Capacitive Proximity Sensing Using the FDC1004 …

1 ApplicationReportSNOA928A March2015 RevisedApril2015 CapacitiveProximitySensingUsingthe FDC1004 DavidWangABSTRACTC apacitiveproximitysensingcan be implementedin a widevarietyof applicationswith the use of TI'sFDC1004and with the flexibilityof the sensordesignin onlyare thereadvantagesof usingcapacitiveproximitysensingcomparedt o alternativedetectionmethods,therearealso guidelinesto followto ensuremaximumperformanceand stabilitywiththe detailthe basicsof proximitysensingand sensortopologyconsiderationsthat of Figures1 FringingElectricFieldsof the IsolatedSensorand the ElectricFlux Densityfor the DistanceComparisonfor Groundedand Distancefor of Tables1 Comparisonof ,HumanHand(GroundedTarget)..64 ProximitySensingRangeBasedon trademarksare the propertyof March2015 RevisedApril2015 CapacitiveProximitySensingUsingthe FDC1004 SubmitDocumentationFeedbackCopyright 2015,TexasInstrumentsIncorporatedSHIELDG NDSENSORSENSORSHIELDG roundedGroundedIsolated SensorParallel FingersBasicsof ProximitySensingUnlikea parallelplatetopologythat workson the principlesof the parallelplatecapacitor,the topologiesfor proximitysensinguse the fringingelectricfieldsto measurethe capacitance,as shownin Figure1.

2 Themajorityof proximitysensingapplicationsuseseitherth e parallelfingersor the the bottomsidesof the electrodesis commonin mostapplicationsto reducethe noiseand strayparasiticcapacitancesin the surroundingenvironmentfromaffectingthe moreinformationon shieldingin capacitivesensing,referto theCapacitiveSensing:Ins and Outsof ActiveShieldingapplicationnote(SNOA926). Figure1. FringingElectricFieldsof the IsolatedSensorand ParallelFingersTopologyModelingthe fringingeffectand workingthroughthe calculationsrequiresthe use of a showsa FiniteElementMethodsMagnetics(FEMM)simul ationof the electricflux densityof theparallelfingerstopologywith a purpleregionsin Figure2 representthe highestdensityof the densityand intensityof the electricfieldsare highestin the regionclosestto the inneredgesof the sensorand groundelectrodes,and will exhibitthe ,a humanfinger(groundedtarget)

3 In betweenthe electrodeswouldcontributemoretowardsthem easuredcapacitancecomparedto the fingernearthe outeredgesof eitherof the FEMMS imulationof the ElectricFluxDensityfor the ParallelFingersTopology2 CapacitiveProximitySensingUsingthe FDC1004 SNOA928A March2015 RevisedApril2015 SubmitDocumentationFeedbackCopyright 2015, of the mostcommontopologiesfor proximitysensingare the parallelfingersand the theirown advantagesand disadvantagesbasedon the type of targetthat is showsa summaryof the best sensortopologyfor differentsystemobjectivesand Comparisonof SensorTopologiesSystemObjectiveTopologyS electionWhySensinggroundedobjectssuchas the humanIsolatedsensorallowsthe majorityof field linestoIsolatedsensorbodyterminateto the humanbodywithoutGNDelectrodeDielectrican d capacitancechangeis smallcomparedtoSensingungroundedobjectsE itherthresholddetectionNo dedicatedGNDelectrodenearbyto terminatefieldHigherproximitydistancesen sitivityIsolatedsensorlinesNo dedicatedGNDelectrodenearbyto terminatefieldHighersensitivityand dynamicrangeIsolatedsensorlinesLessrisk for saturatedmeasurementsParallelfingersIsol atedgroundplaneThe isolatedsensoris the best topologyfor variousmaterials.

4 Sensingrangedetection,and sensitivitybutthe parallelfingerstopologyis capableof performingbetterin primaryconsiderationin selectingthe sensortopologyis how the electrodesare coupledto a FDC1004 Capacitive -to-DigitalConvertercan accommodatea capacitiveoffsetof up100 pF. If the groundplane/electrodeis coupledtightlyso that the capacitancemeasuredbetweenthesensorelect rodeand groundis largerthanthe 100-pFmaximumoffsetcapabilitiesof the FDC1004 ,thecapacitancemeasurementswill alwaysbe takenin the size of electrodesandPCBstackupof the sensordesignto avoidsaturationdue to a parallelfingersdesigncan be less susceptibleto saturationif an isolatedgroundplaneis presentin the systembecausethe GNDelectrodeand groundplanewill not be countlesstargetmaterialsthat can be sensedusingthe capacitiveapproach,but it is possibletogroupthesetargetsinto two categories:groundedtargetsand ungroundedtargetscategories,low and high dielectricconstantmaterialscan be distinguishedin showsthe dielectricconstantsof materialsthathavelow dielectricconstants(closeto the dielectricconstantof air) are limitedto very smallsensingrangessincethe changein capacitance,dictatedby the parallelplatecapacitorequation,is small.

5 Groundedtargetexamples humanbody,metalplates/cases Ungroundedtargetexamples Low dielectricconstant air, plastic,plexiglass,wood Highdielectricconstant varioustypesof alcohol,waterTable2. DielectricConstantsof CommonMaterialsMaterialDielectricConstan tAir1 Alcohol16 123 SNOA928A March2015 RevisedApril2015 CapacitiveProximitySensingUsingthe FDC1004 SubmitDocumentationFeedbackCopyright 2015, DielectricConstantsof CommonMaterials(continued)MaterialDielec tricConstantWood2 6 Waterat 20 50 TopologyAnalysisBothtopologieshavethe capabilityto maindifferencebetweenthetopologieswith detectinggroundedobjectsis the sensitivityof the the locationofthe nearestcommongroundpotentialsource,the isolatedsensortopologyis typicallymoresensitivethanthe parallelfingerstopologyespeciallyat longerdistancesawayfromthe ungroundedtargets,the two topologieshavesimilarperformance,but measurementscan saturatefor the isolatedsensorif the couplingto a groundplaneor groundsourceis largerthanthe 100-pFoffsetrangeof showsplotsof capacitanceversusrangefor groundedand groundedtargets,bothtopologiesshowgoodse nsitivity,with the isolatedsensorshowingslightlybetterperfo rmanceversusthe ungroundedtargets,bothtopologiessufferad ecreasein sensitivity.

6 Showinglowerdynamicrangeand interceptionof the noisefloorat FEMM simulationdatashownin Figure3 doesnot factorin othergroundsourcesin simulationdataonly takesinto considerationthe targetobjectand the electrodes:sensor,ground,and FEMM simulationsuseda sensorsize of 4 cm 1 cm for the isolatedsensorwith a shieldlayerof thesamesize 1 mm parallelfingerstopologywas pairedwith a GNDelectrodethe samesize with a 5-mmgap spacingbetweenthe sensorand shieldlayerspannedfromtheouteredgesof the two electrodes1 mm Capacitancevs DistanceComparisonfor Groundedand UngroundedTargetsOne advantagethat the parallelfingerstopologyhas overthe isolatedsensorwith ungroundedtargetsissensitivitybasedon locationof the the groundsourceis muchsmallerthanthe electrodesand at a distancemuchgreaterthanthe electrodes,the isolatedsensorwill be lesssensitiveat shortersensingrangesthanthe parallelfingerssincethe parallelfingerstopologyhas areferencegroundelectrodepairedwith the FDC1004 SNOA928A March2015 RevisedApril2015 SubmitDocumentationFeedbackCopyright 2015.

7 SensitivityProximitysensingrangeand sensitivityis affectedby a varietyof factors:Sensorstackup,surfaceareaofthe electrodes,the nearestcommongroundpotentialsourceand externalinterference/noise,mostofwhichar e dependenton the and Figure5 displaythe capacitancemeasurementsovertime with a humanhandtarget18 cm awayfromthe paralleland filteringis performedon the takenfrom19 cm awaywith thehumanhandtargetto compareagainstmeasurementswith the hand18 cm datawas collectedusinga sensorsize of 2 cm x 1 cm on a standardtwo-sidedcopperPCBwith the sensorelectrodeon topand the parallelfingerstopology,a groundelectrodewith the samematerialtype and dimensionwas cm awayfromthe ParallelFingersTopologySensingDataFigure 5. IsolatedSensorTopologySensingData5 SNOA928A March2015 RevisedApril2015 CapacitiveProximitySensingUsingthe FDC1004 SubmitDocumentationFeedbackCopyright 2015, SensingRangePerformanceComparison,HumanH and(GroundedTarget)ParallelFingersIsolat edSensor18 cm19 cm18 cm19 cmAverageat 18/19cm (pF) (pF) Cap (fF) comparesthe sensingrangeperformancebetweenthe parallelfingersand humanhandsoonerwith slightlymorenoisemarginthanthe parallelfingerscase,the changein capacitancebetweenthe baselineaverageand the averagedmeasurementswith the handat 19 cm is too smallto distinguishthe differencebetweenthe noiseand smallmovingaveragecan be appliedto the datain real time to help filter out the peak-to-peaknoisealongthe signaland increasethe confidence/reliabilityin detectingthe measuredfor varioussensorareasizesusinga humanhand(grounded)

8 Asthe sensingrangeis basedon a detectionthresholdshift of 3 fF fromthe baselinemeasurement(no targetpresent).A squaretwo-layer,double-sidedcopperPCBwit h standardthickness(62 mils,1-ozcopper)was usedfor the sensorand shieldelectrodewas the samesize as the sensorelectrodeand directlyunderneaththe showsthe sensorstackupusedto measurethe SensorStackupfor SensingRangeTable4 showssensingrangeversustargetsize detailedgraphof capacitanceversusdistancefor the varioussensorareasizesis shownin ProximitySensingRangeBasedon SensorAreaSizeSensorAreaSizeProximitySen singRange(cm2)(cm) size of the shieldand distanceto the sensorelectrodesignificantlyaffectsthe shieldthat is largerthanand closerto the sensorelectrodereducesthe sensitivityand maximumrange,but itlimitsthe amountof interferenceseenby the smallershieldfurtherawayfromthesensorele ctrodehas an oppositeeffecton sensitivityand moreinformationonshieldingeffects,refert o theCapacitiveSensing.

9 Ins and Outsof ActiveShieldingapplicationnote(SNOA926). 6 CapacitiveProximitySensingUsingthe FDC1004 SNOA928A March2015 RevisedApril2015 SubmitDocumentationFeedbackCopyright 2015, determiningthe sensingrangeis the peak-to-peakand RMSnoiseseenon the sensorwill increaseas the sensorareasize increasessinceit acts asa widebandantenna,pickingup any interferencepresentin the (normaldistributionwith meanvalue0), the thresholdlevelshouldbe > 3 to achieve< ,where is the standarddeviationofthe noisefor the dataaboveis fF (at 100 SPS),so a detectionthresholdof > fF is 3- to4-fF thresholdis a validconditionwith plentyof mentionedpreviously,a movingaverageonthe real-timedatacan be usedto help filter out the noisefor a higherSNRand cleanerdetectiontransitionof the targetfor be categorizedbasedon use casesand targetendequipmentin shortrange(< 15 cm) and long range(up to 50 cm)

10 Featuresof the FDC1004allowminimumsensorsize for a givensensingdistanceand maximumsensingdistancefor a (< 15 cm)(up to 50 cm)Usesdisplaywakeupdisplaywakeupcollisi onwarningUsercontrolsdooractivationdoora ctivationemergencybrake/stoppresencedete ctionon/offactivationforeignobjectdetect ionAutomotiveAutomotiveAutomotiveAutomot ivecar doorsensordoorkick sensorAutomaticdoorsand gatesdoorsInfotainmentdisplayIndustrialC onsumerIndustrialConsumerthermostatlapto pselevatorsAudioequipmentproximitysensor scomputerscreensgaragedoorsMP3playersdis playsautomaticdoorsTargetEndHMIrobotsEqu ipmentWhiteGoodsIndustrialIndustrialrefr igeratorHMIthermostatcoffeemachineplumbi ngfixturesWhitegoodssoapdispenserdishwas herstovefan7 SNOA928A March2015 RevisedApril2015 CapacitiveProximitySensingUsingthe FDC1004 SubmitDocumentationFeedbackCopyright 2015,TexasInstrumentsIncorporatedAppendi xASNOA928A March2015 RevisedApril2015 SensorAreaSizeDataFigure7 showsa detailedgraphof capacitanceversusdistancefor the Capacitancevs Distancefor VariousSensorAreaSizes8 CapacitiveProximitySensingUsingthe FDC1004 SNOA928A March2015 RevisedApril2015 SubmitDocumentationFeedbackCopyright 2015, (March2015)to A.