ie3d教程chap8.docx

ie3d教程chap8.docx

《ie3d教程chap8.docx》由会员分享，可在线阅读，更多相关《ie3d教程chap8.docx（24页珍藏版）》请在冰豆网上搜索。

ie3d教程chap8

Chapter8MagneticCurrentModelingofApertureStructures

Sofar,wehavebeendiscussingthemodelingofelectriccurrentinelectroniccircuits.Basically,weareusingpolygonstorepresenttheelectricconductorsandwearesolvingthecurrentdistributionontheelectricconductors.Therearecircuitsorantennasthatutilizingslotsinelectricconductorstoachievespecificperformances.Forsuchcircuitsorantennas,itwouldbebettertomodelthetangentialelectricfieldacrosstheslotsinsteadofthecurrentdistributionontheelectricconductors,becausetheareaofslotsismuchsmallerthantheareaoftheconductors,andthecomputationaleffortwillbesignificantlysmaller.Theoretically,tangentialelectricfieldacrossaslotisequivalenttothemagneticcurrentontheslot.Wewillcallsuchamodelasmagneticcurrentmodel.

Inthischapter,wewilldemonstratehowwecanapplythemagneticcurrentmodelingtosolvesomeproblemsefficiently.MagneticcurrentmodelingisonlyavailableonIE3DPowerPack.ForthoseuserswhoseIE3Dlicensesdonothavethemagneticcurrentcapability,youcanchoosetoskipthischapteroryoucanfollowitforpractice.FollowingtheexamplescanstillhelpyoutopracticeeditingonMGRID.ForthoseIE3Dlicenseswithoutthemagneticcurrentcapability,nosimulationinvolvingmagneticcurrentisallowedevenyoucanstillbuilditonMGRID.Therefore,youwouldnotbeabletocreatethecurrentdistributionfile.

Fromtheelectromagnetictheory,theslotelectricfieldisequivalenttothefictitiousmagneticcurrentdistributionbyEquation（8-1）.

Jm=E⨯n（8-1）

whereJmistheequivalentmagneticcurrentdensity;Eisthetangentialelectricfieldonaslot;nistheunitnormalvectoroftheslot.Ausershouldunderstandthatthemagneticcurrentmodelisbasedupontheassumptionthattheslotisonaflat,largeandverygoodelectricconductor.Inreality,theelectricconductorisalwaysfiniteanditsconductivityisalwaysfinite.Theaccuracyofthemodelwilldependuponhowbigthegroundplaneisandhowgoodtheelectricconductoris.

Section1.ModelingaSlotAntennawithCPW-feedorGapPortFeed.

ThefirstexampleforthischapterisaCPWfedantenna.ItispublishedbyH.S.TsaiandR.A.York,“Applicationsofplanarmultiple-slotantennasforimpedancecontrol,andanalysisusingFDTDwithBerenger’sPMLmethod,”IEEEAP-SymposiumDigest,NewportBeach,CA,1995,pp.370-373.ThestructureisshowninFigure8.1.Thebasicparametersare:

SubstrateThickness=0.813mm,Permittivity=2.2,Conductivity=4.9107s/m,SlotLength=78mm,Width=2mm,CenterConductorWidth=2mm.Thecoloredpolygonrepresentsaslotonagroundplaneevenitlookslikeitisametallicpolygon.

78mm

Figure8.1ACPW-fedslotantennatobeanalyzedusingmagneticcurrentmodeling.

Buildingthepolygontorepresenttheslotshouldbesimplewithwhatyouhavelearned.Therearemultiplewaystobuildit.Youcanbuildtheslotasapath.Youcanalsobuildtheoutershapeasarectangleof78mmby6mm.Then,youbuildasmallerrectangleof74mmby2mmonanotherlayer.YoucanusethesmallerrectangletobuildholeonthebiggerrectanglefortheshapeinFigure8.1.Certainly,youwillneedtodosomeextraworktodefinetheports1and2anditsproximity.Thestructureispre-builtandsavedinfile:

.\ie3d\samples\slot.geo.WewillconcentrateourdiscussiononthesubstratesetupandtheLayerWindow.Figure8.2showsthelayout,theLayerWindowandtheBasicParameters.YoumaynoticewehavedefinedtwosubstratelayersatZ=0.813.Basically,thesubstratesetupmeansthefollowing:

From-∞to0isair.From0to0.813mmisthesubstrate.From0.813to0.813mmisthegroundplane.Then,from0.813mmto∞isairagain.

ItisnotallowedtobuildafinitethickgroundsubstratelayeronIE3Datthistime.However,youcandefineonenormalsubstratelayerbetween2infinitelythingroundplanestomodelthefinitethickgroundplane.

OntheLayerWindow,youseeahorizontaldashlineacrosstheNo.1LayeratZ=0.81300.Itmeansthatthepolygonlayerisaninterfaceofgroundplane.Any2Dpolygononitisassumedtobeaslotontheground.Thepolygonsonthelayouteditormeantheareatobeetchedawayfromthegroundplane.

Figure8.2Thestructure,theLayerWindowandtheBasicParametersdialog.

WehavedefinedadifferentialpairofExtensionforMMICportonit.FormagneticcurrentmodelingExtensionforMMICschemeandtheAdvancedExtensionschemearethesame.Thequestionis,whatdoesthedifferentialpairmean?

Ifthepolygonsaremetallicpolygons,thedifferentialpairmeansitisapairofcoupledstrip.Formagneticcurrentmodeling,thedifferentialpairmeansitisaCPW.WewillexplainthereasoninSection2.

Figure8.3Theslotantennafedbyahorizontallocalizedport.

Isthistheonlychoiceforaportonmagneticcurrentmodeling?

Itcertainlyisnot.Savedin.\ie3d\samples\slot1.geoisthesamegeometryfedbyaHorizontalLocalizedport.

AHorizontalLocalizedportisbasicallyagapport.Ifthepolygonsaremetallicpolygons,thestructureinFigure8.3meansafoldeddipolewithagapportbetweenthetwoterminals.Forthemagneticcurrentmodeling,thepolygonsrepresentslotsontheground.TheHorizontalLocalizedportmeansanexcitationacrossthegapattheportlocation.MoreexplanationcanbefoundinSection2.

HowcanwedefinetheHorizontalLocalizedportonit?

Itissimple.

Step1RunMGRID.Open.\ie3d\samples\slot1.geo.SelectPort->DeleteAllPorts.Wewillseethestructureconsistsoftwopolygons（seeFigure8.4）.Polygon2isthe“foldeddipole”whilethepolygon1isarectanglewithbothendsconnectedtotheendsofthe“foldeddipole”.

Figure8.4Theslot1.geowithouttheportdefined.

Figure8.5TheHorizontalLocalizedPortdefinitiondialog.

TherearetwowayswecandefinetheHorizontalLocalizedportonit.Wearegoingtoshowthebothways.

Step2SelectEdit->SelectPolygonGroup.Windowtherectangle1toselectit.SelectPort->SelectedRectanglesforHorizontalLocalizedPort.MGRIDwillpromptyoutheHorizontalLocalizedPortdefinitiondialog.Itdetectstwopossibledirectionsfortheport.Wewantthepositiveterminaltobeatthelefthandsideofthegapandthenegativeterminaltobeattherighthandsideofthegap（seeFigure8.3）.Weshouldchoosethe2ndPhi=180degree.Pleasenoteyoumaybepromptedwith“1stPhi=180and2ndPhi=0”inyourrun.YoushouldchoosetheonecorrespondingtoethePhi=180.SelectOKtocontinue.WewillgettheportdefinedasshowninFigure8.3.

Ifitisforelectriccurrent,theportdefineasourcetoapplyavoltagebetweenthepositiveterminalandthenegativeterminal.Formagneticcurrentmodeling,thereisanambiguityfortheportdirection.WewilldiscussthisissuelaterinSection2.Let’strythedifferentwaytodefinetheHorizontalLocalizedPort.

Step3TheHorizontalLocalizedPortschemehasbeenimplementedafewyearsago.ThewaytodefineanHorizontalLocalizedPortwejustdemonstratedwastheonlyway.However,manyusersdonotrealizetheexistenceofsuchaschemebecausethewayofdefiningitisdifferentfromotherschemes.Forthisreason,wehaveimplementedawayclosertodefiningportsinotherschemes.

PleaseselectPort->DeleteAllPortstodeletetheportagain.SelectPort->PortforEdgeGroup.TheDe-EmbeddingSchemedialogcomesup（seeFigure8.6）.PleasechooseHorizontalLocalized.SelectOKtocontinue.

Figure8.6TheDe-EmbeddingSchemedialogwithHorizontalLocalizedschemeselected.

Figure8.7Thecomparisonoftheslotantennawith2differentports.

Step4Pleasewindowthepolygon1inFigure8.4.MGRIDwillpromptyoufortheHorizontalLocalizedPortDefinitiondialoginFigure8.5.Selectthe2ndPhi=180degreesandselectOK.MGRIDwilldefinetheHorizontalLocalizedportonslot1.geo.SelectPort->ExitPorttoexitthemode.Pleasesimulateandcomparethes-andz-parametersoftheslot.geoandslot1.geo（seeFigure8.7）.Thes-parametersarealmostthesameexceptthereissomesmalldifferenceatthedip.Thereissomeslightshiftinthez-parametercurves.Itisduetothefactthatthedifferenceinthefeedproximitycausessomedifferenceinthereactancepartintheimpedance.

Section2.DiscussionofPortsonSlotStructures.

InSection1,wehavediscussedhowtomodelaslotantennausingdifferentialfeedschemes.However,whatarethemeaningsoftheschemesdiscussed.Inthissection,wewillillustratethedifferentports.

Inthe.\ie3d\practice\slot.geo,weareusingdifferentialExtensionforMMICport（seeFigure8.2）.Infact,the“+1”portisapplyingavoltagebetweenthemetalonitsrightsidetothemetalonitsleftside.The“-1”portisapplyingavoltagebetweenthemetalonitsleftsidetothemetalonitsrightside.ItisequivalenttoapplyingavoltagebetweenthecenterconductorandthetwosideconductorsontheCPWstructure（seeTable8.1）.Basically,thetwopolygonsarerepresentingthetwoslotsofCPW.

Inthe.\ie3d\practice\slot1.geo,weareusingtheHorizontalLocalizedport（seeTable8.1）.Whatitmeansisthatweareapplyingavoltageacrossthegap.Itisequivalenttoagapportintheelectriccurrentmodel.

Certainly,thefeedproximityisdifferentbetweenCPWfeedandtheGapPortfeedwhenwemodeltheslotantenna.ItisnotsurprisingthattheymayyieldslightdifferentresultsinFigure8.7.

Someusersareconsideringwhetherwecandefineasingle-endedportonapolygonformagneticcurrentmodeling.Theyareconsideringthatthesingle-endedportin“ExcitationacrossaGap”inTable8.1