硕士论文-应用于小麦克风阵列技术的SigmaDelta模数及

上海交通大学硕士学位论文应用于小麦克风阵列技术的Sigma-Delta模数及数摸转换器数字电路部分的设计与实现姓名：卫强申请学位级别：硕士专业：通信与信息系统指导教师：戎蒙恬20070206Sigma-Delta10mbitdBKhzmVMhznspFVΩmµSigma-Delta220061016Sigma-Delta320061016200726Sigma-Delta4Sigma-Delta,(Beamforming),ADCDACDSPMCUCMOSNyquistSigma-Delta5Sigma-DeltaNyquistNyquistSigma-DeltaSigma-DeltaDEMSigma-DeltaDACSigma-DeltaToneSigma-Delta130dBSigma-DeltaSigma-DeltaFIRFIRSigma-DeltaFIRRAMROMSigma-Delta6MatlabROMΣ∆Sigma-Delta7DIGITALCIRCUITDESIGNOFSigma-DeltaVOICEDACANDADCFORTHESMALLARRAYMICROPHONEABSTRACTThetechniqueoftheArrayMicrophonecreateabeamformthatcanmakeadivisionbetweentheusefulvoiceandnoisethroughthedifferentdelayofvoicetothemicrophoneinthearray.So,itcanhavethehighperformanceonechocancellationandnoisesuppress.Includingthenon-linearanddynamicnoisethatisdifferenttohandlebythetraditiontechnique.Butthistechniquehaveavisibledisadvantagethatisinordertogettheenoughsoundwavedelaytocreatethebeamforming,therearemoredistancebetweenthemicrophoneofthearraySothevolumeofequipmentisbigandcannotbeusedontheportablefield.Withthecontinuousdevelopmentofsignalprocessingtowardsthegoalofhighprecision,lowcostandhighdensity,ICcanbedesignedtorealizecomplicateddigitalsystemanddigitalsignalprocess.SoitispossibletodesignanAISCthatintegratedmulti-channelvoiceADC,voiceDAC,DSPandMCUtocreateaSmallArrayMicrophone(RelativeSigma-Delta8totheArrayMicrophone,thedistanceofthemicrophoneinarraycanlessmore).Aswellknow,oneofthekeytechnologiesabouttheSAM(SmallArrayMicrophone)ishowtodesignthevoiceADCandvoiceDACthatcanbeflexibletovarysignalNyquistsample-rateandthenumberofADCchannelwiththelowpowerconsumeandsmalldiesizesolution.Inthisdiscourse,studythevarioustheoriesandmethodologyofthedataconverters,researchthedigitalcircuitdesignandrealizationofaneasyintegrated,flexiblesetting,lowpowerandsmalldiesizeanalogtodigitalanddigitaltoanalogvoicesignalSigma-DeltaconverterswhichcanwellmatchtherequirementofSmallArrayMicrophonetechnology.First,analysesthelimitofthetraditionalNyquistratedataconvertersandtheadvantageoftheSigma-DeltaconverterstohandlethehighprecisionconversionforthelowNyquistsample-ratesignal.Then,analysesthecharacteristicofvariousSigma-DeltamodulatorandaffectiontotheSNR,powerconsumeanddiesizewithdifferentordernumberofthemodulator,oversamplerateandbitwidthofquantization.Baseonthat,willofferthedesignandrealizationofatwoorder,7bitquantizationmodulatorwhichwilluseasegmentednoiseshapingtechnologyinordertoreducetheDEM(DynamicElementMatch)circuitsize.ThismodulatorusingthesimplecircuitwillavoidtheunstableissueofthehighordermodulatorwiththelessidlechanneltoneandpowerconsumeAtsametime,thetheoreticSNAofthismodulatoris130dB.Sigma-Delta9IntheSigma-Deltaconverters,thesignalwillbeprocessedbythecomplicateddigitaldecimationandinterpolationfiltercircuitwithdigitalsignalprocesstechnologyThispartdesignwilldecidetheallsystemdiesizeandpowerconsume.Sothedissertationwillstresstoresearchhowtodesignaneffectivelowcostandlowpowerfilterinboththedecimationandinterpolation,includingthearchitecturedesignandcircuitrealization.Inthephaseofarchitecturedesign,baseonthecharacteristicofthedecimationandinterpolationfilterintheSigma-Deltaconverters,amultistageoptimizationarchitectureFIRwillbeused.Aboutthecircuitrealization,willoffertheoptimizationmethodologyoftheRAMaddressmappingforusingtheROMprogrammethodologytorealizethemultistagedecimationandinterpolationFIRfilterforthelowpowerconsumeandsmalldiesize.Atsametime,someMatlabprogramwaswroteastheassistanttodesignthehighperformanceFIRfilter.Finally,thedissertationwillgivethesimulationresultsofthealldigitalcircuitoftheSigma-DeltaA/DandSigma-DeltaD/AKEYWORDS:ArrayMicrophone,A/Dconverter,D/AconverterSigma-deltamodulator,decimationfilter,interpolationFIRfilter,noiseshapingSigma-Delta151.1,VoIP10.3~1m0.1m,2,,,,Sigma-Delta16ADCDACDSPMCUADCDAC(1)CMOS(2)NyquistDSP16bit(3)(4)FortemediaSigma-DeltaSigma-DeltaADCDAC0.18mµCMOSADCDAC.ADCsf16Khzsf32KhzNyquist.DACsf8Khzsf16KhzNyquist.0.425sf.0.05dB±.ADCSNR96dBTHD92dB16bit[-3276832767].DACSNR96dB;THD86dB;16bit[-3276832767];1.2NyquistADCSigma-DeltaSigma-DeltaSigma-Delta17thermomtercodeSigma-DeltaDACSigma-DeltaADCFIRSigma-DeltaADCSigma-DeltaDACROMRAMRAMROMFIRSigma-Delta18ADC2-1)(tXBf)(KTX)(KTYsfbfNyquistNyquistNyquistADCNyquistADCNyquist)(M)BfSf(tX)(KTX)(KTY2-1Figure2-1StructureofA/DConverter2.1NyquistNyquistADC1.)(ne2.)(ne)(nx3.)(neSigma-Delta19)(ne)(ne0)()]([===∫∞∞−deeepneEme(2-1)222222312)()(]))([(NEdeepmemneEeee=∆=−=−=∫∞∞−σ(2-2)∆≤∆=021)(eep(2-3))(ne),(EE+−N∆NE2=∆nnN2=neE22223⋅=σ76.102.62lg102lg203lg10)(222+=⋅++=nEEdBSNRn(2-4)6dBNyquistADCADCNyquistADCbit10~8bit14~12NyquistCMOS100dBbit1616210.0015%[1][2]CMOSNyquistNyquistSigma-Delta20NyquistNyquist2.2Sigma-DeltaNyquistSigma-Delta2.2.1Sigma-DeltaNyquistADCNyquistt∆t∆sfNyquist2-201)('nx1t1)(netSigma-Delta21∆⋅=sfDDQuantizerIntegratorand1bitD/Ay(n)Sf)('nx)(nx)(ne2-2Figure2-2DeltaModulatorDSigma-DeltaSigma-Delta111)(−−=zzH)(neKX1XD/AZ-1Sigma-Delta2-3111−−Z1−Z2-3Sigma-DeltaFigure2-3First-OrderSigma-DeltaModulator)()()()()(zEzHzXzHzYex+=(2-5)1)1(1)(1)()(−−