D2 vs d2* Explanation in My Terms

E

eturkan

[FONT=&quot]I am a Math lover, recent Purdue MSME graduate and Gage R&R and statistics and Math is what I love since I started my career in QE. I had to spend less time to understand Navier-Stokes equation/tensor notations. I prepared an improvement opportunity list for MSA 4?th Ed and will send it to AIAG soon.[/FONT]
[FONT=&quot]
I'd like to thank everyone on the forum because the information I gain on the forum has been VERY helpful.[/FONT]
[FONT=&quot]As many other users I got confused between d2 and d2* because of misleading statements on MSA'th edition and the change in their methods (constants, d2 usage with small group sizes etc) with the earlier editions.

This is my first post (I believe) and I wanted to explain d2 vs d2* in my perception.[/FONT]
[FONT=&quot]In GRR, We calculate all R_bar in one or another way to determine the variation and, [/FONT]
[FONT=&quot]m= Length of the Range equation that is used [/FONT]
[FONT=&quot]R=(Max x(1),?.x(m))- (Min x(1),?.x(m)))[/FONT]
[FONT=&quot]g=How many unique range calculations do we have for the final R_bar value[/FONT]
[FONT=&quot]g= In R_bar= average (R(1)+...+R(g)) [/FONT]
[FONT=&quot] In a O (3 Operator) x T( 3 trials) x P (10 parts ) Gage R&R study (Prepare your own excel GRR Sheet for a better understanding. It takes 2 hours for a beginner but then you may have a solid understanding)[/FONT]
[FONT=&quot]Repeatability is calculated by first with a range operation for each individual reading set (Trial 1 thru Trial 3 so our Range equation length is 3[/FONT]
[FONT=&quot]Reproducibility is based on the average reading (of all parts and trials) of each operator. Our range equation is [/FONT]
[FONT=&quot]R=Max [ X_BAR_ (OP1) ,X_BAR_ (OP2) , X_BAR_ (OP2) ]- MIN [ X_BAR_ (OP1) ,X_BAR_ (OP2) , X_BAR_ (OP2) ][/FONT]
[FONT=&quot]So m=3 since the length of R is 3 in this example. [/FONT]
[FONT=&quot]g=1 since we only have ?1? unique R equation. [/FONT]
[FONT=&quot](There is a further deduction step due to contamination coming from repeatability but that is not the subject of this topic)[/FONT]

[FONT=&quot]Part to Part variation is calculated by comparing the average of all readings for each individual part [/FONT]
[FONT=&quot]R=Max [ X_BAR_ (PART1) ,X_BAR_ (PART2) ,?, X_BAR_ (PART10) ]- [/FONT]
[FONT=&quot]MIN [ X_BAR_ (PART1) ,X_BAR_ (PART2) ,?, X_BAR_ (PART10 ][/FONT]
[FONT=&quot]So m=10 since the length of R is 10 in this example. [/FONT]
[FONT=&quot]g=1 since we only have ?1? unique R equation.
[/FONT]


[FONT=&quot]I hope it's been helpful.
[/FONT]
 

Miner

Forum Moderator
Leader
Admin
Welcome to the cove. I am a long time ago Purdue grad (1982 BSMET). Good first post, and we hope to see you return often.
 

Bev D

Heretical Statistician
Leader
Super Moderator
:topic:
Nice start!

But I must start with a welcome from a Hoosier! Don't know if you were born in Indiana but I must welcome a purdue grad.

And Miner, you never said you were a boilermaker!

I am not, having graduated from michigan tech...
 

Miner

Forum Moderator
Leader
Admin
Bev, I am also a native Hoosier from southeast of Indy. I didn't realize that you were also.
 

Bev D

Heretical Statistician
Leader
Super Moderator
Miner - I was born in Columbus IN and raised in Huntington. Hoosier through and through.
 

Miner

Forum Moderator
Leader
Admin
Born and grew up in Shelbyville. Worked in Marion, Wabash, Peru and Logansport (same company, different plants). This is scary.
 

Jim Wynne

Leader
Admin
Born and grew up in Shelbyville. Worked in Marion, Wabash, Peru and Logansport (same company, different plants). This is scary.

I was just in Columbus (IN) visiting a customer a few weeks ago. I didn't see any of you people there. :tg:
 
Top Bottom