外文翻譯--數(shù)控機床性能測試設(shè)備的開發(fā)和利用 英文版.pdf

THEDEVELOPMENTANDAPPLICATIONOFPERFORMANCETESTINGDEVICESFORCNCMACHINETOOLSWenyuhJyweProfessor,DepartmentofAutomationEngineering,NationalHuweiInstituteofTechnology,Huwei,Yunlin,Taiwan,632Address:64WenhuaRoad,NationalHuweiInstituteofTechnologyHuwei,Yunlin,Taiwan,R.O.CEmail:J.twAbstract:ThetestingdevicesforperformancetestingonaCNCmachinetoolweredeveloped.Asmall-radiuscontouringdevicewasemployedforcheckingthedynamicperformanceofaCNCmachinetool.Thisdevicewascomposedofalaserdiodeandaquadrantsensor.ForcheckinggeometricerrorofaCNCmachinetool,along-radiuscontouringdevicewasemployed.Theopticencodercontouringsystemswerealsorepresented.Experimentalworkwascarriedoutinthepaper.Copyright2002IFACKeywords:Contouring；Geometricerror；CNCmachinetool；Quadrantsensor；Opticscalemeasuringsystem；Planarencoder1.INTRODUCTIONJ.B.Bryan(1982)developedthefirstballbarsystemforthecontouringtest.However,inthissystem,theuncertaintyishighduetothefrictionbetweenthemasterballsandthemagneticsocketsandnoaccuratecontouringradiuswasgiven.Knappssystem(1983,1983)usedthecirclecomparisonstandarddiscmountedonthetestedtableofthemachinetoolandthe2D-probe.Y.Kakinov(1985),Y.Kakino,YIhara,A.Kamei(1986)andY.Kakinov,Y.IharaandY.Nakastsu(1987).Y.Kakino,Y.IharaandY.Nakastu(1986)providedsomemethodsbyusingtheballbarsystemtocalibrateacoordinatemeasuringmachineandCNCmachinetools.W.Knapp(1986,1988)describedtherulestoreducetheerrorsduetothestick-slipandothers.M.BurdekinandJ.Park(1988)modifiedtheoriginalballbarsystembyemployingafour-rodlinkage.M.BurdekinandW.Y.Jywe(1992)providedamethodtodiagnosethecontouringerrorandtoadjusttheparametersoftheCNCcontrollertooptimizetheperformanceofthetestedCNCmachinetool.J.C.ZiegertandChritopher(1994)describedthelaserballbarsystembyguidingafiberwithalaserinterferometer.WenyuhJywe(2001)providedtheapplicationofthedevelopedlinkagetoverifythecontouringerrorofaCNClatheandalsoprovidedthemethodtocompensatefortheerror.Theabovecontouringsystemsprovidedthegeneralpurposeofacontouringtest.Withthehighcostofthemeasuringdevice,theystillcannotprovidepropercontouringsystemsforvariouspurposes.Thus,inthispaper,somesimpleandlowcostcontouringsystemsaredevelopedindividuallyfordifferenttestpurposes.Theyaresummarizedandrepresentedasfollowed:1.Thelaserdiodeandthequadrantsensorcontouringsystemforsmallradiuscontouring(lessthan4mminthecontouringradius).2.Theopticscalecontouringsystemforlarge-radiuscontouringandforthevolumetricerrortest(Upto300mminthecontouringradius).3.Theopticscalecontouringsystemforthevolumetricerrortest.4.Theself-supportedballbarsystemforcontouringtestundercuttingsituation.2.THELASERDIODEANDTHEQUADRANTSENSORCONTOURINGSYSTEM(QSCS)Fig.1describesthelaserdiodeandquadrantsensormeasuringsystem,composedofalaserdiode,anopticlenses,ashield,aquadrantsensor,asignalpropose,anA/DinterfacecardandaPC.Thequadrantsensoroutput4channelsofvoltageswhilethelaserlightisprojectedonit.Duetothelimitationofthesizeofthequadrantdetector,onlyupto4mmcontouringradiusisavailable.However,thereisgoodenoughforcheckingthedynamicperformanceoftheCNCCopyright2002IFAC15thTriennialWorldCongress,Barcelona,Spainmachinetool.Fig1TheprinciplethequadrantsensorforperformancetestofamachinetoolAVBVCVDVXYDiodeLaserFig2Theworkingprincipleofalaser-diodeandaquadrantsensorAsshowninFig.2,Thelaserdiodeprovidesalaserlightandthelightisfocusedandprojectedonthequadrant.Thequadrantoutputsfourchannelsignalsintermofvoltages.Fromtheprojectedpointonthequadrantsensor,thepositioninX-axisandinY-axiscanbeobtained.LetVA,VB,VCandVDrepresenttheoutputvoltagesofthesensor.Fig.3showsthecalibrationdiagramtoobtainthepositionsinX-axisandinY-axis.XXxBABAbDaVVVV+=+XDXbBABAVVVV+(a)YYYDCDCbDaVVVV+=+YDYbDCDCVVVV+(b)Fig3:ThecalibratingdiagramofanaxisofaquadrantsensorAfterthecalibrating,theconstantswillbesavedinbePCandbeemployedforthefurtherapplications.Fig.4ThetestresultonaCNCverticalmillingmachinewithaFANUC-SeriesOMcontroller3.1TheexperimentaltestsbyusingQSCSAsshowninFig.4,thequadrantsensorworkingrangeisabout6mm6mmafterthecalibration.ACNCverticalmillingmachinewithaFANUCseriesOMcontrolleristested.Thecontouringfeedrateisselectedat10,2100and3200mm/minindividually.A2mmcircularcontouringradiusissetforthreerunsforbothclockwiseandanti-clockwise.Therepeatabilityforeach3runsisverygoodintheactualtest.Thecontouringradiusreduceswhenfeedrateincreases.Atafeedrateof3200mm/min,theradiuserrorisabout-0.7mm.Also,thecontouringerrorresultinFig.4showsa45ellipsepath.AllthesecontouringerrorsinFig.4arecorrespondingtothoseshowninthemanual(1990)providedbyFanucCo.3.THEOPTICSCALEMEASURINGSYSTEMInthissectionanopticscalemeasuringsystemisdeveloped.Basically,thesystemisemployedforcontouringtestonaCNCmachinetool.However,thevolumetricerrortestofemployingthissystemisalsostudied.Thenthemethodtoemploythissystemforsquarenesserrorstestisrepresented.3.1TheopticscalemeasuringsystemforthecontouringtestFig.5indicatestheopticscalemeasuringsystemforCNCmachinetoolperformancetest.Anopticscaleisfixedonthesupportingblock.Thereadingheadisguidedbytheguidingbarandmovedbythemovingbar.Onelinearbearingisfixedbetweenthesupportingblockandthemovingbarwhiletheotheroneisfixedbetweentheopticreadingheadandtheguidingbar.Onemasterballisconnectedtoamagneticbar,whichisfixedontheCNCmachinetoolspindlewhiletheothermasterballandmagneticbarelementisfixedonamagneticbase,whichissetonthetestingCNCmachinetooltable.Fig.5ThestructureoftheopticmeasuringsystemDuringatest,therelativedisplacementbetweentwomasterballsissampled.Foracircularcontouringtest,thecontouringcenterissetattheballcenterofthespindlewhilethetableismovedinthetwoaxesofthemachinetoolatthesametime.Thecontouringmotioncausestheopticscaletomovealongtheopticreadinghead.Therefore,thesamplecanbeobtainedandsenttothePCviaacounterinterfacecard.Furthermore,softwareiswrittentosimple,recordandtoanalyzethedata.Fig.6Thestructureofusingtheopticmeasuringsystemforvolumetricerrortest4.THEOPTICSCALEMEASURINGSYSTEMFORTHEVOLUMETRICERRORTESTAccordingtoISO203-2orANSIB5standard,forCNCmachinetoolpositioningtest,asimplepositiontestcanbeexecuted,ingeneral,byanexpensivelaserinterferometersystem.Fordifferentmeasuringareas,differentsetting-upsarenecessary.AspreviousstudybyWenyuhJywe(2000),aspherevolumetricerrorcanbecarriedoutbyaballbarsystem.Inthatsystem,onlythepositionerrorandbacklashonthesurfaceofthespherecanbeobtainedduetothelimitationoftheworkingrangeoftheLVDT(LinearVariableDifferentialTransformer),whichisemployedontheballbarsystem.Insummary,theopticscalemeasuringsystemtakesbothadvantagesofthetwomeasuringsystems.1.Inonesetting-up,ahollowsphericalvolumetricerrorofaCNCmachinetoolcanbecarriedout.2.Thecostofthisopticscalesystemislow.Itiseasytooperation.3.FordifferentsizesofCNCmachinetools,differentrangesofopticscalesarealwaysavailable.Fig.6showsthesamplingpathofthehollowspherevolumeofaCNCmachinetool.Inone-settingup,thecenterofthemasterballoftheopticscalelinkageissetaboutatthecenterofthevolumeofthetestedCNCmachinetool.Thenpossibleevenspacingarrangementistakentoidentifythesamplingpath.Forexample,thesphereiscutinto5planes.Eachplaneisevenlycutinto6lines(360/60).Eachlineincludes8samplingpoints(240mm/30mm=8).Thusinonesetting-up240samplingpoints(5x6x8=240)canbetaken.Totestthebacklashandtherepeatability3x2runsareneeded.Thus240x6runs=1440samplingpointsaretaken.Ittakes2secondsforeachsampling.Thus,ittakesaboutonehour(includingsettinguptime)tocarryoutthiswork.Fig.7ThetestarrangementforsquarenesserrorstestsofYZandZXplaneInthistestarrangementinFig.7,ahorizontalXYplanecontouringiscarriedout.Then,withthesamesetting-upandthesameopticscalemeasuringsystem,letthespindlemoveupinZ-axisatadistance(offset).Thesecondcontouringtestisthenexecuted.Withthisopticscalemeasuringsystem,itisnotnecessarytochangethelinkageorthesetting-upduringthissquarenesstest.Ifnosquarenesserrorexiststhesameanalyzedinitialsetting-uperrors(eccentricity)inbothcontouringtestswillbeobtained.Underthisarrangement,thesecondsampleddatamustbemodifiedbeforeanalysis.Therefore,)cos()()(AngleSetoutputTransducerdataModified=Where,)2()1(cos1dtestnlinklongofLengthteststlinkshortofLengthAngleSet=Afterthesamplingtheoriginalsampledatawithshortlinkandthemodifieddatawithlonglinkareanalyzedbytheleastsquarefitting.Thetwoindividualanalyzedinitialset-uperrorsareusedforthecalculationofthesquarenesserror.SquarenessatZXplane)(tan121OffsetXX=SquarenessatYZplane)(tan121OffsetYY=Where()istheanalyzedcenterofthefirstcontouringtest11,YXWhere()istheanalyzedcenterofthesecondcontouringtest22,YXOffsetisthemovementdistanceinZ-axisafterthefirstcontouringtest4.1TheplanarencodercontouringdeviceforCNCmachinetoolsTheprincipleoftheplanarencoderAplanarencodersystem,suchasRenishawRGXgridplate,hasbeendevelopedforapplicationssuchassemiconductorandelectronicsmanufacturingequipment.Thesystemusesreadheadwith2orthogonalsensorsthatreadachegure-patternedgridinbothXandYdirectionssimultaneously.Thesystemwithgooddynamicresponsecanprovideupto0.1mresolutioninpositioning.Fig.8showsthearrangementofthecontouringtestbyusingasimpleplanarencoder.Thisplanarencoderprovidespositioninginformationineachaxisfora2Dcontouring.Duringthetest,theplanarencoderissetontheCNCmachinetool.ThereadingheadisfixedwiththespindleoftheCNCmachinetool.Thesoftwareofthecomputerviaacountercardcanreadthesamplingdata.Thissamp