Mike Albright 248-935-4237 albright@signalxtech.com - PowerPoint PPT Presentation

Production NVH Test Tools and Methods Mike Albright 248-935-4237 albright@signalxtech.com www.signalxtech.com

Production NVH Test Tools and Methods Requirements of successful deployment : • Good test machine • low machine self-noise, • stiff repeatable clamping mechanisms, • proper isolation from environmental noise contamination • The ability to exercise the product to optimize defect signatures • Comprehensive sensor suite • Data acquisition system that is fast, reliable, and maintainable • Metric design environment with batch processing capability • Database system with automated archival and reporting

Production NVH Test Tools and Methods Good Test Machine : • Good test machine • low machine self-noise, • stiff repeatable clamping mechanisms, • isolation from environmental noise contamination The “Golden T” Is formed from the elements inside the dashed circle Spindle with Spindle with Integral Encoder Integral Encoder Compliance couplings Static torque Static and dynamic torque Spindle with Torsional acceleration Integral Encoder

Production NVH Test Tools and Methods Optimize Defect Signatures : • The ability to exercise the product to optimize defect signatures

Production NVH Test Tools and Methods Microphone Comprehensive Sensor Suite : Vibration Touch Probe Laser Vibrometer In-line Rotary Torque Sensor

Production NVH Test Tools and Methods Fast, Reliable, Maintainable System : MajX NVH Controller • PXI and LabVIEW RT • Very Fast • Limited services running • Stable execution timing • High uptime reliability • Local data storage • Commercial of the Shelf availability • Scalable I/O • High temp & 24/7 option

Production NVH Test Tools and Methods Interactive metric design environment with batch processing capability : From Signals to Features To Metrics

Meaningful Metrics for Automated Test What Makes a Good Metric? Properties of effective metrics : 1. Discrimination: • Isolate the defect signature 2. Repeatability: • Essential requirement of any metric 3. Dynamic range: • Grade defect severity 4. Distribution: • Normal or other tailed distribution 5. Stability: • No uncompensated change over time 6. Understandable: • Accessible and standardized 7. Diagnostic Aid: • Suggest corrective action

Meaningful Metrics for Automated Test Failed Seat Adjusters Defect Detection • One or more metrics will indicate a defect • Cannot reject good parts . • Cannot let bad ones out of the factory • Classification is Human-driven - initially • Defects become associated with patterns • Metric refinement will be ongoing • Intuitive metric design environment • Batch processing capability • Metric set should be: • Discriminating • Repeatable • Understandable • Suggest corrective action

Production NVH Test Tools and Methods Database system with automated archival, retrieval, reporting, batch processing NVH Controllers Scanner Database Query Reports Job List • Manage Upload Schedule for • Build Action List • Storable report definition • Define individualized all NVH controllers in a domain • Sort Metrics/files • Source data from Query reporting tasks • Retrieve raw data, processed • Storable Queries or batch processing runs • Schedule stored Queries results and meta data • Report source • Automated alarms with • Schedule stored Batch • Populate database electronic notification Processing runs Batch Processing • Report NVH Controller Status • Selection of report types • Schedule reports with • What-if metrics - Cross Correlation automated source data • Re-use datafiles - Regression Analysis update from Query or - Run Charts • Storable Batch def. - Histogram Batch processing jobs • Report source Data Management Functions

Production NVH Test Tools and Methods Production test data serves two communities PASS FAIL Engineering Wants an Explanation Production wants a Solution

Production NVH Test Tools and Methods On the floor – NVH system presence is minimal MajX NVH Controller Signals Final Metric Results

Production NVH Test Tools and Methods But engineers need fast access to detailed analysis of raw signals Without interrupting ongoing production When a part fails, detailed analysis must be immediately available