Measurement System Analysis

Measurement System Analysis (MSA) course is essential for successful Six Sigma DMAIC and DFSS projects. It is also key for implementation of SQC, and efficient process management.
Reliable measurement processes are critical to the success of any effort dependent on measurement data and process analysis, including Six Sigma DMAIC improvement projects, DFSS project, SPC, SQC, Supplier Quality, and business process management and continuous improvement. Without validation that measurements are accurate, repeatable with multiple measurements by the same person, reproducible from person to person (gage Repeatability and Reproducibility or gage R&R), all conclusions are suspect, and process management is therefore fragile and ineffective.

Organizations typically focus on measurement accuracy and calibration, but this course also emphasizes the essential elements of reliable measurement procedures.

1. Measurement
System
ANALYSIS
Measurementsystems
Analysisis adecision
makingprocessonthe
measuringequipment's
usabilityininspectionof
thequalitycharacteristic.
Itevaluatesif a
measurementsystemis
suitableforaspecific
application.

2. MSA
Objectivesofthis training Is your measuring equipment correct?
Toexplainvarioussourcesof
measurementsystemuncertainty.
Assessmentoflinearity,stability,
repeatability,andreproducibilityof a
Measurementprocess(GageR andR)
“Inaccuratemeasurementsmayleadtofalse
signalsoncontrolcharts.”
Supportin settingupameasurementand
Calibrationsystem. isatruestatementandInthepresenceof
significanterror inthemeasurementprocess,a
capableprocessmaybeconfusedwithan
incapableprocess.

3. MSA
MSA is also calledGage Rand R
Knowledgeofvariationisoneofthemost
powerfultoolsa companycouldapplyinthe
questforimprovement.
GageR&RRepeatabilityand
reproducibilityStudiesarethemost
widelyusedtechniquesfor evaluatingthe
variationinameasurementsystem and
determiningif themeasurementsystem is
acceptablefor use.
Itisessentialtoknowhowmuchvariationisas
aresultofthemeasurementprocess.
Whenthemajorsourceofvariationisfromthe
measurementprocess,significanttimeand
moneycanbewastedinfixingandcontrolling
theprocess.
Once ameasurementsystem is found
acceptable,itis equallyimportantto
instituteaformal system tomanagethe
measurementsystem toensurethatit
continuestobereliableanddependable.
Butfirstitsessentialtomeasure

4. MSA
Measurements
Measurementsaredefinedasthe
assignmentofnumbersto materials
andtheircharacteristicssothatit
couldbecontrolled andtheir
relationshipunderstood.
DoUknowaboutthe"fiveM's" that
constitutemostrootcauses?:
•Measurementis aprocess
•Measurementsaredonetounderstand
Measurements
man,machine,materials,methods,
andmeasurement.

5. ors stematicE
MSA
Measurement Error
SystematicErrors(or
offsets):Theseerrors are
definedastheconstant
valuesbywhicha
measurementinstrument’s
readingsareofffromthe
true or referencevalue
(or amastervalue).
Random Errors:Theseare
measurementerrors causedby
differencesamongoperators,
differencesamongthemeasuring
equipment,differencesover time,
or thedifferencesdueto change
intheenvironmental
conditions….
Partto partvariation
variation
Environmentvariation
Instrumentvariation
SystematicErrors
(offsetsor biases)
RandomError
(Characterizedbythevariation)
Instrum
Environm
Operator
Measurementerror

6. MSA
Purpose andRisks
Measurementsystemanalysishelpsreducetwo
types ofrisks associatedwiththemeasurement
ofaprocessandmakingdecisions.
ThepurposeofMeasurement
SystemAnalysisis to qualifya
measurementsystemforuseby
quantifyingits accuracy,precision,
andstability •theriskoffalsealarm
•theriskofmissedopportunities.
Risks

7. MSA
Variations?
Evenwhenallthefactorsina
measurementprocessarecontrolled,
repeatedobservationsmadeduring
precisionmeasurementofany
parameter,evenunderthesame
conditions,are rarelyfoundtobe
identical.Thisis becauseofthe
inherentvariationinany
measurementprocess dueto….
procedurestemfromthefactthatno
identical.Also,differentmethodsof
wouldgiverisetovariation.
inthemeasurement.
Personandprocedure..
Factorsaffectingpersonand
twohumanbeings’visualjudgmentis
measurement—theprocedure—
Instrument..
Allmeasuringinstrumentshavea stated
accuracyoruncertainty.No instrument
canmeasurethetruevalue ofthe
parameter.Thus,theaccuracyor
uncertaintyofthemeasuring
instrumentcontributestothevariation
Environment.
Theenvironmentplaysanimportant
roleinanyprocessofmeasurement.It
mightbepossibletocorrecttheeffect
ofafewenvironmentalconditions,such
astemperatureandheightabovemean
sealevel,tosomeextent.
Workpiece.
Noworkpieceisabsolutelystable.
Thereisalwaysaninherentinstability
inanymaterialorsubstance.However
smalltheinstabilitymightbe,thisgives
risetovariationinthemeasurement
process.
Standard.
Therearedifferentlevelsofstandardin
thetraceabilitychaininordertoprovide
measurementtraceability.Eachof
thesestandards,inturn,introduces
somevariation.Factorsaffectingthe
standardrefertothisvariation.

8. MSA
Resolution
Theverniercalipercan
measurethequality
characteristicsside
length,anouterandinner
diameter,andadepth.
Whenclosetolerancesare
required,measurements
aretakenwitha
micrometerdueto its
superioraccuracyovera
verniercalliper..
Resolution
The resolution,or discriminationofthemeasurementdevicemustbesmall
relativeto thesmaller ofeitherthespecificationtoleranceortheprocessspread
(variation).
Asaruleofthumb,themeasurementsystemshouldhaveresolutionofatleast
1/10ththesmaller ofeitherthespecificationtoleranceortheprocessspread.Ifthe
resolutionis notfineenough,processvariabilitywillnotberecognizedby the
measurementsystem,thus reducingitsefficiency.

9. MSA
Measurement
uncertainty
Everymeasurementis subjectto someuncertainty.Ameasurementresultis
onlycompleteif itis accompaniedbyastatementoftheuncertaintyinthe
measurement.Measurementuncertaintiescancomefrom themeasuring
instrument,from theitem beingmeasured,from theenvironment,from the
operator,andfrom othersources.
Suchuncertaintiescanbeestimatedusingstatisticalanalysisofasetof
measurements,andusingotherkindsofinformationaboutthemeasurement
process.

10. Machines
MSA
Measurement Systems
Therearedifferenttypesofmeasurement
systemsdependingonthesituationeachwith
its ownmethodologyand purpose….
CoordinateMeasuring
VisionMeasuringSystems
OpticalMeasurement
SensorSystems
DigitalScale
FormMeasurement
VibrationMeasures

11. MSAis partofallimprovementprocess,
understandingmeasurementerrorsis
vitaltoimprovement……..
MSA
The MSA Process
WheredoesMSAhappenin atypical Process?
People
Methods
Material
Equipment
Environment
PROCESS
Product
Input Process/System Output
Accurate Measurement
usingMeasurementSystemAnalysis
Identifying
Improvement
Opportunities

12. MSA
Measurement data
Measurement
Variation
Stdev()is ameasure
ofspread( variation)
inthedata.

2
Wheredo errorsand
uncertaintiescomefrom?
Manythingscanunderminea
measurement.Flawsin the
measurementmaybevisibleor
invisible.Becausereal
measurementsarenevermade
underperfectconditions,errors
anduncertaintiescancome
frommeasuringinstrumentsand
theitembeing measured.

13. 
2
MSA
Measurement Factors
Variationshappen
dueto avarietyof
reasons,someof
thesearelisted
here….
Management:trainingprograms,
metrology system,supportofpeople,
supportof qualitymanagementsystem.
Samples:materials,itemstobetested,
samplingplan,samplepreparation.
Process:testmethod,specification.
People:operators,training,education,
skill,care.
Equipment:measuringinstrument,
calibration,fixturing

2

14. MSA
TotalVariance
The totalvariationyoumeasureusinga
measuringequipmentis thesum of
processandmeasurementvariance……
Howgoodisour
measurement
system?
Process
Variance

2
p
Measurement
Variance

2
m
Total
Variance

2
T

2
T
2
p
2
m= +

2=TotalVariance
=PartVariance
=MeasurementVariance
T

2
p

2
m

15. Measurement
System
Errorsareof2 types,
onethat impacts
accuracy andtheother
the precision.
MSA
Measurement Errors
Accuracyis closenesstotheexpectedand
precisionis theabilityto beconsistenttothe
expected….
Accuracy:
Bias
Linearity
Stability
Precision:
Repeatability
Reproducibility

16. MSA
Accuracyvs Precision
Whenmakinganymeasurement,it is
normalpracticetorepeatthe
measurementinordertoensurethat
thedatageneratedisrepeatable.
Thereliablemeasurement:Itis also
importantto makesurethatthedata
generatedis reasonablyaccurateby
takingcareto usemeasuring
instrumentsthatarecalibrated.Then,
whenthesamemeasurement is made
byacustomer, thedatashouldbe
closetothefiguresgeneratedbythe
manufacturer,thatis tosaythedata
shouldbereproducible.Itis onlythen
thatthedatathathasbeengenerated
is considered reliable.
AccuracyRepresentation
Whentheaccuracyofa
micrometerwitharangeof0-25
mmandaleast countof1μis
statedas±4μ,it meansthatif
thismicrometergives areading
of20.255mm,theactualortrue
valueofthemeasurandcanbe
20.255mm±4μ,
i.e.between20.251and20.259
mm.
PrecisionRepresentation
Ifamicrometerisusedto
measurethediameterofasteel
pinanumberoftimesata
certainpoint andthevaluesof
20.253,20.252,20.250,20.251
mmareobtained,thenthe
precisionortherepeatability
ofthemeasurementcanbe
statedas0.003mm
(20.253–20.250mm).

17. 
2
T
2
p
2
m
MSA
Types of measurement
Variation
= +

2
m
Reproducibility
Repeatability,
Linearity
StabilityBias

18. 
2
T
2
p
2
m
MSA
VariationDefined
= +
Biasis thedifference
betweentheobserved
andthereferencevalue.
Theconsistencyof
measurementsover
time

2
m
Repeatability(Equipmentvariation):
variationinmeasurementsunder
exactconditions.
Reproducibility(Appraiservariation):
variationintheaverageof
measurementswhendifferent
operatorsmeasurethesamepart
Ameasureofthe
biasvaluesthrough
theexpectedrange
ofmeasurements
Repeatability&
Reproducibility
Linearity
StabilityBias

19. MSA
Bias
Thereferencevalue,also
knownastheaccepted
referencevalueormastervalue,
is avaluethatservesasan
agreed-uponreferenceforthe
measuredvalues.Areference
valuecanbe determinedby
averagingseveral
measurementswithahigher
levelofmeasuringequipment.
Bias
R
e
f
erence
Va
lue
Observed
AverageValue
Calibrationis ameasurementprocessthatassignsvaluesto thepropertyofan instrumentrelativeto reference
standardsor to adesignatedmeasurementprocess.Thepurposeofcalibrationis to eliminateorreduce biasin
the user's measurementsystemrelativeto the reference base.The mostcriticalelementofanymeasurement
process is the relationshipbetweenasinglemeasurementand thereference basefortheunitofmeasurement.
The reference base is the ultimatesourceofauthorityforthe measurementunit.

20. MSA
Inferencefrom Bias
Observed
AverageValue
Bias
Bias% = Bias/ ProcessVariation
R
e
f
erence
Va
lue
Ifthereis arelativelyhighbiasexaminethe
followingpotentialrootcauses
• Appraisersnotfollowingthemeasurement
procedure
• An error inmeasuringtheReferenceValue
• Instabilityinthemeasurement.
• IftheSPCchartshowsatrend,the
measurementdevicecouldbewearingor
calibrationcouldbedrifting.
Bias =Observedvalue–ReferenceValue
ProcessVariation=6xStandard Deviations(Sigma)

21. MSA
Linearity
Observed
AverageValue
Observed
AverageValue
Bias2
OperatingRange
250mm50mm
Themeasuringinstrumenthasmorebiaswhenmeasuringhighervalues,Bias2is
higherthanBias1…
Bias1
Linearityis the
differenceinthebias
valuesthroughthe
expectedoperating
rangeofthegauge.

22. MSA
Stability
Stability(or drift) is thetotal
variationinthemeasurements
obtainedwithameasurement
systemonthesamemaster
or partwhenmeasuringa
singlecharacteristicoveran
extendetimeperiod.
Time
Stabilityreferstothecapacity of ameasurementsystemtoproducethesame values
over timewhen measuringthesamesample.Aswithstatisticalprocess control
charts,stabilitymeans theabsenceof"SpecialCauseVariation",leavingonly
"CommonCauseVariation"(randomvariation).

23. MSA
Repeatability Repeatabilityis thevariationinmeasurements
obtainedby oneappraiserwithonemeasurement
instrumentwhenusedseveraltimeswhilemeasuring
theidenticalcharacteristiconthesamepart.

24. MSA
Reproducibility
Reproducibilityis the
variationintheaverageof
themeasurementsmadeby
differentappraisersusing
thesamemeasuring
instrumentwhenmeasuring
theidenticalcharacteristic
onthesamepart
Operatorskill-somemeasurementsdependontheskill
andjudgementoftheoperator.Onepersonmaybebetter
thananotheratthedelicateworkofsettingupa
measurement,oratreadingfinedetail byeye.Theuseof
aninstrumentsuchasastopwatchdependsonthereaction
timeoftheoperator.
Threeoperatorsmeasurethesamecharacteristic

25. MSA
Computation
Therearethreemethodstocompute
GageR&R:
•
•
•
RangeMethod
AverageandRangeMethod
AnalysisofVariationMethod

26. MSA
Computation
✓ EvaluateRepeatabilityandReproducibilityseparately
✓ Conductedtypicallywith threeappraisersand10parts
StepsintheVariable
GaugeStudy
(Averageand Range
Method)
✓
Each appraiser measureeachpart3timesinrandomorder
✓
Compute EV(equipmentvariation)-Repeatability
✓
Comupte AV(appraiservariation)-Reproducability
✓
Gage R&R,(R&R)²=EV²+AV²
✓
PartVariation PV
✓
TotalVariationTV²=(R&R)²+PV²8
7
6
5
4
3
2
1

27. MSA
Decisionmaking
Ifrepeatabilityis largecomparedtoreproducibility,then
•
•
•
•
Theinstrumentneedsmaintenance;
Thegaugeshouldberedesignedtobemorerigid;
Theclampingor locationforgaugingneedsto beimproved;
Thereis excessivepartvariation.
Ifreproducibilityis largecomparedtorepeatability,then
•Theappraiserneedsto bebettertrainedinhowto useandreadthegauge
instrument;
Calibrationsonthegaugedialarenotclear;
Afixtureofsomesortmaybeneededto helptheappraiserusethegaugemore
consistently.
•
•

28. MSA
MeasurementCapabilityIndex-P/T
GaugeCapabilityRatio
TP/
PrecisiontoToleranceRatioisthe%of
tolerancethatis taken upbythe
bothmeasurement error which includes
repeatabilityandreproducibility.
Bestcase:10%Acceptable:30%
Tolerance= USL -LSL
5.15*
MS
Tolerance

29. MSA
Decisionmaking
Thoughagageof gagaRR
<30%is usablethetarget
shouldbetoachieve<10%.
<10% 10-30 % >30%
Great
Accept
Measurement
System
Usable
Maybeaccepted
based oncost
NotAcceptable
Measurment
Systemneeds
Improvement
GageRepeatabilityandReproducibility

30. MSA
Sample Output
Least
CountCharacteristic Gauge TV %EV %AV %R&R
Fillerneckholedia. Verniercalipers 0.01 0.1107 35.77 72.88 81.12
Tubecrownthickness Dialvernier calipers 0.02 0.1866 49.19 38.06 62.1
Tubewidth Dialvernier calipers 0.02 0.0365 10.71 43.83 46.87
Inletholedia. Verniercalipers 0.01 0.1057 86.56 20.09 88.83

31. www.qualimations.com
[email protected]