
This thread is carried over and continued in the Current Elsmar Cove Forums 
The New Elsmar Cove Forums 
The New Elsmar Cove ForumsThe Elsmar Cove Forums
Statistical Techniques and 6 Sigma G R&R

UBBFriend: Email This Page to Someone!  next newest topic  next oldest topic 
Author  Topic: G R&R 
AllenLee Forum Contributor Posts: 13 
posted 02 July 2001 11:41 PM
I have a silly question, does the G R&R mean 5.15¦Òm ( Standard deviation of measurement system)? Thanks! Allen  IP: Logged 
Ken K. unregistered 
posted 03 July 2001 11:35 AM
A gage R&R is a method of evaluating several potential sources of variation in a measurement system. The typical gage R&R quantifies the variation due to differences between parts, between operators, and between units treated identically (gage variation). The idea is that the variation between operators and the gage variation should be small compared to the total variation observed. If it is not, you can use the gage R&R to determine whether you need to retrain your operators or improve your gage, or both. If you haven't already, I strongly suggest you purchase a copy of the AIAG Measurement System Analysis Reference Manual, available at http://www.aiag.org/publications/quality/dcxfordgm.html Follow the MSA2 link. Only $11 for nonmembers. It is well written. I also strongly suggest you invest in statistical software capable of analyzing the results properly. Your best bets are MINITAB, StatGraphics, and JMP. I prefer MINITAB. Ken K. IP: Logged 
Rick Goodson Forum Wizard Posts: 135 
posted 03 July 2001 06:02 PM
Ken, You opened the door for another question. Why do you prefer Minitab over StatGraphics or JMP?Regards, Rick IP: Logged 
AllenLee Forum Contributor Posts: 13 
posted 03 July 2001 08:12 PM
Thanks for your response. But my question is whether the G R&R is the 90% of the measurement system distribution (5.15 sigma of measurement system). For I see, in the formula calculating EV and AV, it includes the constent index 5.15, am I right? Thanks! Allen  IP: Logged 
Ken K. unregistered 
posted 04 July 2001 07:04 AM
The formulas use 5.15, which reprsents the number of standard deviations that contain the central 99% of the normal distribution. The only way to make a realistic statement regarding the %contribution of the variation is to obtain variance components using the ANOVAmethod (these are the variances as opposed to the std. deviations). The variances can be added to calculate percentages of the total observed variation. By dividing the R&R variance by the sum of ALL the variances you can determine the %contribution of the measurement system. That is one reason I like MINITAB better than JMP. MINITAB provides both the average & range and ANOVA methods. I prefer the ANOVA method. JMP can do the ANOVA method, but more handwork needs to be done. I used to have the demo for StatGraphics 5, but it has expired and I can't remember much about its GR&R tool. I tend to like MINITAB because overall it has a better laid out set of tools. With respect to the GR&R, StatGraphic's latest version added a onepage report much like MINITAB has had for several years, but I still like MINITAB's better. Ken K. IP: Logged 
Rick Goodson Forum Wizard Posts: 135 
posted 05 July 2001 09:31 AM
Ken, Thanks for the input! Always nice to get the users point of view when evaluating software. Rick IP: Logged 
Ken K. unregistered 
posted 06 July 2001 03:41 AM
If MINITAB was "taken away from me" I'd probably have a tough time deciding between JMP and StatGraphics. I'd probably have to go with SG, since JMP's tools aren't really AIAG (Automotive Industry Action Group) complient for Process Capability, and I work for an automotive industry company. We've used both in the past  StatGraphics through their Windows version 2, and then JMP 3.x. It wasn't until I looked at the StatGraphics 5 Demo that I realized how poor SG's data manipulation tools are. Now that I think of it I did end up using Excel for most of my manipulation when I used SG. MINITAB just rolls over SG in that category. JMP's data manipulation tools are a little better, but not much. The other real advantage MINITAB has is in the area of reliability/survival analysis  their tools are becoming really nearworld class. I'm told their goal is to have world class reliability tools by version 14, which should roll out in 2002. From what I've seen in version 13.3 I'm sure they'll do it. High marks for implementation and ease of use. Ken K. IP: Logged 
All times are Eastern Standard Time (USA)  next newest topic  next oldest topic 
Please Visit the new Elsmar Cove Forums! All these threads are there and much more!