Modeling Of Horns And Enclosures For Loudspeakers (Gavin Richard Putland).pdf

(1273 KB) Pobierz
448917685 UNPDF
Modelingof
HornsandEnclosures
forLoudspeakers
by
GavinRichardPutland,BE(Qld)
DepartmentofElectricalandComputerEngineering
UniversityofQueensland
SubmittedforthedegreeofDoctorofPhilosophy
December23,1994
RevisedNovember1995
AcceptedFebruary6,1996.
ii
Tomyparents,
FrankandDelPutland
Acknowledgments
MyinterestinhorntheorywasinspiredbyDrE.R.Geddes,whosepaper“Acoustic
WaveguideTheory”[18]raisedthequestionsaddressedinChapters3to5ofthis
thesis,suggestedsomeoftheanswers,andprovidedmanyoftherelatedreferences.
ThesedebtsarenotdiminishedbymydisagreementwithGeddes’analysisofthe
oblatespheroidalwaveguide,whichhesubsequentlyrevised[19].Mypaperonone-
parameterwavesandWebster’sequation[43]wasimprovedasaresultofapersonal
communicationfromDrGeddes(October10,1992),whichalertedmetotheexis-
tenceofFresneldi‹ractionfringeswithinthecoverageanglesoffinitehornsathigh
frequencies.
IwishtothanktheExecutiveEditorandReviewBoardofthe Journalofthe
AudioEngineeringSociety fortheirpatienthandlingofthecomplexexchangeof
correspondencethatfollowedmy“Commentson‘AcousticWaveguideTheory’”[42].
Iamindebtedtoananonymousreviewerofmyarticle“AcousticalPropertiesof
AirversusTemperatureandPressure”[44]forcorrectingmyexplanationofthee‹ect
ofhumidityonabsorption,andfordrawingmyattentiontoANSIS1.26-1978[41],
whichiscitedfrequentlyinthefinalversionofthearticleandinChapter9.
Mysupervisor,DrLarrySkattebol,providedcriticalfeedbackonmythreepubli-
cationsandonnumerouspartialdraftsofthisthesis.Ithankhimforhisflexibilityin
acceptingmyresearchproposal,andforhissubsequentprudenceincurbingmyten-
dencytopursuenewlinesofinquiry;withouthisguidance,thisalreadyprotracted
projectwouldhavetakenconsiderablylonger.IalsothankProf.TomDownsand
DrNickShuley,colleaguesofDrSkattebolintheDept.ofElectricalandComputer
Engineering,UniversityofQueensland,fortheircommentsonreference[42].
TheItaliantextofSomigliana’sletter[51](paraphrasedinAppendixA)was
translatedbyBr.AlanMoss,whowasthenagraduatestudentintheDepartment
ofStudiesinReligion,UniversityofQueensland.
TheJapanesetextofSections1to3ofArai’spaper[2]wastranslatedbyAdrian
TreloaroftheDepartmentofJapaneseandChineseStudies,UniversityofQueens-
land.
Thetitlesandauthorsofthemajorpublic-domainsoftwareusedinthisproject
areasfollows.EquivalentcircuitsimulationswereperformedusingSPICE3byTom
Quarles,withthe nutmeg userinterfacebyWayneChristopher.Programswere
compiledbytheGNUprojectCcompiler.Mostdiagramsweredrawnusing xfig
bySupojSutanthavibuletal.,convertedtoL A T E X picture commandsusing fig2dev
byMicahBecketal.,andfinallyeditedastextfiles.Allsourcefileswereedited
with jove byJonathanPayne,andspell-checkedwith ispell byPaceWillissonet
al.ThisdocumentwastypesetandprintedwithintheDepartmentofElectricaland
ComputerEngineering,UniversityofQueensland,usingL A T E X2 " byLeslieLamport.
ThisprojectwassupportedbyanAustralianPostgraduateResearchAward.
iii
StatementofOriginality
Ideclarethatthecontentofthisthesisis,tothebestofmyknowledgeandbelief,
originalexceptasacknowledgedinthetextandfootnotes,andthatnopartofithas
beenpreviouslysubmittedforadegreeatthisUniversityoratanyotherinstitution.
Inathesisdrawingonsuchdiversesubjectsasmechanics,thermodynamics,
circuittheory,di‹erentialgeometry,vectoranalysisandSturm-Liouvilletheory,the
authormustmakeexaggeratede‹ortstomaintainsometheoreticalcoherence.This
mayinvolvederivingwell-knownresultsinamannerappropriatetothecontext;
examplesincludethehierarchyofformsoftheequationsofmotionandcompression
(Chapter2),andthederivationofadmittancesandGreen’sfunctionsfromWebster’s
equation(Chapter3).Furthermore,whenathesiscontainsresultsatvariancewith
earlierresultsintheliterature,theauthorwillbeexpectedtojustifyhisfindingswith
exceptionalthoroughness.Inparticular,hemaybeobligedtoconductmathematical
argumentsatamorefundamentallevelthanwouldnormallybeappropriate;an
exampleismydetailedsolutionoftheheatequationtodeterminethebasicthermal
timeconstantoftheair-fibersystem(Chapter8).Insuchcases,thecontextwill
indicatethattheresultisincludedforthesakeofclarity,cohesionorrigor,andnot
necessarilybecauseofnovelty.
Toavoidexcessiverelianceon“thecontext”,Io‹erthefollowingsummaryof
whatIbelievetobemyprincipaloriginalcontributionsandtheirdependenceon
theworkofearlierresearchers.Thissummaryalsoservesasanextendedabstract.
InChapter2,Ihaveshownhowthenumerousfamiliarequationsrelatedto
theinertiaandcomplianceofaircanbeunderstoodasalternativeformsof
twobasicequations,whichIcalltheequationsofmotionandcompression.
Thehierarchyofformseliminatesredundancyinthederivationsandclearly
showswhatsimplifyingassumptionsareinvolvedineachform.The“one-
parameter”or“1P”formsapplywhentheexcesspressure p dependsona
singlespatialcoordinate ,whichmeasuresarclengthnormaltotheisobaric
surfaces.Theseformsareexpressedwithunprecedentedgenerality,andare
criticaltothemathematicalargumentofsubsequentchapters.Otherforms
leadtothefamiliarelectricalanalogsforacousticmassandcompliance(both
lumpedanddistributed).
Afterareviewofpreviousliterature,Ihaveshownthatthe“Webster”horn
equation,whichisusuallypresentedasaplane-waveapproximation,follows
exactly fromthe1Pformsoftheequationsofmotionandcompression,without
anyexplicitassumptionconcerningthewavefrontshape.The coordinateis
theaxialcoordinateofthehornwhile S ( )istheareaofaconstant- surface
segmentboundedbyatubeoforthogonaltrajectoriestoalltheconstant-
surfaces;suchtubes(andnoothers)arepossibleguidingsurfaces.
iv
v
IhaveshownthattheHelmholtzequationadmitssolutionsdependingona
singlespatialcoordinate u ifandonlyif jruj and r 2 u arefunctionsof u alone.
The jruj conditionallows u tobetransformedtoanothercoordinate ,which
measuresarclengthalongtheorthogonaltrajectoriestotheconstant- sur-
faces.Hence,inthedefinitionofa“1P”pressurefield,thenormal-arc-length
conditionisredundant.UsinganexpressionfortheLaplacianofa1Ppres-
surefield,Ihaveshownthatthewaveequationreduces exactly toWebster’s
equation;thisisasecondgeometry-independentderivation.Ihavealsoshown
thattheterm“1Pacousticfield”canbedefinedintermsofpressure,velocity
potentialorvelocity,andthatallthreedefinitionsareequivalent.
Ihaveexpressedthe1Pexistenceconditionsintermsofcoordinatescalefac-
torsandfoundthat intheelevencoordinatesystemsthatareseparablewith
respecttotheHelmholtzequation ,theonlycoordinatesadmitting1Psolutions
arethosewhoselevelsurfacesareplanes,circularcylindersandspheres(Chap-
ter5).Geddes[18]reportedin1989thatWebster’sequationisexactinthe
samelistofcoordinates,butdidnotmaketheconnectionbetweenWebster’s
equationand1Pwaves.
Withoutusingseparablecoordinates ,Somigliana[51]showedthatthereare
onlythree1Pwavefrontshapesallowingparallelwavefrontsandrectilinear
propagation;thepermittedshapesareplanar,circular-cylindricalandspheri-
cal.Ihaveshown(Chapter5)thattheconditionsofparallelwavefrontsand
rectilinearpropagationareimplicitinthe1Passumption,sothatthethree
geometriesobtainedbySomiglianaaretheonlypossiblegeometriesfor1P
waves.Thisresulthasthepracticalimplicationthatnonew1Phorngeome-
triesremaintobediscovered.
Ihaveproducedanannotatedparaphrase,inmodernnotation,ofSomigliana’s
proof(AppendixA),andadaptedhisproofsoastotakeadvantageofthe
1Pexistenceconditions(Chapter5).InthetheoremsofChapter5andthe
footnotestoAppendixA,IhavefilledinseveralmissingstepsinSomigliana’s
argument.Ifounditmostconvenienttoprovetheseresultsindependently,
althoughrelatedresultsexistintheliteratureondi‹erentialgeometry.
Workingfromthepermitted1Pwavefrontshapesandtheexactderivationsof
Webster’sequation,Ihavegivenwideconditionsunderwhichthatequationis
approximately true,sothattraditionalapproximatederivationsoftheequation
canbereplacedbythemoregeneral1Ptheory.
Ihaveextendedthefinite-di‹erenceequivalent-circuit(FDEC)methodpro-
posedin1960byArai[2].InChapter6,Ihaveshownthatafinite-di‹erence
approximationtoWebster’sequationyieldsthenodalequationsofan L-C
latternetwork(confirmingArai’sunprovenassertionthathisone-dimensional
methodcanbeadaptedforhorns),whileasimilarapproximationtothewave
equationingeneralcurvilinearorthogonalcoordinatesyieldsthenodalequa-
tionsofathree-dimensional L-C network.Ihaveobtainedthesamecircuits
fromtheequationsofmotionandcompressioninordertoshowthat“current”
intheequivalentcircuitisvolumeflux,asexpected.Ihaveshownhowthe
networkshouldbetruncatedattheboundariesofthemodelandterminated
withadditionalcomponentstorepresentarangeofboundaryconditions.
Zgłoś jeśli naruszono regulamin