr its function of frequency because Ra equals Ry = 298 + 730 + 0.0134¢ For example, Qr at 300 ep equals «sssna (14.3) Or ee = sag = aot XIX! SIMPL, NELOSURE: 233 Reference Uo Fig. 8.12 shows that the response at $00 cps eatentially is the same for Or = 1.62 as for its value of 6 at 90 eps ‘Example 8.2. Wide-frequency-range Loudspeaker. A Joudapeaker designed to be operated without a tweeter anit has 8 nominal diameter af 123. and the physieal constant given below 4 = 13.em (effective radius) Cus = 2. 10"+m/newton Mup = 12 BR, = 2ohms Ry = 8 ohms B = 1 weber/m® ta iom Rus = 1 newton-sce/m ‘Sp = xa" = 0.053 m* Sot = 0.00281 me «8, Determine the percentage shift in the frst resonance frequency of the loudspeaker {rom the value for an infinite bafe if a box having ditnensions 1.732 by 1.792 by 1 ft = 3 fis used, , Determine the height of the resonance peak, assuming Ras = 0. ‘. Determine the dimensions for a baffle of the type shown in Fig. 8.9 that will cause a shift in infnite-ha le resonance frequeney of only 10 per een. id. Determine the height of the resonance peak for the bale of (), ‘Solution. a. Fist, let us determine the value of the acoustic-radiation masses for the loudspeaker in an infinite baffle, From Eq. (@.2), O5ie _ (0511.18) May = ei = 49 kg/m Also, Cas = CusSot = (@X 10-9281 X 10" = 5.62 % 10"? m¥/newton Mun _ 0012 ‘Bot ~ O.OmDST “The resonance frequency fo the speaker in an finite ba is found from a. (8.1). 1 Be VEe2 KID FAO) 10 = 08 oa vous “°° ‘rom Eq. (87) we fad the formula for Cas Vv _ @yzss2 x 10-5 Taxi 6.06 X 10-1 mt/ncwton Mav 4.26 ke/mt f the resonance Frequency234 LOUDSPEAKER ENCLOSURES iChawy. 8 ‘That is to say, the new resonance frequency is 48 per cent higher; 20 fa = (1.48)(70) = 108.5 eps , The height of the resonance pes: is to be found from Eqs. (8.28) and (8.29). The total resistance ys given by Eq, (819) for low frequencies. Using Eq. (8.4) and fo = 103.5 ps, we determine the radiation resistance atthe resonance frequency. Stax = 0401076 = (0.01076)(108.5)* = 116 mks acoustic ohms = 356 mks acoustic obma 10% Ra = Goyboney + 856 + 115 = 4090 mks acoustic ohme ‘The value of Mf ia given by Ba. (8.20). From Eq. 8.5), __ (0.1952)(1.18) Ma 52018) = 4.779 kg/mt From Ba, (8), (ossyct.8) _ Man = Cor = 240 ke So Ma = 426-4246 41.77 = 849 kg/m* de a eels. HBA) Ta 20 logue 2, = 20 logie 1.87 = 2.7 db ‘The resonsnce peak lies about 9 db above the reference pressure level. Because Qp is greater than unity, the reference pressure isthe pressure in the flat region © hove the resonance frequency, as we ean see from Fig. 8.12. ‘. For only & 10 per cent shift in resonance frequency, fe = 70 X 11 77 cps From Bq. (8.14) we find that Cas_ Ut Can ag — 1 = O07 502 10 : Can = SSAA 0 8.04 x 10 "To achieve this ratio, the volume of the box either must be quite large, or else we sight use the arrangement of Fig, 89 a8 Was asked for in the statement of the problem "The desired normalized reactance is Xap, _—So___=(10* x 0.058) pe Canoes ~ WBA X BOK X 407 ~ ~00884 From Fig. 8.10 we soe that for this kindof baflea vertical length is required of 02 x 5448 Lom 3 = 0.9, oF 35 in If the 12-n.wide loudspeaker ia to St into a crose-sectional aren of 0.058 m?, the side that itis mounted in will need to be about 13 in. = 0.38m wide Thisgives us, for the other side, 0.053/033 = 0.16 m = 6.34 in Part XIX SIMPLE ENCLOSURES 235, 14 The height of the resonance peak ia found from Qr To determine Qe, we nocd Ra and Xe Rey = 4090 +R We find Ran from Fig, 8.10, (025) _ (4100025) ‘0.088 = 1085 mks nvoustic ohms Raw = which gives us ‘Ry = 4030 + 1938. 6000 mks acoustic ohms My = 4264177 40= 603 kg oy wala = (4844672) z 000 In other words, this arrangement is nearly critically damped. At 77 eps, the reaponse will be down about 5 db from the reference response Example 83. Low-frequency Loudspeaker (Woofer). Many high-fidetity sound ystems employ two loudepeakers. One covers the low-frequency range, and the father covers the high-frequeney range, An electrical network is used to divide the ‘output energy from the amplifier into two frequency regions centered on the “eross- ‘over' frequency. Common crossover Frequencies are 500, 800, 1000, and 1500 eps. Tn this example, we have chosen a 15in. commercial low-frequency unit, designed to be used with « 800 eps erassover network, As no information was available from the manfacturer on the constants of this loudspeaker, they were determined expe ‘mentally, The loudspeaker was then put in a elosed-box bafle, and its frequeney Feoponce waa meacured for comparieon with calculations. The complete procedures tise in eareying out the experimenta are deseribed here, and the results are compared ‘with the computations in order that the reader may get an indication of the reliability of the method, ‘The loudspeaker was placed in closed bos with dimensions V = (Din. X 35 in, X 18 in.).039 X 10°) = O31 ms. “The other constants were determined by the procedures of Par. 87 Rg = 55 ohms Cus = 282 X 10 m/newton 2 = 0.16.m So = 803 X 10 m* So! = 6.45 X 10-t me Rus = 23 mks mechanical ohms Maca = 0.045 kg fu = 30 eps tat = 247 radians/see Bl = 25 webers/m ‘The volume ofthe box oeeupied by the loudspeaker was about 0.01 m?, so that 2058 rom theae quantities we determine the ses of the elements in Fig. 84 Cas = CusSut ~ 1.82 X 10-4 m¥/newton Rue Rag = RE 350 mks acoustic ohms Ma . Wao = MEP = 6.98 kam Bu BE = 9.7 x 10¢ webers}mt236 LOUDSPEAKEH ENCLOSURE [Chap 8 A B:in-thick gound-ahsorbing blanket with a flow resistance of 2000 mks rayls/in, af thickness is placed on the 18- by 35-in, wall ofthe box to reduce the efeeta af stand ing waves at higher frequencies. This material yields a value of Ras that varies DDetween the values given by Iq, (8-10) at low Frequencies to Ban for aa infinite bute st 500 cps. If the amplier resistance is small, Rae may be neglected at low frequencies. Values of ean, caleulated from Bq (8:10) at low frequencies and estimated at higher frequencies, are shown in Table 8:2. At-very low frequencies, for exaimple, 40 eps plywood tox absorbs sound hecause of the frictional losses in the wood as it vibrates, TABLE 82. Determination of Rx T Part XIX1 SIMPLE ENCLOSUNESS 337 I 10 o> Taras 7 Vesna 7S ner Je = 80 ep perenurwarion oF Qr: Ma _ 18M Or TR Re Values of Or, Ma/Re (for > 150 eps), and a /ay are tabulated in Table 8:3. Also shown in Table 8.3 are the valves of the directivity index as determined from Fig. 4.20 for « piston in along tube eee Ry = 190 obme TABLE 8.3. Determination of Ma, Qr. and Ma/Ra f | | Fe] ka | Ras | Rant | Raw . 7 cps x ae 20-150 | wrkoss| ™ |i + RASH) u fe em ee ee was ba lo.oss2-o.sao) o.sse | 0727 | ors | ries | 1.455 20| "0 0582) 356 4] 00] 11,400 | 12,000/ 716 | 1380 Mao 6 | 69s | 69s | 6s | cos | cos oo] 0.201 | 356 | 108 | “@3 } 11400 | 12,000) 746 | 1370 Man ry te es es | ne | eS 50} 0.437 | 386) 242) 158] 11,400] 12,100) 716 ‘| 1470 Mae ra | ovat | ovae | ovae | var | roe 200] 0.382 | 356) 430 | 205] 11,400 | 12,500! 736 —_| 1800 Mal} swat | volar | soar | wer | ost | 9.08 250) 0.728 | 356 72} 465 | 11,400 | 12,990) 716 | 2210 ——! 500] ogre jasn| se | 775) 11.400. | 13,500) 716 | 2820 R as a ka > —|— _ yoo} "1 164 | 390 | 1622} 1sz0/ 11,400 | 14,800 | 716) an60 3 = Eenwliniwn = S00} 1455 | see | 2040 | 2640| 1,400 | 17,000) 716 | 6380 O72 x1] BRIE x OY TR Ox Tor ba © 1, Saw = 0010767. 5.7 x 10-44 6 x 1048.6 x OH? 5 x 10] 6 x 10 For example, at 40 eps, the box might be expected to abeorb one-half as much encray 10.0 per square foot of aren oa the acoustical material placed inside, By 80 ps, however, _ {= _ the acoustical material will aboorb many’ times more than the box. Hence, for the 40 ‘case of the closed box we may neglect the losses in the wood except at the resonance frequency (near 50 cps), where we may estimate io the calculations the effect by tic pling the actual amount of acoustical materia "The quantities Ran and Max (see Pig, 8.1) are found from Fig. §.7 or, for ka <1, Pas. (8A) and (83) “The quantity Ma for ka <1 is found from Faq. (88) and Fig. 8.0, For our box, the quantity Sp/L* ~ 0.0803/0.618 = 0.118. From Fig. 86, we see that B ~ 0.788. ‘The value of May for a > Lis equal to0.92p4¢/Spu times the ordinate of Fig. 5.3. DETERMINATION OF Ray Ma, AND Ca. Those quantities ae given by Tig. (8.10) to (621), Por this example, calculated values of Ry and My are given in Tables 82 snd 8.3, reapectively. Ca is found as follows: a + mt/aewton = on ais — B10 tract a = ssh 182 XDI 10 * Gg b Can” C822 XA = 98E x 107? mY/oewton Can perenaiartow or us wiritaox. The value of we with the closed-box enclosure is from Table 82 and Ca given above, [CALCULATION OF PREQUENCY-RESTONSE CURVE. Theda of Table 3 permit the calculation ‘ofthe relativesound-pressure level response curve fora microphone poston onthe principal axisof the loudspeaker. Figure & 12 4 ued forthe frequency eange below 15D eps and Fig. 6.13 for 200, {psandabove. Theresulsofthecakeulatonsareshown inFig. 8.15. Caleulationsforintermesi- ate resstances ae alto shown. The value of R, = 3 ohms corresponds to constant voltage sross the terminal. R, — 14 ohms ie equal to the nominal impedance ofthe loudspeaker. R, = 20 ‘ohms equivalent to changing B/by a factor of. R, = 130 ohms is equivalent to changing 8 bya factor of 5. ‘Each curve has its own reference pressure, However, to show the effect on the relative level of the reeponse curve (for constant e,) ns the amplifier impedance Ie, was variod, the reference pressures (in decibels) are shifted relative to each other by ference I minus reference 2) = 20 logis Rett Be (Reference 1 ference 2) = 201oge RE RE ab Ifthe power-availuble-ofcieney response (PAE) ia desired, subtruct the directivity, index (DI) of Table 82 from the ealeulated response curves of Fig. 8.13. Tie gives the shane of the PAlig curve298 LOUDSPEAKER ENCLOSURES [Chap. 8 erasuneo rarguesicr-nnsronse CURVE. ‘The frequency-response curve was meas- ‘ured in en ancchoie chamber using a S-obm 50-watt anmplicr as the source, The power available was about 1 watt. The data were taken at a distance of 7 fon the principal axis of the loudspeaker. To be in the far-field of the loudspeaker a distance » (a) | AN tay r ‘i ) Ret 130) oa 1 BH 3 2970 raquacyin les pr sezand sont 1c. 8150 and b, Graphs of the on-ais ound pret level produce by lowte- Gocney uectaduator loudspeaker below the frequency af tim resonance. The ire carves are ealculatad’” Solid curves with porn are massed. Graphs 8#e plated elaive to the fefetence premre for f= 8 china Diferencs betyeen Eetnions aad data st fequencis above 100 cp are duet coe reronances The Sim reeonancerceurs at about 900 ep seater than 7ft would heve been desirable, bt the mal sn of the avallableanechoie Shumber prevoted this, ‘The dats for Ry ~ 8,34, 30, and 90 ohn with «constant “alue ofthe open-losit ample voltage care also shown by th crn and atid Tne in Fig 818 "The agteemeat between theory and experiment at frequencies Below 100 ope i ccalety winch indicates that te diaphragm, waa vibrating at «rig son. Part XX) BASS-NEFLEX ENCLOSURES: 239 [Retween 100 and 500 eps the loudepeaker radiates more sound along the principal saxs than the theory predicts, assuring the directivity index for 8 1Sin. piston in & long tube Polar plots of the sound pressure level were also made with the microphone held in one position and the loudspeaker (and baffle) rotated, The measured directivity Index is compared with the theoretical in Table 8.4. The differences between the TABLE 84. Comparison of Measured DI for a 15-in. Loudspeaker in a Closed- box Baflle to that for a 15-in, Piston in a Long Tube. Comparison of the DI flerences is Also Made with the Differences between the Curves of Fig. 8.15 Directivity index, db Differences in DI | Differences from Fig. 8.15, db sirectivity indexes are also compared with the diflerences between the theoretical and measured curves of Fig, 815. Comparison of these curves shows thnt not only was the loudspeaker more directive than expected but it also radiated more poser by the amount indicated ia the fifth columa. The increased directivity is aasumed to be ‘due to the conical shape of the diaphragm, whereas the increased power output is attributable to the breakup of the cone into higher-order resonances, A sunilur diflerence between measured and calculated curves was also found for the in. loudspeaker of Fig 78. At about 500 cps, the major rim resonance ocours, with the expected sharp dip in response, As he loudspeaker was not designed to operate above $00 cpa, no data were taken beyond that point. part XX Bass-reflex Enclosures 89. General Description. ‘The bass-reflex enclosure is closed box in which an opening, usually called the port, has been made.*-* The area of, the port is commonly made equal to or smaller than the effective area of +A. L. Thuras, “Sound Translating Device,” U.S. Patent No, 1,869,178, July, 1037 (Ged 1930) HF. Olson, “Elements of Acoustical Bnginecring,” 24 ed., pp. 164-186, D. Van Nostrand Company, Ine., New York, 1947, *D. J. Plach and P. B. Williams, Loudspeaker Enclosures, Audio Engineering, 36: 12g, (Buly, 1951) J. J. Baruch and H.C. Lang, “Some Vented Enclosures for Loudspeakers ” Unpublished report of Acoustics Laboratory, Massachusetts Institute of Technology, Cambridge, Mass, January, 1953210 LOUDSPEAKER ENCLOSURES Chap. 8 the diaphragm of the driving unit, A common construction of this type of loudspeaker is shown in Fig. 8.16, When the diaphragm vibrates, part of its displacement compresses the air inside the box and the remainder of its displacement moves air outward through the port. Thus the port isa second “diaphragm,” driven by the back side of the loudspeaker diaphragm. The port is, at low frequencies, equivalent to a short length of Box with compliance eed as oases totem rt ex ms = eecneane saraencana Fie. 816. Baserefler bale. The port has an area S,, and the diaphragm has an rea Sp. The inner end correction for the tube is included in the magnitude of Bar. tube with an acoustic reactance and a series acoustic resistance. This tube has an end correction on the inner end and a radiation impedance on the outer, or radiating, end. We shall assume for the remainder of this analysis that ka < 0.5. In other words, we are restricting ourselves to the very low frequency region where the radiation from both the port and the loudspeaker is nondirectional. Fic. 8.17, Analogous acourtial rcut for «loudspeaker box witha port. ‘The wal tine vloity of the diaphragm Uthat of the port in Uy and thet of the ox 0 8.10. Acoustical Circuit. The acoustical circuit for the box and the port is given in Fig. 8.17. The compliance and resistance of the box are Can and Ran. The mass loading on the back side of the diaphragm is Mas. The mass and resistance of the air in the port that penetrates the side of the box, including the inner end correction, are Mar and Rar respectively. Finally, the series radiation mass and resistance from the front side of the port are, respectively, May and Quay. The values of these quantities are Mayas in Bq. (83); Ray asin Eq, (8.10); Canasin Ee (8.7); Marasin Eq, (8.5), but with ayinstead of a thatis, Max = 0.28/¢si Rane as in Ba. (8.4); and Pact XX} BASS-t LEX ENCLOSUK oat Mar = (C+ 00a.)po/ra;* = acoustic mass of the ait in the port it Kilograms per meter’. This quantity includes the inner end Rap = aconstie resistance of the ain the portin mks acoustic obms [Sce Eq. (5.54). Use the number (1) in the parentheses.) fu = Density of air in kilograms per eubie meter (normally about L18 kg/m) lective radius in meters of the port in the vented enclosure. IF the port is not circular, then let as = -/8,/x, where Sy is the effective orea of the opening in square meters S, = ra;* = effective area of the port in square meters. 1 Tength ofthe tube or the thickness ofthe wall ofthe enclosure. in which the port is out in meters In ease the port is comprised of a number of identical small openings or tubes, the following procedure is followed: Let NV equal the number of such openings in the enclosure. For each ‘opening the acoustic mass and resistance including Mas and Qtans are Ma = (6+ 1.7a3)p0/(rast) kg/m! [see Eq. (5.57)] ‘Ry = acoustic resistance of each opening in mks acoustic ohms {see Eq. (6.50) as — effective radius of each opening in meters ‘The total acoustic mass and resistance for the N identical openings are Mar + Mar = Ma/N kg/m* Saas + Rar = Ra/N mks acoustic ohms ‘The directivity factor for a group of holes is shout equal to that for a piston with an area equal to the area within a line circumscribing the entire group of holes. 8.11, Electro-mechano-acoustical Circuit. The complete circuit for a loudspeaker in a hass-reflex enclosure is obtained by combining Figs. 84 cman | Mtaniget i SN Rp en oa Mo Gigli Ue Way By Mae Fie, 8.18, Complete clectro-mechano-ncoustiealcirevit for a bass-refex loudspeaker ‘The total force produced at the voice eoil by the electric current is pSp, where Sp is the area of the diaphragm. The volume velocity of the digphragm is U/, and that fof the port is Uy. The box is assumed to be unlined. If the box is ined with a sound absorbing material, s renstance must be inserted in sores with Mag. Note that Mar includes the inner mass loading for the port and 8.17. ‘To do this, the portion of the circuit labeled “‘box’* in Fig. 84 is removed, and the circuit of Fig. 8.17 is substituted in its place. The resulting circuit is shown in Fig, 8.18.22 LOUDSPEAKER ENCLOSURES [Chap 8 SUMMARY OF BASS-REPLEX DESION 1. To determin the alum ofthe box and the typeof port to goin 4a. Find the values of Myo, Cus, and Sp trom Par. 87 (pages 228 to 281) Approximate values may be obtained from Fig. 85. 2 Determine Man = Mivn/Sct and Cue = CwsSe « Balct the volume of baffle box you desire, making it renonably large, if posi bie, Calculate Cap fom Ba, (8) 4. Determine Map from Bq, (8), and compute Ma ~ 0.23/a, where o i the effective radian of the diaphragm in meter. Determine Ma trom Eq (8.51), and calculate May = MsCus/Cas. The quantity Ave isthe scoustic tomas of the port Tes composed of two parta May fod Mav 1. Devide on the aren of the port, Unvally the ares is Uetweon 0.5 and 1.0 tia Sp. Calculate May = 0.23/as, where aris the effective radius ofthe port in eters. Ifthe ports not round, ax = V/Sy/a, where Sis the port area in square eters Subtract May from May to get Map. {Determine the length & ofthe tube thet goes into the port from f= O852f4rSp ~ 0.60, A, Study Pars, 8.18 and 8.17 (pages 252 to 254) for construction, adjustment, and performance. 2. To determine the response ofthe loudspeaker af frequencies below the rim resonance {shout 500 eps): 12 Finil the element sites for Figg. (8 18) and (B19) from the text. . Determine the reference sound pressure at distance r from Faq. (8.64). «. Determine the values of Qs, Qu, and Qs from Fags. (8.56) to (8.50) and the ratio Cas/Can. 4. Detarmine the sound-prosgure response in decibels relative to the reference sonnd pressure at the three eritical frequencies from Figs. 822 to 8.24, The values, fof the critieal frequencies are found from Figs. 8.20 and 821 ‘The quantities not listed in the previous paragraph are ‘open-circuit voltage in volts of the audio amplifier. flux density in the air gap in webers per square meter (1 weber/m? = 10* gauss). length in meters of voice-coil wire. ‘output electrical resistance in ohms of the audio amplifier. electrical resistance in ohms of the voice col. ‘effective radius of the diaphragm in meters Moo/So? = acoustic mass of the diaphragm and the voice col in kilograms per meters. Cas = CusSot = acoustic compliance of the diaphragm suspension in meters* per newton, Ras = Rus/So? = acoustic resistance of the diaphragm suspension jn mks acoustic ohms, Mj = acoustic-radiation mass for the front side of the loudspeaker diaphragm = 0.195p/a kg/m*. Note that we assume the loud speaker unit is equivalent to a piston in the end of a tube Pact XX} BASS-REF 243 aes = 0.01076" = acoustic-radiation resistance for the front side of the loudspeaker diaphragin in mks acoustic ohms (see Fig. 5.7 for ka > 1.0). If the port is closed off so that U,, the volume velocity of the air in the port, equals zero, then Fig. 8.18 reduces to Fig. 8.4, At very low free ‘quencies the mass of air moving out of the lower openings neatly equal to ‘that moving into the upper opening at all instants. In other words, at very low frequencies, the volume velocities at the two openings are nearly ‘equal in magnitude and opposite in phase. 8.12. Radiated Sound. The port in the box of a bass-reflex baffle is generally effective only at fairly low frequencies. At those frequencies its dimensions are generally so small it can be treated as though it were @ simple source. The loudspeaker diaphragm can also be treated as a simple source because its area is often nearly the same as that of the opening, Referring to Eq. (4.3), we find that the sound pressure a distance r away from the bass-reflex loudspeaker is Ds pit ps = BE ert — Uys) 47) where p; and ps = complex rms sound pressures, respectively, from the diaphragm and the port at distance r. average distance of the point of observation from the diaphragm and the port. Note that r is large compared with the diaphragm and port radii. rand rs = actual distances, respectively, of the point of obser= vation from the diaphragm and the port. Uz = complex rms volume velocity of the diaphragm. U, = complex rms volume velocity of the port, Note that ‘the negative sign is used for U, because, except for phase shift introduced by Ras and Cs, the air from the port moves outward when the air from the diaphragin moves inward, Also, the complex rms volume velocity necessary to compress and expand the air in the box is Uns U.-U, (8.48) If we now let r1 = 72 = r by confining our attention to a particular point in space in front of the loudspeaker where this is true, we get (8.49) where ¥ ia phase angle equal to kr ~ 2/2 radians,244 LOUDSPEAKER ENCLOSURES (Chap. Since Ue a, we have simply that iol = lel (3.50) Amazing as it seems, the sound pressure produced at faraway points equi- distant from cone and port of a bass-reflex loudspeaker is directly proportional to the volume velocity necessary to compress and expand the air inside the box! ‘Av very low frequencies, where the reactance of Cas is very high, U, becomes nearly equal to U, and the pressure, measured at points r= approaches zero. In fact, the two sources behave like a dipole so that the radiated sound pressure decreases by a factor of 4 for each halving of frequency. In addition, if we are below the lowest resonance frequency of the circuit of Fig. 8.18, the diaphragm velocity Ue Fic. 8.19. Simplification of the ciruit of Fig. 818, The generator preaoure pe equals eeBi/ iy + RedSo. halves for each halving of frequency. Henee, in this very low frequency region, the sound pressure decreases by a factor of 8, which is 18 db, for each halving of frequency. Note that this decrease is greater than that for a loudspeaker in a closed box or in an infinite batt. 8.13. Resonance Frequencies. In the interest of simplifying our analysis, let us redraw Fig. 8.18 to be as shown in Fig. 8.19. ‘The new quantities shown on that circuit are defined as follows: Ma = Mao Mat Mas 851) » Be 52 Pa = Ge RpSe + Bas + Stan (8.52) + Me (8.58) + Rane (854) ‘The resonance frequency of the vented enclosure (ws = 1/-/Cxalfa) may or may not be set equal to that for the loudspeaker itself. As is common practice, jet vive resonance frequencies of the enclosure and the loudspeaker equal wy that is, let MiCas = Marl a (8.55) Part XX] BASS-REFLEX ENCLOSURES 245 At the angular frequency as, the impedance of the enclosure becomes very large, and nearly all the radiation takes place from the port. In other words, U, approaches zero, and U', becomes large. ‘As further steps in handling the circuit of Fig. 8.19, let us define Q's for the various parts of the circuit as follows Os Fe Sea 6.56) (37) (8.58) 4 (8.59) GoCaa(ar + Raa) ‘The total impedance Z, looking to the right of the terminals 1 and 2 + ea(4 gb nals), (8.60) Cas (861) Also, we shall need Us in terms of U.. (+98) Us a (8.62) C7Q) + iMlwFe0) = CooFe} It is difficult to handle these equations if Qs is small, There is no advantage, furthermore, to a bass-reflex box over u closed box if Qs is small. If Rar is large, the box becomes the same as though it were tightly closed. If Ras is large, then Us is small and it. would be better to use a closed-box baffle. We shall limit our studies to Q, greater than 2.5,246 LOUDSPEAKER ENCLOSURES IChap. 8 We shall also limit our studies to cases where the resistance is principally in either the port or in the box but not in both. "Then, Qs equals either Qu oF Q. Under these conditions there are generally three frequencies ws, ww, and ‘aw at which the reactance to the right of terminals 1 and 2 becomes zero for given values of Quand Cas/Can. One of these frequencies, aw, is near the resonance frequency uw. Another frequency, wx, occurs at a value greater than ue, ‘The third frequency, w1, occurs at a value less than we "The upper and lower resonance frequencies are found from Ea. (8.61) as follows: Set Qe © and Qr= Qu. Then, oo/ws and wx/ee are determined. ‘The results are plotted in Figs. 8.20 and 8.21, respectively. Next, set Qu = © and Q; = Q.. Then, ws/ax and on/oare determined, ‘The results are plotted in Figs, 821 and 8.20, respectively. In other words, Fig. 8.20 gives the lower resonance frequency for Q. = = and the upper resonance frequency for Qu = «©. Figure 8.21 gives the upper resonance frequency for Q. = « and the lower resonance frequency for Que If the value of Qs is greater than about 10, the two resonance frequencies, «1 and wy, are symmetrically located on either side of we for @ given value of Cxs/Caa. That is to say, wo/un * on/0 "The resonance frequesicies for londapeaker systems where the serie and parallel branches of the circuit of Fig. 8.19 do not have the same angular resonance frequency are more difficult to find, If we set Qs = © ‘and if the angular resonance frequencies of the two parts of the circuit, fre wand ws, we find that, for zero reactance to the right of terminals Land 2 of Fig. 8.18, #1) (8.63) Can” Ma i where wo is the geometric-mean angular frequency equal to Vai: and Go = Gow, 1 = Wo/G, and cou/io9 = K = wo/ay. Note that this equation is symmetrical about w just as Eq. (8.61) is for Qe = ©. B14. Reference Sound Pressure. ‘The concept of reference sound pressure has been discussed in Part XIX and here is expressed as follows; eBloo RF Re) Mere Ipc! (8.64) By way of review, pe is the sound pressure that would be produced at r if the frequency were above the critical frequency wy and if the loudspeaker were nondirectional 8.16. Radiated Sound Pressure at the Critical Frequencies. We expect that there are important fluctuations in the response curves at the three critical frequencies wy, ww, and w,, At wy = wx, the radiation ‘occurs principally from the port, and the radiated pressure there depends Part XX] am afin 100 Q, antes 10. UN 21a 16-18" "25-27 2a 26 28 30 ajo tor Qe= 00 fet '¢ Qu 2 Fra, 820, Value of two ofthe zoro-reactance frequencies ofa bast-reflex loudspeaker system as a function of Cas/Can = Mar/Mas with Qs as the parameter. wn ia the ‘eaonance frequency of either the series or parallel branch of Fig, 819. ‘The region eng ttaedaseaime antsy tara a tent eset, The ge Seater nce at he tthe dhl tnd cho cafe ar Gam of ele scent ‘athe edd gow cufeer Gu'2 = or wale a TT T] 1 : | LH t | | 4 | MH Pa | | a i | Decrees Li anderen | 18" "Is -20 22 Fa 25 28 30 efter Qe=o> . afi Qu => Fo, 821. Value of one of the reosteartancefrequenins of « basses loudspeaker ftom maa Tomcton of CCan = Mso/ Mas mith Qy asthe parameter, ni te ‘sonuncefroqueny of ether the vere othe paalel brench of ig 819.8 LOUDSPEAKER ENCLOSURES, ic both on Q) and Qs. Regardless of the values of Qs and Cas/Cany the upper frequency for zero reactance, wn is about the same as though the port were closed of. Thiscan be seen by comparing the curve of Fig. 821 marked Q; = 2 with the eurve for Qs = ©. I ey/ao = 1 is small, the if © a @ Fic, 82. elatve response a the lower criti fronueney, foray = Ona a fone on of Qe and Qu Bow unlined response of the loudspeaker with vented enclosure wil be neatly the same 1s that for the same loudspeaker in an infinite bafle, This condition requires a large box (large Cus) which obviously approaches being. an infinite bale, For Ras * 0 (Qe © 02), No Acoustical Material in Box, We find the ‘magnitude of the ratio of the rms sound pressure Lo the reference pressure Past XX BASS-IEFLEX ENCLOSURES 9 at the three frequencies, w,, wy, and on, from Eqs, (8.50), (8.62), and (64) and by solving for C7, from Fig. 819 with the aid of Eq. (8,60). FOR W = wi. The pressure ratio at the lowest eritieal frequency is » EF Oatlaciont 2) ott Bier ak = 1. Plots of this equation, n decibels, are given in Fig. .22 for four values of (665) 14 z aw am C EH Gf eH ® 7 : T l | = | | + | ac 1 Gafeai9 | | ; Guleu20 mor ts SHEL io "0 ‘a 10 © @ ie 8.23 Raltive raponte at the middle cite! Frequency, we © ue for Raw = 0 ses fonction of Qu and Gy Box urine vous = wy swe. ‘Tho pressure ratio atthe middle critica frequency ba) Be |” CuaiCae F HOD Plots of this equation, in decibels, are given in Fig. 8.28 for four valwee of Ca/Con. (6:66) cx Thea the hg nel foe Veo Pe (os wet \[ Qu wefan? MO - 8) + Ste) Re erate | (6.07) This equation, in decibels, is plotted in Big 8.24 fur Four values of e/a250 LoupsPHAKEH ENCLOSURES icimp. 8 Patt XX} ASS-REFLEX ENCLOSURES 251 For Rar * 0 (Qu), Acoustical Material in Box. "The magnitude This equation, in decibels it plotted in Fig. 825 forfour values of of the ratio ofthe rns und pressure tothe reference sound pressure is alos found by wong the same equations and figure 8 for Qe = = Tone = ew + wy. The pret ratio at the mile erica! fequeney riqt o- ; a Pa) 1 cr Lely Fel" Callan 1700. eo) 1 y 4 “Thin equation, in decibels, is ploted in Fig. 8.26 for four values of 2 - Calan a | | sol l T | \ { Hg t | 4 -x If »| - +t 10 . : @ l| | WAL tu i | pe tttl. Hoag | aa i ae @ ® oA am : - : my “oa te “oe 10 4 © a @ / fn, $24, live epee atthe upper ea fnquoney, wn er Ray = 0 a8» 1 vor «= wz, The pressure ratio at the lowest critical frequency is | | 7 LL eS Seppe _. (6.68) cat iene es oe © «LOUDSPEAKER ENCLOSUR Ss ct ror w = we. The pressure ratio at the highest critical frequency is (wn/wo)Q ~ oe. —___ (8.70) Pe wo) oo? wot fou le VO =) " al, iS aia +E QA (ws?/wn? — 1) This equation, in decibels, is plotted in Fig. 8.27 for four values of 1816. Performance. With the formulas and charts just given, it is possible to calculate the response of the loudspeaker in a bass-reflex Fa = 19 t =| 1 janet of abet et be (@) [eas/Can=20 f al sh fe) a 3) pi. 88, Rata ea hemi eel frquney, a fo Rar # 0 CORRS at ‘enclosure at the three critical frequencies. A complete example is given in Par. 8.18. From Fig 8.19, we see that for frequencies below wo, radiation from the port (proportions! to —Us) is out of phase with the radiation from the diaphragm (proportional to U,). AS a result, the response at very low frequencies is usually not enhanced by the addition of the port. Above the resonance frequeney a, radiation from the port isin phase with that from the diaphragm, with a resulting enhancement over the closed-baffle response. ‘The amount of the increase in response generally averages About 5 db over a frequency range of one to two octaves. ‘An important reason for using a bass-reflex enclosure is that the loudspeaker produces less distortion at frequencies between wy and 2ue for s given acoustic power radiated than would be the ease if the box were Closed, ‘The assimption on which this statement is made is that the Pant NI HASS-HEPLEX ENCLOSURES: 253 motion of the air in the port is distortionless even though the amplitude of vibration is large. ‘Tiss true generally because there is no suspension ‘or magnetie circuit in the port in whieh nonlinear effects can occur. A. : . ; yy LH ~ 7] * cgareaa| x {4 Lu , aig (a) (by) a g : nyt r| / ; Oe ~ © (ce) a (d) io, 8.27. Relative response at the upper ertieal frequency, ow for Rar £0 as a function of Q.and Que Box ied, men large loudspeaker diaphragm usually is superior to a small one because the amplitude of its motion is less, thereby reducing nonlinear distortion. ‘The port should be placed close to the loudspeaker to avoid an irregular direotivity pattern up to as high a frequency as possible and to achieve the widest possible frequency range of enhancement,An advantage of a bass-reflex enclosed loudspeaker is that, where space is a factor, a properly tuned bass-refiex system helps to offset. the effect of the small enclosure volume. For a small enclosure volume, it is, best to tune the port exactly to the loudspeaker resonance, For a very large enclosure, it is permissible to tune the port to a fre~ quency below the loudspeaker resonance. This permits the use of a port that is not over about 1.5 to 2.0 times the effective speaker area. 8.17. Construction and Adjustment Notes. Tho box should be very rigid in order to resist vibration. The joints should be tight—-glued or caulked—and the larger panels should be braced by gluing reinforcing strips to them. The access side should be screwed on securely. In constructing the box to achieve a desired volume, account should be taken of the volume of air displaced by the loudspeaker and by the inward extension of the port if such occurs. As we said earlier, the volume of air displaced by the loudspeaker (in cubie meters) equals about 0.4 times the fourth power of the advertised diameter (in meters) ‘When the cabinet, has been completed and the loudspeaker has been installed, the correctness of the tuning may be determined by connecting ‘an audio oscillator with an output impedance about 100 times that of the loudspeaker to the electrical terminals. Next, connect a voltmeter across the loudspeaker terminals. Then vary the frequency of the oscillator until the two eritieal frequencies w, and wy are observed as peaks in the voltmeter reading. ‘These should occur at the calculated frequencies if the design is correct ‘The resonance frequency of the enclosure can be increased by deereas- ing the enclosed volume Vz. ‘This can be accomplished by putting blocks of wood inside through the port opening. The port area may be decreased to lower the resonance frequency of the enclosure. 8.18, Example of Bass-reflex Enclosure Design. In the previous part we discussed in detail the design of a closed-box bafile for a low-frequeney (woofer) loudspeaker. We presented methods for the determin tion of its physical constants, and we showed a comparison between measurements and calculations, In this part we shall use the same loudspeaker unit and box. A port will be introduced into the box that resonates with the box compliance to the same frequency as the series branch of the circuit of Fig. 8.19, that is, VM Cas, the element sizes are as given in Example 8.3 (page 235), except that ueney a» is different for the analysis here, viz., JVM Cas = 10°// T0378) = 231 radians/see fo = wo/2n = 36.8 eps Cas/Can = 182/214 = 0.85 Mar = CasMa/Can = (0:85)(10.27) = 874 keg/mot | Past XX} BASS-R EX ENCLOSURES 255 ‘The area of the port must be chosen. If the circumference of the port is less than a half wavelength, the port behaves as a simple source and it will radiate the same amount of power regardless of its size for a given volume velocity. Arbitrarily, let us select the area S, to be a little more ‘than one-half of the area of the diaphragm, say, 8, = 0.055 m? Calculate the acoustic mass at the outer end of the port. O.6axee _ 0.690 Ma = SS, VS (0.6)(1.18) ~ Fro.05s ee ‘The remaining mass of the port is Mae = Mar ~ Ma = 8.74 ~ 1.7 = 7.04 kg/m! eros = (4 0.6as)o0 _ 0.18) 7.04 3, toss + LT So f, the length of the tube behind the port, is _ 7.04 — 1.7 ¢ = BOS M = 025 m ‘This says that the length of the tube behind the port should be about 9.8 in, ‘The acoustic resistance of the port is found from Eq. (5.54). 18. creas ye 9 028 Rar = AIS yantres) x 10 (225 +1) = V5 (012) mks acoustic ohm ‘To get an average value of Mar, let us take Rar = VBI 0.12 ‘The resistance from the front side of the port is, 1.82 mks acoustic ohms Stans = E22 = Shams = O.0107F" Let Stan and Wane equal their value at w = wy, thatis, fo = 36.8. Then, ‘ray = Ranz = (0.0107)(1355) = 14.5 mks acoustic ohms Finally, rar = Rar + ans = 16.3 mks acoustic ohmsand, for later use, 9.7 x10 5 Ras = GEESE + 856 + 15 mks acoustic ohms (231)(8.74) Qe = SUED 2 114 ‘This value of Q is very high. Let us see whether or not Q, is smaller. ‘The box contained one 18- by 35- by 3-in. piece of PF Fiberglas board with a flow resistance of about 2000 mks rayls/in., or 6000 mks rayls for Bin, Hence, from Par. 8.9, 6000 gop Ran = Go.a05) ~ 4920 a 4920 8 CBI) AA? X 1O-"NDO)? F C7 4920 — = 5 ft20,g ~ 90 mks acoustic ohm 10" : - ese *” and ¢ 1)(R74) 16390 ~* Hence, we may assume that Qe * Q. Qe Rar Mar Po gt po Fic, 828. Cirouit showing the port volume velocity for an external source of sound ‘when the loudapeaker cone it locked so that it cannot move. ‘The sound pressure wis that sound pressure which would exist at the eatrance to the port if the port ‘Wore blocked off go that Uy were zero, ‘The magnitude of the volume velocity Up equals [palCane, Tt is assumed that Raw < 1/aCaa. ‘The Q. of the box may be measured by replacing the loudspeaker unit by a rigid board and placing the box in free space, along with a source of sound of variable frequency. Reference to Figs. 8.19 and 8.28 shows that if U.=0, then —U, = Us. Placing a microphone inside the box measures the sound-pressure drop ps across the compliance Cs. A plot is then made as a function of frequency of the product of this sound pressure and the frequency of the oscillator. This plot is proportional to the volume velocity U,, ‘Then, Qe is found by dividing the frequency (cycles per second) for the peak of the curve by the width of the curve (cycles per second) at the points where the volume velocity is 0.707 of its peak value. Part XX} BASS-REFLEX ENCLOSURES 257 This measurement made on our box yielded a value of Q, = 7, which is much lower than that calculated, A plausible explanation for this is that the plywood absorbs energy internally when itis flexed. According to information in the literature, a 3/4-in. plywood box absorbs af 40 cps about half as, much sound energy per square foot as does the 3-inthick layer of Fiberg- lass at the same frequency. Thisis, of course, a function of panel size, type of glue, and bracing, so that at best only a crude guess can be made. At frequencies above 100 cps, the 3-in.-thick layer of Fiberglass will absorb much more energy than the wood box. Since the area of the walls of the box is about six times that of the Fiberglas, the box absorption is equivalent to adding an additional three times as much material, making the total to be four times as much area and volume of acoustical material as that now assumed. An approximate revised Ray would be Ray © 830 mks acoustic ohms Henee, ‘This number is very close to the value of 7 measured; so we shall assume Qe QaT From Figs. 8:20 and 8.21 we find the three critical frequencies oy = wy = 231 radians/see fo = 36.8 ops oy = 1.55u9 = 398 radians/sec Su = 57.0 eps on = 7p = 148 radians/see fu = 23.6 eps ‘To determine the relative sound-pressure response at the three critical frequencies, we use Figs. $25 to 8.27. First, we shall determine Q, at these three frequencies for the four values of amplifier resistance Ry, namely, 3, 14, 30, and 130 ohms. ‘The values of Ra, and Qs are given in Table 8.5, TABLE 8.5, Values of Ray and Q, vs. Amplifier Resistance Rohm 30 s.r x 108 | 1.09 x10" om | 22258 LOUDSPEAKER ENCLOSURES {Chap. 8 We obtain the values of the sound-pressure-level response relative to the reference sound-pressure-level response at the three critical frequencies given in Table 8.6. TABLE 8.6. Sound:pressure-level Response at the. to the Reference Sound-pressut al Frequencies Relative evel Response Comparison of the numbers given in Table 8.6 with the calculated ‘curves of Fig. 8.15 for a closed box shows that the average enhancement of the response at the upper two critical frequencies was about 6 db. By ‘comparison, the measurement on this particular loudspeaker showed average increase in response of approximately 4 db. No explanation is offered for this 2-db difference. Data reported by other observers usually show an enhancement of the order of 6 db. It is possible that transmission through the side walls of the box, which would tend to produce an ‘out-of-phase wave, contributed to the difference in the results. Time did not permit enclosing the box in a sand or a concrete jacket to determine ‘the magnitude of this effect. CHAPTER 9 HORN LOUDSPEAKERS, pant XXI Horn Driving Units 41, Introduction. Horn loudspeakers usually consist of a moving-coil driving unit coupled to a horn. When well-designed, the large end of the horn, called the “mouth,” has an area sufficiently lange to radiate sound efficiently at the lowest frequency desired. ‘The small end of the horn, called the “throat,” has an area selected to match the acoustic impedance ‘of the driving unit and to produce as little nonlinear distortion of the acoustic signal as possible. ‘Hom loudspeakers are in widespread use in cinemas, theaters, concert halls, stadiums, and arenas where large acoustic powers must be radiated and where control of the direction of sound radiation is desired. ‘The efficiency of radiation of sound from one side of a well-designed direct- radiator loudspeaker was shown in Chaps. 7 and 8 to be a few per cent. By comparison, the efficiency of radiation from a horn loudspeaker usually lies between 10 and 50 per cent. ‘The principal disadvantages of horn loudspeakers compared with the direct-radiator loudspeakers are higher cost and larger size ‘Before proceeding with an analysis of the horn loudspeaker, it should bbe mentioned again that the radiating efficiency of a direct-radiator loudspeaker can be increased at low frequencies by mounting several unite side by side in a single baffle. ‘The mutual interaction among the radiating units serves to increase the radiation resistance of each unit substantially. For example, two identical direct-radiator loudspeakers very near each other in an infinitely large plane baffle, and vibrating in phase, will produce four times the intensity on the prineipal axis as will one of them alone. Direct-radiator loudspeakers used in multiple often are not as sat factory at high frequencies as one horn loudspeaker because of the diffi. culty of obtaining uniform phase conditions from different direct-radiator diaphragms. ‘That is to say, the conditiors of vibration of a loudspeaker 259260 HORN LOUDSPEAK nS. [ehup. 9 one are complex, so that normal variations in the uniformity of cones result in substantial differences in the phases of the radiated signals of different cones at high frequencies. A very invegular aud unpredictable response curve and directivity pattern result. ‘This problem does not arise with a horn where only a single driving unit is employed. When two or more driving units are used to drive a single hhorn, the frequency range in which the response curve is not adversely affected by the multiplicity of driving units is that where the diaphragms vibrate in one phase. 9.2, Blectro-mechano-acoustical Circuit. The driving unit for a horn loudspeaker is essentially a small direct-radiator loudspeaker that couples to the throat of a flaring horn as shown in Fig. 9.1. In the next part, ‘we shall discuss the characteristies of the horn itself. In this paragraph Mouth of theo Pro, 9.1, Cross section of « typical horn loudspeaker with an exponential eross section ‘or this design, the radius of the throat is 0.2, the radius of the mouth 3.4, and the length 5.0 (arbitrary units). we restrict ourselves to that part of the frequency range where the complex mechanical impedance Zur looking into the throat of a horn is a pure resistance, Ay = pecSr mks mechanical ohims on Lu where py = density of air in kilograms per cubic meter ¢ = velocity of sound in meters per second ot = 406 mks ohms at 22°C and 10° newtons/m? ambient pressure ‘Sy = area of the throat in square meters zur = mechanical mobility at the throat of the horn in mks mechanical mohms (mobility ohms) Cross-sectional sketches of two typical driving units for horn loud~ speakers ure shown in Fig. 9.2, Each has a diaphragm and voice coit with * An authoritative discussion of horn loudspeakers is found in HF. Olson iments of Acowstical Fagineoring,” 24 ed., Chap, VII, D. Van Nosteand Ine., New York, 1947, Part XX) HORN DRIVING UNITS 261 8 total mass Mus, a mechanical compliance Cys, and a mechanical resistance ys = 1/rus. The quantity rus is the mechanical responsiveness of the diaphragm in mohms (mobility ohms) Behind the diaphragm is a space that is usually fled with a soft this space acts as a compliance Phase conection pg acoustical material. At low frequene! Cup Which can be lumped in with the compliance of the diaphragm. At high frequencies the reactance of this space becomes small so that the ‘ space behind the diaphragm becomes LA. Fis et a mechanical radiation resistance = Run = I/res with a magnitude also ‘equal to that given in Eq. (9.1). This resistance combines with the mechanical radiation resistance of the throat, and the diaphragm must develop power both to its front and its back. Magnetic ease Loaot tng Obviously, any power developed be- o hind the diaphragm is wasted, and at, ble pce high frequencies this sometimes be- sound cher comes as mich a3 one-half ofthe total Te ple on generated acoustic pover. 2) Sal In front of the diaphragm there is an airspace with compliance Cis. AL Hy low frequencies the air in this space CA behaves like an incompressible fluid, yy {| Vy that is, wCws is small, and all the air B displaced by the diaphragm passes into t the throat of the horn. At high fre- wae I eee quencies the mechanical reactance of aca hin airspace becomes suficintly ow | oO yp Ge, the air becomes compressible) so Fi. 2.2. Crom section of two typics that all the air displaced by the dia Lommavyina. unin he agphem phragm does not pass into the throat through a mall cavity with a mechan of the horn ical compliance Cus. ‘The voice coil has an electrical resistance Ry and inductance L. As stated above, zer is the mechanical mobility at the throat of the horn, By inspection, we draw the mobility-type analogous circuit shown in Fig. 9.3. In this cireuit forces “flow” through the elements, and the velocity “drops” across them. The generator open-circuit voltage and resistance are ¢, and 2. The electric current is ¢; the linear velocity of the voice coil and diaphragm is 1; the linear velocity of the air at the Uroat of the horn is wr; and the force at the throat of the horn is fr. As before, the area of the diaphragm is Sp, and that of the throat is Sy.262 HORN LOUDSPRAK ERS eh 93. Reference Power Available ficiency. In the middie-frequency range many approximations usually ean be made to simplify the analogous circuit of Fig. 9.3. Because the driving unit is very small, the mass of the diaphragm and the voice coil Mw» is very small. ‘This in turn usually, means that the compliance of the suspension Cys is large in order to keep the resonance frequency low, Also, the responsiveness of the suspension rvs usually is large, and the reactance Cx is small. Hence, in this fee (eee Ci S08 Fig, 9.8. Electro mechano-acovstical analogous cirouit of the mobility type for the Griving unit, assuming that the mechanical impedance at the horn throat is cS, that is; he mechanical mobility is Zur = V/ocSs- () Fic. 0, Simplified analogous circuits of the mobility type for the driving unit in the region where the motion of the diaphragm is rxitance-cantrolled by the bora queney range, the circuit reduces essentially to that of Fig. 94a, where the responsiveness behind the diaphragm is <5; mks mechanical mohms 12) jas; tks mechanical moh @2) With the area-changing and electromechanical transformers removed, we get Fig, 9.4b, where the radiation responsiveness at the throat is mks mechanical mohms (03) As before, Sr is the area of the throat and Sp is the area of the diaphragm in square meters. We have assumed here that the cavity behind the diaphragm in this frequency range is nearly perfectly absorbing, which may not always be true. Usually, however, this circuit is valid over a considerable frequency range because of the heavy damping provided by the responsiveness of the horn rwr. Also, rwr usually is smaller than Pact XI] HORN DRIVING UNITS 203 rus 80 that most of the power supplied by the diaphragm goes into the horn Solution of Fig. 9.46 gives us the reference current ¢ (Ry + Re) + A) ‘The maximum electrical power available to the loudspeaker from the generator, assuming It, fixed, is ‘Max power available 95) ‘The reference power available efficiency (PAE,«) is equal to the reference power delivered to the horn, \i:/*B%rwr, times 100 divided by the ‘maximum electrical power available. PABw 100) Rerun Be 5 ~ OF S/S) Co From Eas. (9), (9.4), and (9.6) we get oe 4008 ret Bier R a) + Se/So)P 94. Frequency Response. The frequency response of a complete horn loudspeaker, in the range where the throat impedance of the horn is ‘a resistance as given by Bq. (9.1), is determined by solution of the circuit of Fig. 9.3. For purposes of analysis, we shall divide the frequency range into three parts, A, B, and C, as shown in Fig. 9.5. Mid-frequency Range? In the mid-frequency range, designated as B in Fig. 9.5, the response is equal to the reference PAE given by Eq. (9.7) Here, the response is “flat” with frequency, and, for the usual high- frequency units used in cinemas with 300 eps cutoff frequencies, the flat region extends from a little above 500 to a little below 3000 eps. In this region the velocity of the diaphragm is constant with frequency, rather than decreasing in inverse proportion to frequency as was the case for a direct-radiator loudspeaker Resonance Frequency. It is apparent from Fig. 9.3 that since «Ca is small, zero reactance will occur at the frequency where ae (9.8) 2s VM wal * See also Fig. 7.2 p. 188, of reference 1. Note that Olson’s efficiency is the ratio 100raeBXP/ (Ree + rwrB*P), assuring that ave > rr264 HORN LOUDSPEAKERS chap. In practice, this resonance usually is located in the middle of region B of Fig. 9.5 and is heavily damped by the responsiveness rr, 50 that the velocity of the diaphragm is resistance-controlled Low Frequencies. At frequencies well below the resonance frequency ‘de response will drop off 6 db for each octave decrease in frequency if the B — _ 109 20 E | x e z 2 é 2 10] . fro 3 & 3 +6 DB per oe i (Gine6) : DB pec octae z “Ger Conant) | =12 DB peace i (Guspopety Ses) 0 iw icy Toobo Frequency Fic. 9.5, Power availble efficiency of » horn driver unit a the frequency’ regan ‘where the mechanical impedance of the horn at the throat is pure retstance pw Sr ‘The ordinate isn logarithmic resle, proportional to decibel throat impedance is a resistance as given by Bq. (9-1). This case is shown as region A in Fig. 9.5. In practice, however, the throat impedance Zr of the horn near the lowest frequency at which one wishes to radiate sound is not a pure resistance. Henee, region A needs more eareful study. RetRe a SrSp__ fy a Fra, 9.6, Analogous eireut for @ hora driver unit in the region where the diaphragm would be stiffness controlled if the horn mobility were infinite, “The actual value of the methanieal mobility of the horn at the throat is wr. Let us simplify Pig, 9.3 so that it is valid only for the low-frequency region, well below the resonance of the diaphragm. Then the inductance L, the mass Mu, the compliance Cyn, and the responsivenesses ries and. rua may all be dropped from the circuit, giving us Fig. 9.6, Solving for the mechanical mobility at the diaphragm of the driving unit yields ue _jaualSr/So)eme_ ue jal + (9.9) Pact XT} HORN DRIVING UNITS 265, where Cus = fuse us + Cus @.10) ur = mechanical mobility at the throat of the horn with area Se ‘The mechanieal impedance at the diaphragm of the driving unit is the reciprocal of 24, wae where Zr = 1/2wr = mechanical impedance at the throat of the horn with area Sr. ‘As we shall show in the next part, the mechanical impedance at the throat of ordinary types of horn at the lower end of the useful frequency range is equal to a mechanical resistance in series with a negative compliance. That is to say, 1 Zur Rar +5 (9.12) ‘The German in ®ur indicates that this resistance varies with frequency. Usually, its variation is between zero at very low frequencies and pacSr [as given by Eq, (9.1)} at-some frequency in region A of Fig. 9.5. Hence, the mobility zwr = 1/Zws is a resistance in series with a negative mass reactance. In the frequency range where this is true, therefore, the reactive part of the impedance Z 4. catt he canceled out by letting [see Eqs, (9.11) and (9.12)} Stitt (tga Soom ~ ta (ots +a) ‘Then, @.4) where twe is the acoustic responsiveness of the throat of the horn at low frequencies transformed to the diaphragm, ‘The power available efficiency for frequencies where the approximate circuit of Fig. 9.6 holds, and where the conditions of Eq. (9.13) are met, is 400 BU tw. (FR) FBV aT oe PAE, ‘The responsivencss tu. usually varies from “infinity” at very low frequencies down to Sr/(Si%pe¢) at some frequency in region A of Fig. 9.5. High Prequencies. Mt very high frequencies, the response is limited principally by the combined mass of the diaphragm and the voice coil ‘uo. If Cw were zero, the response would drop off at the rate of 6 db per octave (see region C of Fig. 9.5). It is possible to choose Car to266 HORN LOUDSPEAKI (Chap. 9 resonate with Mus at a froqueney that extends the response upward treyond where it would extend if it were limited by Muo alone. We ean Understand this station by deriving a circuit valid for the higher fr- Gjuencies as shown in Fig. 0.7. It is seen that a damped antiresonance Seoure at a selected high frequency. Above this resonance frequency, the reaponte drops off 12 db for each octave increase in frequency (see region C of Fig. 9.5) ‘Because the principal diaphragm resonance [Eq. (0.8) is highly damped by the throat resistance of the horn, itis posible to extend the region of fla response of «driver unit over @ range of four octaves by proper choice Aeon BC BY nee Fic. 9.7. Analogous circuit for a horn driver unit at high frequencies where the diaphragm mass reactance ia much larger than its compliance reactance, of Cu: at higher frequencies and by meeting the conditions of Eq. (9.13) at lower frequencies. 9.5. Examples of Horn Calculations Example 91, As on example of the reference power availuble efficacy of = theater horn driver unit designed to operate in the frequency range above 500 eps, let tu insert typieal values of the loudspeaker constants into Eq. (9.7). Wehave Ry = Ry = 24 ohms ‘B= 19,000 gauss = 1.9 webers/m* = 249m {& webers4/m? Sah em! = 3.14 X 10-tm? Sp = 283 cmt = 283 X 10-¢ mt S1/So = 0111 Sr/Soh = 3:14 x 10° (28.8 10-0 = 39.2 m* Per = 89.2/406, 0.0065 mks mechanical mohms jee Eq. (93)] (400) (24) 0.0965) (44) PARw = [425 + (8)(L1IDP 10,750 10350 — 123% Solution. To increase the PAE itis seen from this example that the electrical resistance Re should be reduced as far as possible without decreasing the length 1 of tthe wire on the voice coil. Within given space limitations, this ean be done by wind- Ing the voice coil from wire with a reetangular cross ection rather than with x circular ross section. "This means that the voiee-coil mass wil be increased. | Reduction of ‘Re further will domand wire of larger eross section which wil require a larger ai g9p, {rith a corresponding reduction in B. Also, the voice coil must not become too large fs ith mass wil Limit the high-frequency response. We note further that the value of fice and hence the ratio Se/Sp%, would teem to need to be large for high efficiency. However, if Sr/So! becomes too large, reference to Fig. 9.40 shows that too much power willbe dissipated in uw and the eficiency will below. A compromise is needed to effort the desired frequency response at high efficiency and minimum cost, The ‘optimum value of S-/Sp is ealeulated as shown in the next example Part XU] HORN DRIVING UNITS. 267 Example 92, Determine the optimum value of the area of the throat of the horn for use with the driving unit of the preceding example. Caloulate the reference PAE for this optimum condition, Solution, Letus express the ratio of the throat area to the diaphragm arca as ‘Then Eq. (0.7) becomes eo $00R a8 /oeSo [+i +0(R +R +22] "To determine the maximusn value of PAB, we differentiate the equation with respect to-a.and equate the result to zero, This operation yields Ry + Re Ot = BF Re + Ppa Invertiog, So) Bu a Substitution of the constants from the previous example gives us 50 “4 Dan RT aOOTERAT TO OF dan, = 0.558. Hence, Gr) 5.7 X 10-t mt Substitution of @ = 0.856 into the equation for PAE above gives, (400) (24)(0.555)(48) % 104 CEE Ay caonensy [18+ 058s (a + SERIE] PAW = By comparison with the previous example, we ace that, by proper choice of the ‘throst area, ve ate able nearly to double the reference (mid-Trequency) efficiency. Tn Addition, let us ese what would happen ifthe output impedance of the power amplifier were made very low (eay, Ry = 1 ohm), Se 82 a 208 sp = BSI ant 10 sa BOLI ayy — ‘Thia moult shows that for maximum effieney the amplifier impedance should epproxi- imately equal the loudspeaker impedance268 HORN LOUDSPEAKERS (Chap. 9 parr XXII Horns 0.8, General Description, A hor is infest an acoustic transformer, Rag aeuetialaee etme atone ne the fiat of cone resonances dacuned in Part XVIII A legearea Shuphragm har a radiation ispedance toot i more near redette over ig assay ras a ie ease eee Ca ea trea diaphragm (eee Fig 53)" Ana rele more hover i todd at eae ea geass ae Tn designing n hom fora pereular appiction we usully wih toes the parameters so as to radiate the maximum amount of acoustic power over the desired frequency range with suitably low nonlinear distortion, Once we have stated the frequency range, tolerable distortion, and desired radiated power, we can choose the driving unit and then proceed to calculate the throat and the mouth diameters and the length and shape aa SPE oer rien cireumference large enough so that the radiation impedance is nearly resistive over the desired frequency range. Reference to Fig. 5.3 shows esas se enn and Nusa ore a oe Sees ee iw aeated ect ee ator ae eee ea coramateraaleatete ieee uicems ecsrarn er atarcgenee iret far cea an cae a od dea circumference C or mouth area Sw, C= 2Ve8u >d (9.16) where isthe lngest wavelength of sound that is to be radiated efficiently. Sete eee types of longitudinal setion are poestble fora homme may be J bar, cont, ypert o stim, Tem agen real SF For a hor to be a autsfatory of the horn in meters, and Sr is the pear the throat end should increase throat area in square meters, gradually with x (see Fig. 9.8). If The oe comtten atape of longitudinal section mening the requ ae ae aa aee ay (oa vie nn cnet cto Part XX} WoRNs 269 tional area at any point x along the axis is given by the formula S = Sr (9.17) where S = cross-sectional area at z in square meters ‘Sr = cross-sectional area of the throat in square meters flare constant in inverse meters distance from throat in meters If the horn is a number of wavelengths long and if the mouth eireumfer- ence is larger than the wavelength, we may call the horn “infinite” in length. This simplification leads to equations that are easy to understand and are generally useful in design, Theoretical Considerations. ‘The general differential equation for sound propagation in an exponential horn whose length 1 is “infinite” is ap a _ 2p ae a “a? (9.18) ‘where p= sound pressure at a point along the length of the horn, (It is, assumed that the pressure is uniform aeross the eross section, of the horn.) ¢ = speed of sound fiare constant. Ubviously, m determines the magnitude of the second term of the equation above, which expresses the rate at which the sound pressure changes with distance down the hom. If m= 0, Hq. (9.18) becomes the equation for propagation in a cylindrical tube, Ze, a horn with zero flare A solution to By. (9.18) for the steady state is Dt) = Pyare AVE (9:19) whore 5 (9.20) ‘The volume velocity U at any point 2 is [see Ba. (2.58)] 8 [m, VER : Impedances. The acoustic impedance Zs, at a point x along the horn where the cross-sectional area is S, is 2a rks acoustic ohins (9.22)270 HORN LOUDSPEAKERS [Chap. 9 At the throat, where S = Sz, the mechanical impedance Zu ZarSr? is my pense Par Em pas sft = Fy + 50m = Rur-+j—— mks mechanical ohms (0.23) ‘The acoustic mobility z4 = 1/Z, at the throat is U_ Sr mai Sm to Ba Sef gS (024) The real and imaginary parts of Z4 and 24 behave alike with frequeney and differ only by the magnitude ($/pyc)* and the sign of the imaginary part. Notice that this derivation is not restricted to a circular cross section. Limitations on cross-sectional shapes will be discussed later. Let us see next how varying the flare constant m affects the acoustie impedance Z4 at any point along the horn where the area is S. Flare Constant and Throat Impedance. When the flare constant m is greater than 4r divided by the wavelength (m > 2k, low frequencies), the acoustic resistance Rar and the acoustic reactance Xr, at the throat of the horn where the area is Sr, are Par = 0 Xu = (% [RE -1) When the flare constant m equals 4x divided by the wavelength, the acoustic resistance and reactance are (9.25) Rar = 0 Xar= (9.26) For all cases where m is less than 4x divided by the wavelength (m < 2k, high frequencies), the acoustic resistance and reactance at any point 2 along the horn where the cross-sectional area is $ are Ray = EA 7 ae = etm DES, ~ a8; (9.27) Xar where Car = 28x/pue%m, For very high frequencies, the reactance approaches zero and the resistance approaches pw/Sr or pw/S in general. This is also the impedance for a plane progressive sound wave in a tube of uniform cross section S, Part XXII] HORNS om Clo Frequency. The special case of m = 4/A occurs at a frequency which we shall designate ., where L°% (928) ‘This frequency f. is called the culo’ frequency because for frequencies lower than this'no power will be transmitted down the horn, ée, the impedance at all positions along the horn is purely reactive [see Eq. (9.25). 13 uf 210 5 os| < Zos| g Bos I} Fos|_ [ dad wey 5 ¥ | B03} ol fy ‘on oe 18 1 v0 Yo io, 9.0, Potof the quantities A and B, whith ae defined by the relations given on the graph ‘To obtain the acoustic impedance atthe throat of the horn in terms of the eutoff frequency, we observe thal f/f = m/2k. Substituting in Eg (0.22) yields san i= G)vof)-teeste am where S; = throat area in square meters pot ~= characteristic impedance of air in mks rayls ‘fe = cutoff frequency J = driving frequency Graphs of two quantities A and B that are directly proportional to the resistive and reactive parts of the acoustic impedance at the throat of an infinitely long exponential horn are shown in Fig. 9.9. ‘The quantities A and B also are directly proportional to the real and imaginary parts of ‘the acoustic mobility at the throat. ‘The relations among A, B, Rar, Xan rar, and Zar are given on the graph.272 HORN LOUDSPEAKERS. [Chap.9 When the frequency is greater than approximately double the cutoit frequeney f., the throat impedance is substantially resistive and very near its maximum value in magnitude. Finite-length Horn. ‘The equation for the acoustic impedance at the throat of an exponential horn of finite length was given by Eq. (5.80) For exact caleulations, that rather complicated equation must be used whenover the bell diameter is not large or when the horn length is short ‘To illustrate what the words “large bell diameter” and “long length” mean, let us refer to Fig. 9.10. If the cireumferenee of the mouth of the horn divided by the wavelength is less than about 0.5 ( 05. Tn the frequency region where the circumference of the mouth to wavelength ratio lies between about 1 and 3, the exact equation for a finite exponential horn (Eq. (5.80)] must be used, or the results may be estimated from b and ¢ of Fig. 9.10. When the length of the horn becomes less than one-fourth wavelength, it may be treated as a simple discontinuity of area such as was discussed in Par. 5.11 (pp. 139 to 141). Obviously, if one chooses a certain mouth area and a throat area to obtain maximum efficieney, the length of the horn is automatically set by the flare constant. m, which is in turn directly dependent upon the desired cutoff frequency Nonlinear Distortion. A. sound wave produces an expansion and a compression of the air in which it is traveling, We find from Eq. (2.6) that the relation between the pressure and the volume of a small ‘'box”” of the air at 20°C through which a sound wave is passing is 0.726 ‘ ae (9.30) P where V = specific volume of air in m*/kg = 1/o5 P = absolute pressure in bars, where | bar = 10* newtons/m* Phis equation is plotted as curve AB in Fig. 9.11 Assuming that the displacement of the diaphragm of the driver unit is sinusoidal, it acts to change the volume of air near it sinusoidally. For Pact XU) HORNS 273 on, e oa cog Ft +. +5 09254. T 7 reat an ial esistonee HH iN NIAY reactance, | | || Tass [| ot a CE) i 2 Pin $ 7 | é aT = alone g q i z EEO : 6 8 5 @) fl 6 Bo 2 OO075A fe ostaeal @ 2 fo cion. oom t= 1.09 \ Resistance | | | 07s 050] + Sif Resatnee 02s acs ‘ CER ae 0 @) Frequency =//. © ie. 910. Graphs showing the variation n specie acute impedance a the throat of fost ‘pont home a2 fancton of frequency with tll ameter asthe parameter, The oto iene, e/tean the set ameter ~ OO17S fsbo arhel onsale Irene ae (a) C0236, (B) CHOA,, (2) COMA, (2) C=O94, and (0 SS Themouthortnehern issued tobe ermnnedinan inte Pl, (Air Ol)274 HORN LOUDSPEAKERS Chap. 9 large changes in volume, the pressure built-up in the throat of the horn is no longer sinusoidal, as can be seen from Fig. 9.11. ‘The pressure wave so generated travels away from the throat toward the mouth. Pin bors i | 1205 a.3600 Soop ume Vn MY Ka | t 1 PerAbsoute pressure in bars 1 VeSpecte volume 1 aif Mee ye 140s “ Fra. 9.11. Plot of the gas equation PY? = 1.26 x 10%, valid at 20°. Normal atmospheric presture (0.76 m Hg) is shown as P, = 1 ber If the horn were simply a Tong eylindrieal pipe, the distortion would increase the farther the wave progressed according to the formula (ai assumed)? Pea PAN Pry a trig Pe Pi By Po Po m9 sound pressure ofthe fundamental frequency in newtons per square meter px = rms sound pressure of the second harmonie in newtons per suere meter Py = atmospheric pressure in nowtons per square meter Ee ofe = 2n/d = wave number in meters y= 1A for air 2 = distance the wave has traveled along the eylindrial tube ia meters (9.31) where pr *A.L, Thuras, R, T, Jenkins, and H.T. O'Neil, Bxtrancous Frequencies Generated in Air Carrying Intense Sound Waves, J. Acoust. Soe. Amer, 6: 173-180 (1935). “1. H, Black, A Physical Analysia of the Distortion Provluved by the Non-linearity of the Medium, J. Acoue, Soe, Amer, 12: 266-267 (1940). Part XXII) HORNS 275 Equation (931) breaks down when the second-harmonic distortion becomes large, and a more complicated expression, not given here, must be used. In the case of an exponential horn, the amplitude of the fundamental decreases as the wave travels away from the throat, so that the second- harmonic distortion does not increase linearly with distance. Near the throat it increases about as given by Bq. (0.31), but near the mouth the pressure amplitude of the fundamental is usually so low that. very little additional distortion occurs. ‘The distortion introduced into a sound wave after it has traveled a distance x down an exponential horn for the case of a constant power supplied to unit area of the throat is found as follows: 1, Differentiate both sides of Eq. (9.31) with respect to 2, so as to obtain the rate of change in ps with z for a constant py. Call this equation (9.314). 2. In Eq, (9.31), substitute for px the pressure pre, where pr is the rms pressure of the fundamental at the throat of the horn in newtons per square meter and m is the flare constant. 3. Then let pr = -V/Trpez, where Iris the intensity of the sound at the throat in watts per square meter and pc is the characteristic acoustic impedance of air in mks rayls. 4. Integrate both sides of the resulting equation with respect to =. ‘This yields Per cent second-harmonie distortion 50 + 1) WTroe fy 7 Po fe m3] (9,32) For an infinitely long exponential horn, at normal atmospheric pressure and temperature, the equation for the total distortion introduced into a wave thal starts of sinusoidally at the throat is Per cent second-harmonic distortion = 1.73 f VTz x 10-* (0.88) where f = driving frequency in eycles per seeond cutoff frequency in eyeles per second I = intensity in watts per square meter at the throat of the horn Equation (9.33) is shown plotted in Fig. 9.12. Actually, this equation is nearly correct for finite horns because most ofthe distortion occurs near the throat Equation (9.33) reveals that for minimum distortion the cutoll frequency fe should be as large as possible, which in turn means as large a flare constant as possible. In other words, the horn should flare out rapidly in order to reduce the intensity rapidly as one travels along the horn toward the mouth276 MORN LOUDSPEAKERS Unfortunately, a high onto frequency ie not a feasible solution fon home that are designed to operate over a wide frequency range. Tn this case, it is necessary to operate the horn at low power at the higher fro quencies if the distortion is to be low at these frequencies. This goal is achieved automatically to some extent in reproducing specch and music, because above 1000 eps the intensity for these sounds decreases by abouy 4 factor of 10 for each doubling of frequency 109 a nea] TZ Fat ot ry cof ry Ty ae | eae e 2 ; P<
rr. From Eq. 07), par = 4009(0000.05)(15)" Tass0.05)-+ cay ~ 72% ‘Ass trial, let us make Sp/Sr = 2.0, ‘Thon rr ~ 0.025, and PAE = 24%. Finally let Sp/Sp = 2. Then, rwr ~ 0.1, and PAB = 15.8% eis seen that the ratio ofthe throat and diaphragen areas may be made equal with, little loss of effcieney, thereby making our horn of reasonably short length. ‘The length of our horn is found, from Eq, (0.17), 0.93 aa ~ me = 2.94 201 2 = 3 135m = sate ‘The intensities for a horn with # throat area of 0.049 m* are as follows, assum cuniform pressure diteibistion: Premeney | Power, wae | Wauelanta _}—__.| a 70-250 | 158x104 | 322x104 250-400 1.58 x10 | 322x106HORN LOUDSPEAKERS (Chap. © Lotus set the upper limit of operation at 600eps. Thea f/fe ~ 10. Extrapolation of the line for 10 in Fig, 0.12 t0 2.22 X 10-* shows that the per cent seeond-harmonic distortion in the horn will be about 0.02 per cent, which is negligible. In fact, the power level could be increased 30 db before the distortion would become as large as 1 per cent. ‘This calculation would seem to indicate that the low-frequency unit could be ‘operated successfully above 600 ops. Hovtever, it seems from experience that for ‘paychological reasons the changeover from the low-frequency to the high-frequency hhorn should occur at a frequency below 600 cps for best auditory results. 53° + 38 END VIEW Fro. 9.17. Plans for a simple stenight exponential horn with a cutoff frequency of 60 cps, a throat area of 0.089 mt, and « raouth area of 0.98 m™ Tot neat ‘what value the total complianea in the drivingainit cimuit-amght ta have ifivis to halance out the maas reactance of the horn at frequencies below the diapbragea resonaace frequency. From Eqs. (8.12), (8.13), and (8.23), 282 282 Soper ~ SohPom 2 4 » gana ~ BE X 10° newton ‘The quantity Cys includes the combined compliance of the loudspeaker and the tnclogure behind it. Reference to Fig. 8.5d shows that this is reasonable value of compliance to expect from a loudspeaker of this diameter. In case the compliance is not correct, we can vary the size of the thront, or even m somewhat, in order to achiove the desired value for Cus. “Tweo possible horns for our design are the straight square horn shown in Fig, 917 fr the folded hora of the Klipach typet shown in Fig. 9.18, If the straight horn is used, it will probably be necessary t0 put it partially above the ceiling or below the finor in order to make its presence nonobjectionable in the room, Solution for Highsrequency Horn. As a cutofl frequency, let us choose Sx Cus = $5 Cus = Cue Jo = 300 ep6 We shall use an clerical erosover network of 500 eps which wil make efetive ‘so of both horns ana good ehaoe of feequcney from the standpoint of the py ‘ology of stning “The are constant is occ Ba. (9.28) = toh 300 29) Loom P,W, Klipsch, A Low-frequency Horn of Small Dimensions, J. Acoust Soe. Amer, 48: 157 1s amit), Port XXII] HORNS, 283 Let us assume that the driver unit isthe one discussed in Example G-1, For this, Spm 814 X 10-4 HoRZONTAL SECTION ‘ecton2-20 VERTICAL SECTION TRANSVERSE SECTION {ecton 3:3) Fra. 9.18 Sketches for a Klipsch type of folded exponential horn, ‘This particulay hora is about 40 in. high and has smooth response below 200 eps. "The horn should radiate sound well at 400 eps, so that the mouth-opening are: should be, if possible, greater than that given by Eq. (9.16), aa Su = Gap FeC400)! ‘As we learned in Chap. 4, in order to neta wide directivity pattern, say 30° over: wide range of irenucnetes, the horn should have a evrved mouth. Let us select evign that is shont (in. i hesghtarvd has a eizcular curved mouth with an are lengt? fof 30 in, The mouth area for these dimensions is 180 in%, or 0.1163 m4, which | Alowble that called for shove2nd HON LOUDSPEAKERS IChep.9 “The length of the horn is found from Eq, (@.17), (1109) x 10 Gain x smo me = 591 521 ose m = 2130 7 = 32) ost m = 215 ‘The horn will have the shape and cross section shown in Fig. 9.19. ‘The cutoff frequeney is at 300 eps and ie far enough below the 500 eps erossover frequency’ sn that the throat impedance will be resistive over the entire useful range fof the loudepeaker; hence thete i no real need to halence (aC) against the mass = LoneTUoINAL SECTION onisers Paw Fic. 9.19. Plan for simple straight exponential horn with a cutoff frequency of 800 cpa, a throat area of 3.14 x 10-*m%, and a mouth area of 11.0. 10°? m?. The dividers guide the wave along the horn and (end to produce w plane wave front of ‘uniform intensity at the mouth, reactance of the horn, Noverthees, let us enleulate the value of Cas from Eat (912), (248), and (28), 28 Cu Spam eam aH x 10D) 5.1 X 10" m/nowton ‘The magnitude of this compliance, alto, is not an unreasonable value to achieve ia 8 Griving unit of the type given in Example 9. Finally, let us determine the power-handling capacity of this horn, The intensity at the throat of the horn Z— in the 8000- to 8000-cps band will equ 4x watt /emt At our upper desiga frequency of 6000 eps, which gives us £ = 200 _ 29, ie we ar fromm Fig. 9.12 thatthe second-harmonic distortion wil equal about 0.4 percent ‘This islow distortion, and we conclude that our desga i satisfartory CHAPTER 10 SOUND IN ENCLOSURES part XXIII Sound Fields in Small Regularly Shaped Enclo- sures 10.1. Introduction. ‘The study of sound in enclosures involves not, only a search into how sounds are reflected backward and forward in an enclosure but also investigations into how to measure sound under such conditions and the effect various materials have in absorbing and con- trolling this sound. Also, of great importance in applying one’s engineer ing knowledge of the behavior of sound in such enclosed spaces is an understanding of the personal preferences of listeners, whether listening in the room where the music is produced oF listening at a remote point to ‘a microphone pickup. Psychological criteria for acoustic design have occupied the attention of many investigators and must always be borne in mind. This chapter is confined to physical acoustics. Psychological factors will be discussed in Chap. 13, which deals with psychoacoustical phenomena and criteria for acoustic design. ‘Two extremes to the study of sound in enclosures can be analyzed and understood easily. At the one extreme we have small enclosures of simple shape, such as rectangular boxes, cylindrical tubes, ot spherical shells, In these cases the interior sound field is describable in precise mathematical terms, although the analysis becomes complicated if the walls of the enclosures are covered in whole or in part with acoustical absorbing materials. AL the other extreme we have very large irregularly shaped enclosures where no precise description can be made of the sound field but where a statistically reliable statement can be made of the average conditions in the room. This is analogous to a study that a physician might make of a particular man to determine the number of years he will live, as opposed to a study of the entire population on a statistical basis to determine how long a man, on the average, will live. As might be expected, the statis. 205286 SOUND IN ENCLOSURES Chap. 10 tical study leads (o simpler formulas than the detailed study of a par ticular case. 10.2. Stationary and Standing Waves. One type of small regularly shaped enclosure, the rigidly closed tube, has heen discussed alre Part IV. ‘This case provides an excellent example of the acoustical tion that exists in large enclosures. First, we noted that along the x axis of the tube the sound field could be described as the combination of an outward-traveling wave and a backward-traveling wave. Actually, the outward-traveling wave is the sum of the original free-field wave that started out from the source plus the outward-going waves that are making their second, third, fourth, and so on, round trips. Similarly, the backward-traveling wave is a combination of the first reflected wave and of waves that are making the return leg of their second, third, fourth, and so on, round trips. These outward- and backward-traveling waves add in magnitude to produce what is called a stationary wave! if the intensity along the tube is zero or what is ealled a standing wave if there is absorption at the terminating end of the tube so that, power flows along the tube away from the source Gantensity not equal to zero) 10.3. Normal Modes and Normal Frequencies. We saw from Bq, (2.48) that. whenever the driving frequency is such that sin kl 0, the pressure in the tube reaches a very large value. ‘That is to say, the pressure is very large whenever a0.) ‘Then, because we have (10.2) (10.3) where ergs see (10.4) fa = nth resonance (normal) frequeney of the tube As = ¢/fa = nth resonance (normal) wavelength of the tube Waquation (10.3) tells us that the pressure is very large whenever the length of the tube equals some integral multiple of a half wavelength 2). ‘The condition where the frequeney equals ne/2l so that a very large sound pressure builds up in the tube is called a resonance condition oF & "Phe definitions for standing snd stationary waves are found in“ Amerean Stands lard Acoustical Terminolegy,” 22411961, American Standards Association, Ine., New York NY Part XXL] SOUND FIELDS IN SMALL ENCLOSURES 287 normal mode of vibration of the air space iu the tube. ‘The frequency fa of a normal mode of vibration is called a normal frequency. ‘There are an infinite number of normal modes of vibration for a tube because n can take on all integral values between 0 and infinity. We may look on the tube, or in fact on any enclosure, as a large assemblage of acoustic resonators, each with its own normal frequency. In the elosed-tube discussion of Part IV, we made no mention of the effect on the results of the cross-sectional shape or size of the tube. It was assumed that the transverse dimensions were less than about 0.1 wavelength so that no transverse resonances, ‘would oceur in the frequency region of interest. te If the transverse dimensions are greater than one-half wavelength, we gl haveasmalloom which, ifrectangular, 7g te can be described by the dimensions pation asaicores shown in Fig. 10.1. Waves can travel Sloane’ ©” {oF * Teetenaslar in the room backward and forward between any two opposing walls. They can travel also around the room involving the walls at various angles of incidence. If these angles are chosen properly, the waves will return on themselves and set up stationary or standing waves. Each standing wave is a normal mode of vibration for the enclosure. Tt would be interesting in such rectangular enclosure to solve mathe- matically and to deseribe exactly the distribution of sound as determined by the strength and type of source. ‘That study is beyond the scope of this text. We shall describe, however, the simplest eases and suggest extra reading for those interested. ‘The number of modes of vibration in a rectangular enclosure is much greater than that for the rigidly closed tube whose diameter is small eom= pared with « wavelength. In fact, the normal frequencies of such an enclosure are given by the equation, 4 4 ye (10.5) integers that can be chosen separately. ‘They may take on all integral values between 0 and @. dimensions of the room in meters. speed of sound in meters per second. +P. M. Morse, “Vibration and Sound,” 2d ed, Chap. VIIE, McGraw-Hill Book Company, Ine, New York, 194%, *P. Mf Morse and RH. Bolt, Sound Waves in Rooms, fe, Md, Phys, 16: 69-150 (1944)

Acoustics Leo Beranek PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Acoustics Leo Beranek PDF

Uploaded by

Copyright:

Available Formats

You might also like