PicoQuant MultiHarp 150 - T3 Mode

PicoQuantGmbH MultiHarpSoftwareV.3.1.0.0
transferratemustbesharedbytheinputsused.Forallpracticallyrelevantphotondetectionapplicationstheef-
fectiverateperchannelismorethansufficient.Ifthereareveryintenseburstsattheinputofthefront-endFIFO
itmayhappenthateventsarelost.Thisisindicatedtothesoftwarebymeansofahardwareflagsothatthe
usercanbeinformedofsuchlosses.Theusermustthendecideifthelossescanbetoleratedforthegivenex-
periment(seealsosection8.1onwarnings).
Formaximumthroughput,T2modedatastreamsarenormallywrittendirectlytodisk,withoutpreviewotherthan
countrateandprogressdisplay.However,itisalsopossibletoanalyzeincomingdata”onthefly”.TheMulti-
Harpsoftwareprovidesareal-timecorrelatorforpreviewduringaT2modemeasurement(seesection5.3.7).
Othertypesofreal-timeprocessingmustbeimplementedbycustomsoftware.TheMultiHarpsoftwareinstalla-
tionprovidesdemoprogramstoshowhowT2modefilescanbereadbycustomsoftware(seethefolder
filedemo underthechosensoftwareinstallationfolder).Theimplementationofcustommeasurementpro-
gramsrequirestheMultiHarpprogramminglibrary,whichisprovidedasaseparatesoftwarepackageonthe
distributionmediaorasdownload.AlternativesforadvancedT2datacollectionandanalysisaretheSymPho-
TimeandQuCoasoftwaresuitesofferedbyPicoQuant.SymPhoTimeisfocusedontypicallifescienceapplica-
tionswhileQuCoaisorientedtowardstypicalquantumopticsapplications.
5.3.3. T3 Mode
InT3modethesyncinputisdedicatedtoaperiodicsyncsignal,typicallyfromalaser.Asfarastheexperimen-
talsetupisconcerned,thisissimilartoclassicTCSPChistogramming.Themainobjectiveistoallowforhigh
syncrateswhichcouldnotbehandledinT2mode.AccommodatingthehighsyncratesinT3modeisachieved
asfollows:First,thesyncdividerisemployedasinhistogrammingmode.Thisreducesthesyncratesothatthe
channeldeadtimeisnolongeraproblem.Theremainingproblemisnowthatevenwiththedivider,thesync
eventratemaystillbetoohighforcollectingallindividualsynceventslikeinordinaryT2mode.Consideringthat
synceventsarenotofprimaryinterest,thesolutionistorecordthemonlyiftheyarriveinthecontextofapho-
toneventonanyoftheinputchannels.Theeventrecordisthencomposedoftwotimingfigures:1)thestart–
stoptimingdifferencebetweenthephotoneventandthelastsyncevent,and2)thearrivaltimeoftheeventpair
ontheoverallexperimenttimescale(thetimetag).Thelatterisobtainedbysimplycountingsyncpulses.From
theT3modeeventrecordsitisthereforepossibletopreciselydeterminewhichsyncperiodaphotoneventbe-
longsto.Sincethesyncperiodisalsoknownprecisely,thisfurthermoreallowsreconstructingthearrivaltimeof
thephotonwithrespecttotheoverallexperimenttime.
EachT3modeeventrecordconsistsof32bits.Thereare6bitsforthechannelnumber,15bitsforthestart–
stoptimeand10bitsforthesynccounter.Ifthecounteroverflows,aspecialoverflowrecordisinsertedinthe
datastream,sothatuponprocessingofthedatastreamatheoreticallyinfinitetimespancanberecovered.The
15bitsforthestart–stoptimedifferencecoveratimespanof32,768×RwhereRisthechosenresolution.At
thebestpossibleresolutionof5psthisresultsinaspanofabout164ns.IncaseoftheMultiHarp150Nmodels
(80psresolution)thecoveredspanis2621ns.Ifthetimedifferencebetweenaphotonandthelastsyncevent
islarger,thephotoneventcannotberecorded.Thisisthesameasinhistogrammingmode,wherethenumber
ofbinsislargerbutalsofinite.However,bychoosingasuitablesyncrateandacompatibleresolutionR,it
shouldbepossibletoreasonablyaccommodateallrelevantexperimentscenarios.Rcanbechoseninawide
range,startingwiththedevice'sbaseresolutionandthencontinuingbyrepeateddoublingofthetimebinwidth.
DeadtimeinT3modeisthesameasintheothermodes.Withineachphotonchannel,autocorrelationscanbe
calculatedmeaningfullyonlystartingfromlagtimeslargerthanthedeadtime.Acrosschannelsdeadtimedoes
notaffectthecorrelationsothatmeaningfulresultscanbeobtainedatthechosenresolution,allthewaydown
tozerolagtime.Thisrequirescustomsoftware.
The32biteventrecordsarequeuedandforwardedtothehostPCinthesamestaggeredFIFOarchitectureas
describedinthesectiononT2modeabove.Accordingly,asustainedaveragecountrateofupto90Mcpsis
possibleinT3modetoo,whilenowthesynceventsdonotconsumeanytransferbandwidth.
Formaximumthroughput,T3modedatastreamsarenormallywrittendirectlytodisk.However,itisalsopossi-
bletoanalyzeincomingdata”onthefly”.Onesuchanalysismethodistheon–linecorrelationimplementedin
theMultiHarpsoftware(seesection5.3.7).Otherspecializedanalysismethodsmustbeimplementedviacus-
tomsoftware.TheMultiHarpsoftwareinstallationprovidesdemoprogramsshowinghowT3modefilescanbe
read(seethefolderfiledemounderthechosensoftwareinstallationfolder).Theimplementationofcustom
measurementprogramsrequirestheMultiHarpprogramminglibrary,whichisprovidedasaseparatesoftware
packageonthedistributionmediaorviadownload.AnalternativeforadvancedT3modedatacollectionand
analysisistheSymPhoTimesoftwaresuiteofferedbyPicoQuant.
Page37