Gage Repeatability Reproducibility for Continuous Data Instructors Name
Gage Repeatability & Reproducibility for Continuous Data Instructor’s Name
Introductions n n n Your Name Your Position What type of measurements you take in a given day
Agenda n n n Introduction…………………. . ……(30 minutes) Theory of gage R & R……. . (50 minutes) How to plan and execute a gage R & R study. . . ……………………. . (20 minutes) --- Break --- n How to calculate the total gage R & R for a measurement system…………. . . ……(75 minutes) --- Break --- n How to improve a gage R & R result …………………. . (60 minutes) --- Break --- n Conclusion & Wrap-Up………. . . …. …(20 minutes)
Theory of Gage Repeatability & Reproducibility
What do we expect from data? n Accuracy – Calibration Accurate, but not precise n Precision – Repeatability (Equipment) – Reproducibility (Operator) Precise, but not accurate
Accuracy AND Precision n n A good measurement system must provide data that is BOTH accurate and precise. How do we know? By conducting a GAGE REPEATABILITY AND REPRODUCIBILITY STUDY
Measurement vs Measurement System n n Measurement - value or data Measurement System - everything involved in performing the measurement – measurement tool – part or sample – operator – procedure – environment Gage R & R takes into account the entire measurement system, not just the measurement.
Measurement Resolution n n The resolution of a gage is the smallest unit of measure the gage is able to read. A gage should have a resolution of less than 10% of the total specification tolerance. Example: Specification: +/- 0. 050 Q: What is the maximum gage resolution? Total tolerance: 0. 050 + 0. 050 = 0. 100 10% of total tolerance: 0. 100 * 0. 10 = 0. 010 A: The gage should have a resolution of 0. 010 or less.
Measurement Variation n n Total Variation = Actual Part-to-Part Variation + Measurement Variation can come from anywhere within the measurement system: Man (operator) Method Material (part) Measurement Machine (tool) Mother Nature (environment)
Exercise 1: Length of a Pen n n Group Size: 3 - 4 participants Materials: 6” plastic ruler, Exercise 1 in Student Manual Task: Record all possible sources of measurement variation if your group was asked to measure the length of pen. Duration: 10 minutes
Total Gage R & R n n The total gage R & R of a measurement system is the final number calculated from a gage R & R study. Total gage R & R has two parts: – Repeatability or Equipment Variation (E. V. ) – Reproducibility or Appraiser Variation (A. V. ) Total Gage R & R = E. V. 2 + A. V. 2
Judging the Total Gage R & R n If Total Gage R & R is: – < 20%: ACCEPTABLE – Between 20% and 30%: CONDITIONAL – > 30%: UNACCEPTABLE
How to Plan and Execute a Gage R & R Study
Planning n The typical gage R & R study uses – 2 or 3 operators – 10 parts n n Each operator will measure each part two or three times. In general, more data is better.
Selecting Parts n n Do NOT just select 10 parts at random! Samples should cover the full range of the specification and beyond. Example: Specification: 1. 000 +/- 0. 025 Q: What should be the range of the samples? Full specification range: 0. 975 - 1. 025 A: The samples should cover the full specification range and beyond, so they should cover approximately 0. 950 to 1. 050. Bonus Q: What should the gage resolution be for this case?
Tips for Data Gathering n n n Stay as close to the existing process as possible. Use a scribe to record data. Identify samples. Randomize the order of the samples for measurement. Do not round the data.
Time to take a… BREAK!
How to Calculate a the Total Gage R & R for a Measurement System
Tips on the Gage R & R Worksheet n n Be able to perform all calculations by hand (a computer is not always available). Save a master copy of the blank spreadsheet (Read-Only). Shaded cells contain formulas. Enter data in white cells only. Record all background information at the top of first page of the worksheet.
Three Ways to Calculate n Measurement Unit Analysis – Raw calculation of the data. n Percent Tolerance Analysis – Relates measurement variation to the specification. – Product Acceptance n Percent Study Analysis – Relates measurement variation to the total process variation. – Process Control
Special Cases n Discrete data (go/no go, count, categorical) – Checks for percent agreement among all operators. – Goal: 100% agreement n Destructive testing (torque, tensile strength) – Instead of repeated measurements, subgroups of sequentially processed parts are used. – Part variation within the subgroup should be minimized.
Exercise 2: Dowel Stick Gage R & R n n Group Size: 3 - 4 participants Materials: Participant Kit, Student Manual Task: Conduct a gage R & R on the measurement system for one dimension of the dowel stick. The dimension to be measured will be assigned by the instructor. Duration: 45 minutes
Time to take a… BREAK!
How to Improve a Gage R & R Result
Recall from earlier. . . n n Repeatability - Equipment Variation Reproducibility - Appraiser Variation Tolerance Analysis - Product Acceptance Study Analysis - Process Control
Improving a System n High percent repeatability? – Operator judgment / opinion – Awkward sample – Gage calibration – Specification n High percent reproducibility? – Operator training – Sample state or position – Specification
Exercise 3: Dowel Stick Gage R & R (2) n n Group Size: 3 - 4 participants Materials: Participant Kit, Student Manual Task: Discuss how to improve the measurement system as it was used in Exercise 2. Confirm those improvements by performing a Gage R & R study on the new measurement system. Duration: 55 minutes (10 for discussion, 45 for Gage R & R)
Exercise 3: Dowel Stick Gage R & R (2) Percent Improvement = [(Exercise 2 Result) - (Exercise 3 Result)] (Exercise 2 Result)
Time to take a… BREAK!
Conclusion & Wrap-Up
Review n n The theory behind gage R & R studies How to plan and execute a gage R & R study How to calculate the total gage R & R How to improve the result of a gage R & R study.
Exam n n There is an exam associated with this course! Date: x Month yyyy Duration: 90 minutes Passing Score: 80%
Evaluations n n Anonymous Please be candid! Thank you!!
- Slides: 33