How to Measure the Loop Transfer Function of …

1 ApplicationReportSNVA364A October2008 bodeplotwithonlyanaudiogenerator(orsimpl esignalgenerator) explainedinaneasytofollowstep-by-stepman nersothata powersupplydesignercanstartperformingthe semeasurementsin a :Analyzinga Caseswitha TypicalBodePlotofa October2008 RevisedApril2013AN-1889 HowtoMeasuretheLoopTransferFunctionofPow erSuppliesSubmitDocumentationFeedbackCop yright 2008 2013, a Bodeplotis ideallymeasuredwitha networkanalyzerbutcanalsobeobtainedwithv erysimpleequipmentavailablein :SettinguptheCircuitThefirststepis tobreaktheloopofthepowersupplyin ordertohavea pointtoinjecta ,theloopisbrokenatthelowimpedanceoutputn odeabovethehighsidefeedbackresistorin goodinjectionpoint, showsa ,A andB,a smallresistanceshouldbeputin goodvalueis a 20 resistortheadjustedoutputvoltagewillbene gligiblyinfluencedbutthenodestoinjecta parallelwiththeR1feedbackresistorshoulds tayinparallelwithR1andnotbemodifiedbythe introductionofthe20 soundmechanicalmeasurementstructureis shownin Figure2.

2 Thefeedbackpathis brokenonthePCBanda 20 resistoris is relativelyeasytocutthefeedbacktraceonthe PCBandreconnectit througha 20 twistedcableasshownin Figure2 is smalltwistedpairworksverywellforminimizi ngnoisepickupwhichis criticalfora October2008 RevisedApril2013 SubmitDocumentationFeedbackCopyright 2008 2013, SwitchModePowerSupplyCircuit3 SNVA364A October2008 RevisedApril2013AN-1889 HowtoMeasuretheLoopTransferFunctionofPow erSuppliesSubmitDocumentationFeedbackCop yright 2008 2013,TexasInstrumentsIncorporatedStep2 :TheInjectionTransformerTheinjectedsigna lhastobea smallsignalin ,yethastobelargeenough,sothatitcanbereco gnizedonceit , typicalinjectedsinewavesignalshouldbebet ween30mVand100mVacrossthemeasurementresi storasdescribedin neededin stillconsidereda smallsignalin ,whichoffera veryflatvoltagetransmissionovera networkanalyzer, changedmanually, transformercanbeusedwhichdoesnotshowa flatvoltagetransmissioncurvesuchasa verylowcoststandardtransformerintendedfo ra showssucha , Figure6,andFigure7, thetransformeris.

3 PreparingtheSignalGeneratorThesignalgene ratorshouldprovidea is , welltriggeredpictureofthesinewavesonecan triggeroffathirdchannelwhichis , ,thiswaythetriggerwillnothavetogetreadju stedwithchangingamplitudesofchannelA :PreparingthePowerSupplyThepowersupplyha stobeputin theconditionin is alwaysa goodideatoperforma voltagemodeconverterswithoutinputvoltage feedforward, ,makesuretoattachthesignalinjectiontrans formeracrossthefrequencyinjectionresisto rofthepowersupplyandtoattachthetwovoltag eprobesofchannelA October2008 RevisedApril2013 SubmitDocumentationFeedbackCopyright 2008 2013, :TakingtheMeasurement7 Step6:TakingtheMeasurementAftersettingup theequipmentandpoweringeverythingup,thes copewillprobablyshowonelineononechannela nda thereis nosinewavevisible,eitherthescopeis notsettothehighestamplituderesolution(ty pically20mVperdivision) sinewaveis visible,changingthefrequencyofthesignalg eneratorwillleadtoachangein amplitudeonchannelA certainfrequencywhenthesinewavesofchanne lA andchannelB thepointwherethegainoftheloopis showschannelA andchannelB , , andchannelB givesthegainatthefrequencythesignalgener atoris helpswiththevoltagedifferenceandthecorre spondingvaluein dBaccordingtotheformula.

4 DB= 20 October2008 RevisedApril2013AN-1889 HowtoMeasuretheLoopTransferFunctionofPow erSuppliesSubmitDocumentationFeedbackCop yright 2008 2013,TexasInstrumentsIncorporatedStep7:A nalyzinga fullbodeplot,thesignalgeneratorfrequency is ,it ,it is quitedifficulttoseea voltagerelationshipbetweenchannelA andB ofa ,it mightbedifficulttoseetheexactdBvalue,but a quantitativeobservationsuchas thegainis veryhighandprobablyabove30dB designwhichdoesnotoscillateoris notin stabledesignformeasurementstobetaken, largecapacitancefromthecompensationpinto groundin designswheretheerroramplifieris a transconductanceamplifierandfromthecompe nsationpintotheFBpinin designsweretheerroramplifieris a capacitorvalueof1 F willseta poleatverylowfrequenciesandforcethegaint odropquicklysothatthe0dBcrossoveris ata ,thephasemarginatverylowfrequenciesis usuallyenoughtoyielda onlygoodforgettinga.

5 Analyzinga BodePlotThemostimportantmeasurementis providesinformationaboutthebandwidthofa powersupplyaswellasthephasemarginofa combinationofthelevelofDCgainaswellasa versuschannelB atthisfrequency,onemeasuresthephasemargi nofthesystemwhichis designis,a minimumphasemarginof45to50degreesis showsa typicalbodeplottakenwitha is takenfroma TypicalBodePlotofa Step-DownVoltageModeSwitchingRegulator6A N-1889 HowtoMeasuretheLoopTransferFunctionofPow erSuppliesSNVA364A October2008 RevisedApril2013 SubmitDocumentationFeedbackCopyright 2008 2013, , thechannelwhichmeasurestheinjectedsignal goingintotheloopcanalsobeattachedtotheco mpensationpinofa powersupplydesignsuchasshowninFigure6. Inthissetupthetransferfunctionoftheplant is is easytoselectoptimizedcompensationcompone ntsfora desiredbandwidthandphasemarginofa Converter10 MeasuringinCaseswitha SpecialOutputStageAsmentionedin steponeofthisdocumentit is necessarytohaveaninjectionnodein theloopwhichhasa lowimpedancenodeononesideanda ,thebestpointtoinjectthissignalis pointis a regulatinga currentthrougha measuredwitha highsidecurrentsenseblockwhichis sensingtheLEDcurrentandis translatingit intoa placein thefeedbackpathwhichhasonesidelow-impeda nceandtheothersidehigh-impedance,theoper ationalamplifierOP1is addedintothefeedbackloopasa thisamplifieris selectedtobefasterthanhalftheswitchingfr equencyoftheconverter.

6 Realnetworkanalyzerwhichsweepsthefrequen cyofaninjectedsignalautomaticallyandcomp utesthephasedifferenceandthegainbetweenc hannelA andchannelB ontheflyis ,measuringthecontrolloopofa powersupplycanbedonewithbasiclabequipmen tsuchasanoscilloscope,a simplesignalgeneratoranda powersupply,veryusefuldatacanbetakenwhen performingthesemeasurementssuchasthephas emarginofa ,a October2008 RevisedApril2013AN-1889 HowtoMeasuretheLoopTransferFunctionofPow erSuppliesSubmitDocumentationFeedbackCop yright 2008 2013, October2008 RevisedApril2013 SubmitDocumentationFeedbackCopyright 2008 2013,TexasInstrumentsIncorporatedIMPORTA NTNOTICET exasInstrumentsIncorporatedanditssubsidi aries(TI)reservetherighttomakecorrection s,enhancements,improvementsandotherchang estoitssemiconductorproductsandservicesp erJESD46,latestissue,andtodiscontinueany productorserviceperJESD48, (alsoreferredtohereinas components )

7 AresoldsubjecttoTI s ,in accordancewiththewarrantyin TI s ,testingofallparametersofeachcomponentis productsandapplications, ,eitherexpressorimplied,is grantedunderanypatentright,copyright,mas kworkright,orotherintellectualpropertyri ghtrelatingtoanycombination,machine,orpr ocessin licensetousesuchproductsorservicesora licensefroma thirdpartyunderthepatentsorotherintellec tualpropertyofthethirdparty,ora TIdatabooksordatasheetsis permissibleonlyif reproductionis withoutalterationandis accompaniedbyallassociatedwarranties,con ditions,limitations, is solelyresponsibleforcompliancewithallleg al,regulatoryandsafety-relatedrequiremen tsconcerningitsproducts,andanyuseofTIcom ponentsin itsapplications, hasallthenecessaryexpertisetocreateandim plementsafeguardswhichanticipatedangerou sconsequencesoffailures,monitorfailuresa ndtheirconsequences, , ,TI s goalis , FDAC lassIII(orsimilarlife-criticalmedicalequ ipment)unlessauthorizedofficersofthepart ieshaveexecuteda enhancedplastic solelyattheBuyer's risk,andthatBuyeris solelyresponsibleforcompliancewithallleg alandregulatoryrequirementsin , , , :TexasInstruments,PostOfficeBox655303,Da llas,Texas75265 Copyright 2013,TexasInstrumentsIncorporat