关闭

关闭

关闭

封号提示

内容

首页 app_eng.pdf

app_eng.pdf

app_eng.pdf

上传者: 国内外语言学习 2012-07-26 评分 0 0 0 0 0 0 暂无简介 简介 举报

简介:本文档为《app_engpdf》,可适用于IT/计算机领域,主题内容包含iithAnniversaryReaderBonusAskTheApplicationsEngineerANALOGDEVICESTECHNICAL符等。

iithAnniversaryReaderBonusAskTheApplicationsEngineerANALOGDEVICESTECHNICALREFERENCEBOOKSPublishedbyAnalogDevicesNonlinearCircuitsHandbookTransducerInterfacingHandbookMixedSignalDesignSeminarNotesAmplifierApplicationsGuideSystemApplicationsGuideLinearDesignSeminarNotesPracticalAnalogDesignTechniquesHighSpeedDesignSeminarNotesADSPFamilyApplicationsHandbookPublishedbyPrenticeHall(availablefromAnalogDevices)AnalogDigitalConversionHandbookDigitalSignalProcessinginVLSIDSPApplicationsUsingtheADSPFamily(Vols)AskTheApplicationsEngineerAnalogDevices,Inc,AllrightsreservedWORLDWIDEHEADQUARTERSOneTechnologyWayPOBox,Norwood,MA,USATel:(),Fax:()PrintedintheUnitedStatesofAmericaInformationfurnishedbyAnalogDevices,Inc,isbelievedtobeaccurateandreliableHowever,noresponsibilityisassumedbyAnalogDevices,Inc,foritsuse,norforanyinfringementsofpatentsorotherrightsofthirdpartieswhichmayresultfromitsuseNolicenseisgrantedbyimplicationorotherwiseunderanypatentorpatentrightsofAnalogDevicesUseinlifesupportapplications:ProductssoldbyAnalogDevicesarenotdesignedforuseinlifesupportequipmentwheremalfunctionofsuchproductscanreasonablybeexpectedtoresultinpersonalinjuryordeathBuyerusesorsellssuchproductsforlifesupportapplicationatbuyer’sownriskandagreestodefend,indemnify,andholdharmlessAnalogDevicesfromanyandalldamages,claims,suits,orexpenseresultingfromsuchuseAllbrandorproductnamesmentionedaretrademarksorregisteredtrademarksoftheirrespectiveholdersSpecificationsaresubjecttochangewithoutnoticeAskTheApplicationsEngineerthAnniversaryReaderBonusiiiTABLEOFCONTENTSPagePrefaceivMultitroubles,Troublefromthestart,AboutlogcompensationresistorsJamesBryantWhenitcomestotrimmingJamesBryantVFconvertersJamesBryantUsedcorrectly,highspeedcomparatorsprovidemanyusefulcircuitfunctionsJohnSylvanOpampissuesJamesBryantOpampissuesNoiseJamesBryantOpampissuesNoise(continued),HowgrainyiscurrentflowJamesBryantSeminarsandSupportChrisHydeSupplydecoupling,nonidealitiesofresistorsJamesBryantPCboardtracksandgroundplanesJamesBryantVoltagereferencesJamesBryantGrounding(again)WaltKesterTimereferences,MoreonTrimmingJamesBryantConfusedaboutamplifierdistortionspecsWaltKesterHighfrequencysignalcontaminationJamesBryant,withHermanGelbachAReaderNotes(reHighfrequencysignalcontamination)LeroyCordillUsingsigmadeltaconverters,partIOliJosefssonUsingsigmadeltaconverters,partII,andquestionsonnoiseinconvertersOliJosefssonMusta“bit”converterbebitmonotonicandsettletoppmDaveRobertsonandSteveRuscakSettlingtimePeterCheckovichInterfacingtoserialconverters,partIEamonNashInterfacingtoserialconverters,partIIEamonNashCapacitanceandcapacitorsSteveGuintaCurrentfeedbackamplifiers,partIErikBarnesCurrentfeedbackamplifiers,partIIErikBarnesIndexvivthAnniversaryReaderBonusAskTheApplicationsEngineerPREFACEANDINTRODUCTIONThisspecialissueofAnalogDialogueisofferedasabonustoourfaithfulreadersandanencouragementtonewreadersWehavereprintedherethepopularseries,entitled“AskTheApplicationsEngineer”,fromitsinceptioninthroughNumber*inWearecelebratingAnalogDialogue’sthsequentialyearinprint,justconcludedDuringthoseyears,AnalogDialoguehasinformed,enlightened,andoccasionallyeducatedmorethanagenerationofengineers,scientists,andtechniciansaboutAnalogDevices’sproducts,practice,andideasforhighperformancesignalprocessingbyanalog,digital,andmixedmeansThisthanniversarySelection,AskTheApplicationsEngineer,joinsthecollectionofarticlesinourthanniversarybonusedition,TheBestofAnalogDialoguetoReadersofthatpagecollection(stillavailableuponrequest)willrecallitsmanyarticlesrelatingtopracticeaswellasarticlesaboutlandmarkproductsOurintentinoriginatingthe“AskTheApplicationsEngineer”serieshasbeentosupplementthatmineofinformationwithasourceofwhathascometobecalledFAQfrequentlyaskedquestions(andanswers)TheyarechronicledbyafewofourApplicationsEngineers,butarebasedontheaccumulatedwisdomandexperienceofourworldwideApplicationsEngineeringstaff,interactingwithtensofthousandsofcustomersworldwideThroughoutthepagesofthisvolumeyouwillfindreferencestoAnalogDevicesproductsandpublications,justastheyappearedintheoriginaleditionIfyouareinterestedinreceivingcopiesofpublicationsavailablefromAnalogDevicesorinformationaboutspecificproducts,youmayconsultourWebsite,http:wwwanalogcom,orphoneourHelpline,ANALOGD()inNorthAmerica,()elsewhere,oroneofourlocalofficesDatasheetsareavailableinNorthAmericabyhourautomatedfaxbyphoningtheAnalogFaxnumber,Requestsforpublicationsmayalsobefaxedto()Naturally,anytimelymaterialisonitswaytoobsolescencethedayitappearsinprint,soit’spossiblethatatthetimeyoureadaboutaparticulardevice,updatedorimprovedversionshavebecomeavailableForupdatedinformation,phoneourHelplineoroneofourlocalofficesThearticlesinthispublicationappearinchronologicalorder,startingwith“AskTheApplicationsEngineer”Butinorderforthebooktobeusefulasareferencesource,wehaveprovidedanIndexatthebackitwillpermitthereadertofindmaterialrelatingtoaparticularinterestwhereveritappearsWehopeyouwillfindthispublicationhelpfulandwillwelcomeyoursuggestionsofquestionstobeansweredinthefutureDanSheingoldanalogcomEditor,AnalogDialogue*Withoneexception:Wehavenotincluded“AskTheApplicationsEngineer”,whichwasdevotedtoadescriptionofa"ComponentSelectionGuidediskettethatwasnovelandusefulinitstimehowever,ithaslongbeenobsoleteinthisdayoftheWebsite,theCDROM,andtheselectiontreeinprintJustabouteverythingelsehasturnedouttoincludeinformationoftimelessvalueAnalogDialogue()AskTheApplicationsEngineerbyJamesBryantMULTITROUBLESQMymultiplexedADCsystemismisbehavingABeforeyougoanyfurther,haveyougroundedallunusedmultiplexerchannelsQNoButhowdidyouknowABecausethefloatingterminalisoneofthecommonestcausesofproblemsinsystemscontainingCMOSmultiplexersUnusedMUXinputsandoutputs(whetherintegratedintoamultiplexedADCorpartofaselfcontainedMUXchip)canpickupsignalsfromstrayfieldsandinjectthemintothedevice’ssubstrate,turningonspurioussubstratedevicesThen,evenwhentheunusedchannelisturnedoff,theperformanceoftheonchannelmaybebadlydegraded(attheunlikelyextreme,theinjectionmayturnonaspuriousfourlayerdeviceanddestroysomechips)WheneveraMUXisused,allitsinputsandoutputsmustbeconnectedtoapotentialbetweenitssupplyrailsThebestwaytodealwithunusedchannelsistogroundthem,buttheymaybeconnectedtoamoreconvenientpotentialwithintherailsbTROUBLEFROMTHESTARTQTosavepower,myADCispowereduponlytomakeameasurementThesystemisveryaccurateincontinuousoperation,butunpredictablewhenpowerisstrobedWhyAWhenanADC’spowerisswitchedononlytoperformaconversion,itmaymisbehaveforthreereasons:slowreferenceturnon,randominitiallogicstates,andsystemlatchupForvariousreasonsthermalstabilization,capacitancecharging,slowstartingofregenerativecurrentmirrorsusingPNPtransistorsinbandgapreferencesitisnotuncommonforsomevoltagereferencestohaverelativelylargeerrorsformanymillisecondsafterpowerupSucherrorsinanADC’sexternalorinternalreferenceduringconversionleadtoinaccurateresultsAtturnon,atypicalADC’slogicwillbeinarandomstateforaconversiontriggeredatthattime,theADCmaynotbeabletoperformcorrectlyWithoneconversiontriggered,thelogicshouldreturntoitscorrectpreconversionstatebutcasesexistwheretwoconversioncyclesarenecessarybeforetheADCiscertaintoperformavalidconversionHence,agoodgeneralruleistoperformtwo“dummy”conversionsafterpowerupbeforerelyingontheresults(ItisalsowelltorecallthatsomeADCsreactbadlytohavingaconversiontriggeredbeforethepreviousconversioniscompletewhenthishappens,oneortwo“dummy”conversionsmaybeneededtoreturnthelogictoaknownstate)IfanADC’sexternallogicisarrangedsothattheendoftheADC“Busy”signalstartsadelaywhichendswiththestartofthenextconversion,itisimportanttorealizethatiftheconverterpowersupintheBusystate,theBusysignalmayremainlatchedupuntilaconversionStartpulsehasbeenreceivedInthiscase,suchasystemcannotselfstartIftheBusysignalisalwayspresentonpoweruptheproblemisalmostcertaintoberecognizedandaddressedduringthedesignofthesystembutiftheBusysignalisonlyoccasionallypresentonpowerupthesystemmaylatchunpredictablyAsarule,controlsignalstoanADCduringstartupshouldnotdependonthelogicalstateofBusybABOUTLOGCOMPENSATIONRESISTORSQDesignsoflogarithmiccircuits*,includingthoseusingtheADYZXmunit:(forexample,FigurefromtheADMultifunctionUnitdatasheet†)callfor“kTqcompensationresistors”WhataretheyandwheredoIgetthemATheVBEdifferenceacrosstwoopposedsiliconjunctions,onecarryingacurrent,I,andtheotheracurrent,IREF,is(kTq)ln(IIREF)Here,kqistheratioofBoltzmann’sconstanttothechargeonanelectron(about,KV),andTistheabsolutetemperatureinkelvinsAlthoughemployingsimilarjunctionsinisothermalpairseliminatestheeffectsoftemperaturesensitivityofreversesaturationcurrent,thekTqtermisstilltemperaturedependentToeliminatethisdependencyintheapplication,thelogarithmicvoltagemustbeusedinacircuitwhosegainisinverselyproportionaltotheabsolutetemperatureofthejunctionsOverareasonablerangeoftemperaturesnearC,thismaybearrangedbytheuseofagainsettingkΩresistorhavingapositivetemperaturecoefficientofapproximately,ppmCandkeepingitatthesametemperatureasthejunctionsA,ppmCresistorisavailablefromTelLaboratories,GHarveyRoad,Londonderry,NewHampshire(),Telex:(),designatedQ,andfromthePrecisionResistorCoInc,,thSt,Largo,Florida()Telex:,asthePTAnalogDevicesofficesinmostEuropeancountriesareawareoflocalsuppliersoftheseresistorsb*MuchusefulinformationaboutlogarithmicandotheranalogfunctioncircuitscanbefoundintheNonlinearCircuitsHandbook,publishedbyAnalogDevices($),POBox,NorwoodMA†SeeTheBestofAnalogDialogue–,ppAskTheApplicationsEngineerbyJamesBryantWHENITCOMESTOTRIMMINGQIneedsomeadviceabouttrimmingoffsetsandgainsADon’t!unlessyoumustGoodalternativesinclude(a)usingheadachefreedevices,components,andcircuitsthatmeetthespecswithouttrimming(b)takingadvantageofdigitaltechnologyinsystemapplicationstomaketrimcorrectionsinsoftwareSavingsprovidedonoccasionbytrimpotentiometers,inconjunctionwithlooselyspec’ddevices,canturnouttobeillusorywhenyouconsidertheeffectsofcircuitdesign,temperature,vibration,andlifeonperformanceandstabilityaswellasadditionalpaperworkandcomplexitytrimmingentailsQNevertheless,howdoItrimtheoffsetandgainerrorsinanalogcircuitryAInthecorrectorderandwiththecorrectinputsIfyouconsiderthetransfercharacteristicofthecircuitbeingtrimmedthemethodtouseisgenerallystraightforwardThesimplifiedidealtransfercharacteristicofalinearanalogcircuit(suchasanamplifier,ADCorDAC)isgivenbytheequation:OP=KIP()whereOPisoutput,IPisinput,andKisascalefactor(Notethatthissimplificationhidesanenormousnumberofissues:quantizationerrorinanADCdimensionalityofKiftheinputandoutputareindifferentformsegvoltageincurrentoutintentionaloffsetsandmanyothers)Inareal(nonideal)circuit,offsetandgainerrors,OS(referredtotheinput)andK,respectively,alsoappearintheequation,whichbecomes:OP=(KK)(IPOS)()OP=(KIP)(KOS)(KIP)(KOS)()Equations()and()areincompleteinthattheyassumeonlyoneoffsetattheinputbutthisisthemostcommoncaseSystemswithseparateinputandoutputoffsetswillbeconsideredlaterFrom()weseethatitnotpossibletotrimgaindirectlywhenanunknownoffsetispresentOffsetmustbetrimmedfirstWithIPsetat,theoffsettrimisadjusteduntilOPisalsoGainmaythenbetrimmed:withaninputneartofullscale(FS),thegaintrimisadjustedtomaketheoutputobeyequation()QButwhataboutbipolarADCsandDACsAManyADCsandDACsmaybeswitchedbetweenunipolarandbipolaroperationsuchdevices,whereverpossible,shouldhavetheiroffsetandgaintrimmedintheunipolarmodeWhereitisnotpossible,orwheretheconverteristooperateonlyinthebipolarmode,otherconsiderationsapplyAbipolarconvertermaybeconsideredasaunipolarconverterwithalargeoffset(tobeprecise,anoffsetofMSBonehalfoffullscalerange)Dependingonthearchitectureused,thisbipolaroffset(BOS)mayormaynotbeaffectedbythegaintrimIfitissoaffected,equation()becomes:OP=K(IP–BOS)()Inthiscaseoffsetistrimmedatanalogzero,afterwhichgainistrimmednearFSpositiveornegative,butusuallypositiveThisisnormallythemethodusedforDACswherethebipolaroffsetiswithintheDACIfthebipolaroffsetisnotaffectedbythegaintrim:OP=KIP–BOS()HereoffsetistrimmedatFSnegativeandgainistrimmedat(orveryneartoseebelow)FSpositiveThismethodisusedformostADCsandforDACswherebipolaroffsetisobtainedbytheuseofopampsandresistorsexternaltotheDACNaturally,themethodsuggestedonthedatasheetshouldalwaysbefollowed,butwhereadatasheetisunobtainable,ingeneral,offsetshouldbetrimmedatanalogzeroforDACsandFSnegativeforADCsandnearFSpositiveforbothQWhydoyoukeepsaying“near”tofullscaleAAmplifiersandDACsmaybetrimmedatzeroandfullscaleInthecaseofaDAC,all’sthelargestdigitalinputpossibleshouldproduceanoutputLSBbelow“fullscale,”where“fullscale”isconsideredassomeconstanttimesthereferencethisfollowssincetheoutputofaDACisthenormalizedproductofthereferenceandthedigitalinputADCsarenottrimmedatzeroandFSTheoutputofanidealADCisquantized,andthefirstoutputtransition(fromto)takesplaceLSBabovethenominalvalueofall’sThereaftertransitionstakeplaceeveryLSBincreaseinanaloginputuntilthefinaltransitiontakesplaceLSBbelowFSAnonidealADCistrimmedbysettingitsinputtothenominalvalueofadesiredtransitionandthenadjustinguntiltheADCoutputflickersbetweenthetwovaluesequallyTheoffsetofanADCisthereforetrimmedwithaninputcorrespondingtothefirsttransition(ie,LSBabovezerooraboveFSnegativewhichis“near”zeroor“near”FSnegative)andthegainisthentrimmedatthelasttransition(ieLSBbelowFSpositivewhichis“near”FSpositive)ThisprocedureresultsinaninteractionbetweenthegainandoffseterrorsduringoffsettrimbutitshouldbetooslighttobesignificantQArethereanyotheranomaliesresultinginaneedtotrim“near”,ratherthanatfullscaleASynchronousvoltagetofrequencyconverters(SVFCs)areliabletoinjectionlockingphenomenawhentheiroutputfrequencyisharmonicallyrelatedtotheirclockfrequency,ie,whentheiroutputisverycloseto,orofclockfrequencyFSforanSVFCisclockfrequencyThepresenceofatrimtoolcanexacerbatetheproblemItisthereforeadvisabletotrimthegainofanSVFCataroundofFSAnalogDialogue()QWhataboutcircuitsrequiringboth“input”and“output”offsettrimACircuitssuchasinstrumentationandisolationamplifiersoftenhavetwostagesofdcgain,andthegainoftheinputstagescanbevariableThusatwostageamplifier,withaninpu

用户评论(0)

0/200

精彩专题

上传我的资料

每篇奖励 +2积分

资料评价:

/12
仅支持在线阅读

意见
反馈

立即扫码关注

爱问共享资料微信公众号

返回
顶部