E6S-058 - Rule out the Ruler Part 5B- Gauge R&R Variable MSA
Intro: Welcome to the E6S-Methods podcast with Jacob and Aaron, your source for expert training, coaching, consulting, and leadership in Lean, Six Sigma, and continuous improvement methods. In this episode number 58, “Rule out the Ruler – Part 5B,” we go through the do’s and don’ts, and our own personal preferences, (for better or worse), when planning and setting up a crossed variable gauge R&R. Here we go.
***Rule out the Ruler - Measurement (Decision) System Variation***
Part 5B
I How to perform Type-2 Crossed GR&R
a. Steps (by the book)
i. Ensure calibration of the gauge
ii. Select 10 parts from the process representing full process variation
iii. Select 2+ fully trained operators
iv. Each operator measures the parts in random order 2 times
1. (10 parts, measured 2 times, by each of 2+ operators) = 40+ measurements
v. Analyze Results
vi. Implement any improvements
vii. Re-run Gauge Study
viii. Document findings
b. Steps (our commentary on the same steps)
i. Ensure calibration of the gauge,
1. and/or Type 1 Gauge study
ii. Select 10 parts from the process representing full process variation
1. It Depends! What’s the goal?
a. Measurement Variability relative to customer spec: choose from process, but also known good & bad around the lower and upper specification limits
i. Necessary when relying on a gauge to determine good or bad product.
b. Measurement Variability relative to process variability, choose straight from the normal operating process (randomly spread over some time frame)
i. Helpful when using a gauge to improve a process
c. Deeper dive into measurement variability (Aaron’s preference), fewer parts, measured more often.
iii. Select 2+ fully trained operators
1. Or 3+
2. Each operator who is expected to use a gauge should have proof of capability, perhaps in the form of a GR&R comparison to standard. Uncovers Operator*Part interactions
iv. Each operator measures the parts in random order 2 times
1. About the parts & measurements
a. By the book: (10 parts, measured 2 times, by each of 2+ operators) = 40+ measurements
b. Aaron’s preference (5+ parts, measured 3 times, by 2+ operators = 40+ measurements
i. Can better study variation within operators (repeatability on single parts) and get a truer average. Cannot get a standard deviation or a true average on only 2 measurements per part
ii. Sometimes it’s difficult to track down 10 parts.
iii. Easier sell to operators, get their buy-in to complete the study
2. About the randomness
a. Recall a fully randomized study has different operators jumping in and out of the study, performing the test in a random sequence until all measurements are done
i. Sometimes it’s too inconvenient to keep switching back and forth between operators – Randomize within operators
ii. Sometimes it’s too inconvenient or costly to keep switching back and forth between different ends of the measurement scale
b. With any move away from full randomness, you risk confounding your results with some change that happens over time.
i. If Operator C is way different than. You won’t know if it is due to Operator C or some time variant that coincided with Operator C’s work. (Parts destroyed, gauge damaged, temp/humidity or fatigue factor etc.)
v. Analyze Results (Next Episode)
vi. Implement any improvements
vii. Re-run Gauge Study
viii. Document findings
Outro: Thanks for listening to episode 58 of the E6S-Methods Podcast. Stay tuned for episode number 59 for the analysis of a real-life gauge R&R study and an overview of the AIAG rules, closing out our “Rule out the Ruler” series with part 5C. If you would like to be a guest on podcast, contact us through our website. Follow us on twitter @e6sindustries and join a discussion on LinkedIn. Subscribe to past and future episodes on iTunes or stream us live on-demand with Stitcher Radio. Find outlines and graphics for all shows and more at www.E6S-Methods.com. “Journey Through Success”