数字预失真关键技术-ppt2

第二章数字预失真技术理论第一节线性化技术概述第二节数字预失真基本原理第二节数字预失真线性化系统第三节数字预失真器及其参数辨识第三节数字预失真器及其参数辨识第四节数字预失真器参数辨识算法第五节峰均值比及削峰技术简介第六节数字预失真器的仿真第六节数字预失真器的ADS仿真线性度和功率效率2功率回退3高功率射频功放•HighPowerRFPA’s(10W)usemultipledriverstagestoamplifyaninputsignalinputsignal.•DifferentPAarchitecture’s(ClassA,AB,C,etc…)offervariousdegreesoflinearity,costandefficiency.–RFPA’sarenotoriouslyinefficient–Airisaconvenientbutpoortransmissionmedium.yp•RFPA’saredesigned(tuned)forspecificfrequencyrangeandbandwidth–MCPA~=widebandRFPA,doesnothavetoprocessmultiplecarriers•PAGainisusuallyfixed–sopre-ampsmayberequiredtodrivethePAinput.RFout=RFin==AntennaPATXBoardIfGain=30dB,50dBm(100W)RFin20dBm@800MHzDACIF-RFDUCFromBaseband50OhmTypicalAPre-Amp3to4gainstagestypicalInput4线性化、峰值因子减小•LinearizationtechniquesallowaPAtobeoperatedathigherpowerwithminimalIMDincreasesthusgreaterefficiencywithminimalIMDincreases,thusgreaterefficiency•Recenttechnologicaladvanceshavemadedigitalpre-distortionthefocusofresearcheffortsCtftdti(CFR)fthithffiifthPA•Crestfactorreduction(CFR)furtherincreasestheefficiencyofthePAbyreducingthepeak-to-averageratio(PAR)ofthetransmittedsignalTheoreticalPerformanceofClassABPAPre-DistortionNoNoYesYesCFRNMdMdYTheoreticalPerformanceofClassABPACFRNoModerateModerateYesTxPower10W10W10W10WPAR12dB9dB9dB6dBBackoff15dB12dB9dB6dBPAPowerRating320W160W80W40WEfficency5%9%18%30%Efficency5%9%18%30%PowerDissipation120W101W45W7W5线性化基础线•Thetrade-offbetweenefficiencyandlinearityistheprimaryconcernforPAtheprimaryconcernforPAdesigners•APAoperatingatahighpercentageofitspowerpercentageofitspowerratingrequiresexternallinearizationtomaintainlinearity•ThelinearizationofthePAreducesback-off,thusincreasingefficiency6线性化技术总览线性化技术总览7线性化的优势TechnicalComparisonofBFLPAvs.LPAComparisonFactorBruteForceLinearPowerAmplifier(BFLPA)LinearizedPowerAmplifier(LPA)„P1db=56dBm(400Watts)„P1db=50dBm(100Watts)TypicalPerformances„P1db=56dBm(400Watts)„Back-off=13dB„C/IM3=46dBcP=43dBm(20Watts)„P1db=50dBm(100Watts)„Back-off=7dB„C/IM3=46dBcP=43dBm(20Watts)„P=43dBm(20Watts)„Efficiency≈5%„P=43dBm(20Watts)„Efficiency≈15%PowerCPDC=400WattsPDC=133WattsConsumptionPDC400WattsPDC133Watts8线性化的优势WithlinearizationWithoutlinearizationNetworkPlanningTradeoffs9预失真、前馈、后馈三种线性化技术比较预失真、前馈、后馈三种线性化技术比较WideNarrowUltra/MediumFrequencyFeedforwardFeedbackRFPredistortion/DigitalPredistortionUltraGoodMedium/UltraLinearityWideNarrowUltra/MediumbandwidthLowHighHigh/HighPowerEfficiencyHighMediumMedium/MediumComplexityIntrinsicallynonadaptiveIntrinsicallyadaptiveIntrinsicallynonadaptiveAdaptationogg/goececySSPAsforPAsSSPAsforBaseStationsPAsforMobileStationsTWTAs&SSPAsforSatcomapplications10线性化技术比较线性化技术比较CorrectionCapability*CorrectionBandwidthRelativeCostLinearizationTechniques25-35dB100MHzHighFeedforwardEnvelopeFeedback10-20dB5MHzMed5-10dB25MHzLow10-20dB50MHzMedAnalogPre-DistortionAdaptivePre-1020dB50MHzMedAdaptivePreDistortion*IMDCtibd8TCtiRdPh*IMDCorrectionbasedon8-ToneContinuousRandomPhase11第二章数字预失真技术理论第节线性化技术概述第一节线性化技术概述第二节数字预失真基本原理第二节数字预失真线性化系统第三节数字预失真器及其参数辨识第四节数字预失真器参数辨识算法第四节数字预失真器参数辨识算法第五节峰均值比及削峰技术简介第六节数字预失真器的ADS仿真12数字预失真原理示意图PredistorterRFoutRFinRFoutRFin|G||G||G|PinPinPin∠G∠G∠GPinPinPin13数字预失真技术数学推导•Pre-distortioneffectivelyperformsamathematicalinversionoftheVolterraPAmodel•Theoutputofthepre-distortionprocessorisdescribedbythefollowing•ThePAislinearizedwhen14数字预失真技术数学推导•Digitalpre-distortion(DPD)hasbecomeaneffectivelinearizationtechniqueduetotherenewedpossibilitiesofferedbyDSPtechniqueduetotherenewedpossibilitiesofferedbyDSP•AdaptivePDdesignsusefeedbacktocompensateforPAvariations•Look-uptablesareupdatedtoachieveoptimalpre-distortionbycomparingPDinputtoPAoutputg•ThePDfunctionisexpressedasacomplexpolynomialwhere15数字预失真频谱•Digitalpre-distortion(DPD)requiresfeedbackforsample-by-sampleadaptation5timesthatofthesignalbandwidth•Multi-carriersystemsusesignalbandwidthsofupto20MHztodaythusMulticarriersystemsusesignalbandwidthsofupto20MHztoday,thusthefeedbackbandwidthmustbe100MHztocompensate3rdand5thorderIMD•Least-mean-square(LMS)isapopulargradientbasedoptimizationalgorithmthatrequireswidebandfeedbackalgorithmthatrequireswidebandfeedback1616数字预失真频谱17第二章数字预失真技术理论第一节线性化技术概述第二节数字预失真基本原理第二节数字预失真线性化系统第三节数字预失真器及其参数辨识第三节数字预失真器及其参数辨识第四节数字预失真器参数辨识算法第五节峰均值比及削峰技术简介第六节数字预失真器的仿真第六节数字预失真器的ADS仿真自适应数字预失真系统框图DiitlAlPredistorterD/AConverterQuadratureModulatorPowerAmplifierDigitalAnalogIinIpdQpdPredistorterD/AConverterModulatorAmplifierLOQinQpdPredistortionLOEstimatorA/DConverterQuadratureDemodulatorIoutQout19自适应数字预失真关键事项„ADCssamplingspeedmustbeatleast5timesthesignalbandwidthandhave12bitsresolutionhave12bitsresolution.„Modulationcanbeperformed@[e-mail].„AnalogueimplementationoftheMOD/DEMODintroducesamplitudeandphaseunbalancesÎdigitalMOD/DEMODispreferable@thecostofincreasingtheprocessingpower/timeneeded.„LimitedACPRimprovementduetoPAmemoryeffectsÎcompensateformemoryeffectsatbasebandprocessing.„ThePAmonitoring/modellingloopmustbeindependentfrommodulationlloop.„Forhighvolumeproduction,atailormadeASICcanbeacosteffectivesolution