1、07FTM06Using Barkhausen Noise Analysis for Process andQuality Control in the Production of Gearsby: S.J. Kendrish, T.J. Rickert and R.M. Fix,American Stress Technologies, Inc.TECHNICAL PAPERAmerican Gear Manufacturers AssociationUsing Barkhausen Noise Analysis for Process andQuality Control in the P
2、roduction of GearsStephen J. Kendrish, Theo J. Rickert and Robert M. Fix, American StressTechnologies, Inc.The statements and opinions contained herein are those of the author and should not be construed as anofficial action or opinion of the American Gear Manufacturers Association.AbstractThe use o
3、f magnetic Barkhausen Noise Analysis (BNA) has been proven to be an effective tool for thenon-destructive detection of microstructural anomalies in ferrous materials. Used as an in-process tool forthe detection of surface temper burn, heat treat defects and stresses, BNA is a quick comparative andqu
4、antitative alternative to traditional destructive methods.Applications of BNA as a quality control tool in gear production have existed for nearly two decades. Thispaper presents examples that demonstrate how BNA is used to evaluate changes in microstructuralproperties. Quantitative results correlat
5、e BNA test values to X-Ray diffraction values for the detection ofchanges in surface residual stress. Other quantitative analysis correlates BNA test values to surfacehardnessvaluesforthedetectionofre-temperingburn. QualitativeresultscorrelateBNAtestvaluestoacidetch patterns/colors for the detection
6、 of re-tempering burn defects.Roboticallyautomatedsystemscantestgearteeth andprovide immediatefeedback forprocess controlandquality assurance.Copyright 2007American Gear Manufacturers Association500 Montgomery Street, Suite 350Alexandria, Virginia, 22314October, 2007ISBN: 978-1-55589-910-31Using Bar
7、khausen Noise Analysis for Process andQuality Control in the Production of GearsStephen J. Kendrish, Theo J. Rickert and Robert M. Fix,American Stress Technologies, Inc.IntroductionIn the manufacturing process of grinding hardenedgear teeth to final dimensions, there exists the pos-sibility of therm
8、al damage to the ground surface.This thermal damage results in a changeof thesur-face microstructure and/or surface residualstresses, which, in turn, leads to poor performanceand premature failure of the end product. In the ex-treme, grinding can produce rehardening tensilestresses and cracking. The
9、 practice of testing forthis damage has been implemented in manydestructive and non-destructive forms.BarkhausenNoiseAnalysis(BNA)isa provennon-destructive method to accomplish this task. Unlikemany other methods, BNA is a quick, efficient andsafe way to provide feedback with regards tochanges in mi
10、crostructural properties of the groundgearsurfaces. Itisquick,inthatrealtimedataisgiv-enwithlittleornomanualanalysisrequired. Itiseffi-cient,inthatitdoesnotrequireanysurfaceprepara-tion. It is safe, in that it does not use hazardouschemicals.With the proper initial setup and studies, BNA is auseful
11、tool in monitoring the metallurgical effects ofthe grinding process and ensuring product quality.Barkhausen noiseBarkhausen Noise is produced when a ferromag-netic material is influenced by an external magneticfield. As the ferromagnetic material becomes mag-netized, magnetic domains grow or are con
12、sumedundertheinfluenceoftheexternalfield. Thechang-ing domain volumes emit electromagnetic signalsthat are inductively detected. The aggregate resultof signals received is Barkhausen Noise.The apparatus for creating and detecting Barkhau-sen Noise consists of a processing unit that sendsan AC curren
13、t to an electromagnet. An inductivepick-up collects the resultant electromagneticpulses and sends them back to the processing unitfor filtering and amplification. The results reportedherein are in terms of the RMS amplitude of theBarkhausen Noise or BNA value.In the machined, hardened and properly g
14、roundgear, the microstructure remains hard and the sur-face stresses are compressive. With grinding in-ducedheatatandabovetheoriginaltemperingtem-perature of the material, the resultantmicrostructure becomes softer and the favorableresidual surface stresses are less compressive.Initial setup and ver
15、ificationSince Barkhausen Noise provides data relating tothe magnetic and microstructural properties of thematerial, it is useful as a comparative tool. In rela-tion to Barkhausen Noise, softer material is easierto magnetically influence by an external field andthe BNA values are higher in compariso
16、n to a hardmaterial. Once the baselines for known acceptableparts are established, deviations from the normalcurvescanbeevaluatedanddeemedacceptableorrejected. There are no NIST traceable standards.Thetraceabilityof thesystem (CPUandsensors)isdependent on master set-up parts. Periodicallyverifyingth
17、atthevaluesofthemasterpartshavenotchangedover timeensures thatthe systemis stillinalignment.VerificationmustalsobemadetocorrelatetheBNAvalues to some other measure of material propertyor degree of retemper burn. For retemper burn,BNA values are frequently compared to nital etchresults, which is the
18、traditional method used to testforretemperburns. Othernotablemethodsusedforcorrelation are fatigue life, residual stress andhardness.Instrumentation and techniqueThe Barkhausen data for the results presented inthis paper were obtained using an AST RoboScan600 (Figure 1). A robotically articulated ar
19、m holdsandmovestheBNAsensor,whichimprovesrepeat-2ability. The gear is manually installed into the hold-ing device, and inductive proximity sensors or cam-eras locate the surfaces to be tested. Pre-installedprograms can test any number of teeth on the frontand rear flanks. In this study, the front an
20、d rearflanks are labeled A and B, respectively. The testarea consists of 80% of the total tooth width at theapproximate pitch diameter on the tooth depth(Figure 2). Greater resolution can be obtained byslowing the speed of the robots scans.Figure 1. RoboScan 600.Figure 2. Robot during test.The gear
21、used for this study is shown in Figure 3. Itis an induction hardened and finish ground helicalgear that consists of 39 teeth. The overalldiameteris approximately 195 mm and the flank width isapproximately 29 mm. Sample output of the Ro-boScan tests are shown in Figure 4. The tabulateddata is calcula
22、ted from each scan on the graph.Maximum, minimum and average values, amongothersforeachtoothflankareprovidedimmediatelyafter each scan. In this study, the gear was testedfive times to verify repeatability.Figure 3. Test gear.Figure 4. Sample output graph withtabulated data.SincehigherBNAvaluescorres
23、pondtoretemperedor softer parts, the maximum values obtained foreachflankareofimportance. Agraphicalrepresen-tation of the maximum values for each tooth isshown in Figure 5. For an in-depth analysis, thefront (A) and back (B) flanks of tooth #12 and #32will be examined. Figures 69showtherawtestdata
24、for each flank, tested five times.3Figure 5. Maximum BNA values for eachflank.Figure 6. Tooth 12, A flankFigure 7. Tooth 12, B flankFigure 8. Tooth 32, A flankFigure 9. Tooth 32, B flankQuantitative setup and analysisAs part of an initial study or setup, BNA values arecorrelated to a quantitative, t
25、raceable means ofanalysis. For this study, X-Ray diffraction and sur-face hardness readings were performed. The sur-face residual stresses and hardness values on se-lectgearteethweremeasuredintwo relevantspotson each flank (see Figures 10 - 13) that align withthex-axislocationsofapproximately20and80
26、inthe scans shown in Figures 6 9.The surface stress values and hardness valueswere then compared and correlated to the obtainedBNA values.4Figure 10. Tooth 12, A flankFigure 11. Tooth 12, B flankFigure 12. Tooth 32, A flankFigure 13. Tooth 32, B flankOnceacorrelationisestablished,andifthematerialand
27、 processing are constant, BNA can then asso-ciate a surface residual stress and/or hardness tothe tooth flanksin afraction ofthe timeit wouldtakefor X-Ray diffraction or hardness testing. EachBNA scan takes approximately 1.5 seconds perflank. See Figures 14 and 15 for quantification ofthe correlatio
28、n.Figure 14. Residual surface stress as afunction of BNA valueFigure 15. Hardness as a function of BNAvalueQualitative setup and analysisAlso, aspart ofan initialstudy orsetup, BNAvaluesare correlated to chemical etch results. With theBarkhausen data obtained for all the tooth flanks,the test gear w
29、as then acid etched. Looking at thesame teeth used for the X-Ray correlation, Figures16 and 17 show the nital etched teeth with dark-ened, retempered regions as well as un-burned,light regions.5Figure 16. Etched tooth 12 B flankFigure 17. Etched tooth 32 A flankBy establishing a visual relation betw
30、een maximumBNA values and degree of darkness of the etchedsurface, a correlation can be made. From this cor-relation, accept/reject criteria can be established.See Figure 18.Once this relationship relating BNA values to thesurface hardness of the gear teeth is established, amanufacturer can quickly
31、determine if gears arebeingproperlygroundwithoutanythermaldamage.Theprocessofnitaletchingalonewouldhavetakenmuch more time and possibly cause the gear to beunfit for use.Figure 18. Relation between BNA and etchcolorConclusionsBarkhausen Noise Analysis is a quick and effectivemethodfordetectingretemp
32、erburnandmicrostruc-tural defects in gears. The times for checking theteeth with BNA are considerably short in compari-son to the comparativemethods usedfor setupandverification for this study.All of the ground flanks of a 39 tooth gear weretested with automated Barkhausen Noise inapproximately five
33、 minutes. Aside from normalcleaning of the part, no special preparation was re-quired. No chemicals were necessary in theperforming of the test, and the test did not create ahazardous environment.References1. ANSI/AGMA 2007-B92, Surface Temper EtchInspection after Grinding.2. SAE J784, Residual Stress by X-RayDiffraction
