Need help with Gage R&R

[Kaji Ryoji]

Thanks for taking some time here to answer my questions.

Here is my problem.

I've been charged with doing a Gage R&R on a Part using a CMM and a fixture.

I've been asked to do a 3-10-3 and return the results using Datapage.

I understand all of the trace fields and how to use them except the "Trial Field"

Do I need to make a New trace field for each of the three measurements in order for Datapage to work properly?

I'm not very good with Datapage and this is the first R&R I've had to do so any help is appreciate.

Thank you,

Kaji

[Matthew D. Hoedeman]

10 parts, 3 trials, 3 operators.

Use 3 tracefileds:
Operator
Trial
Sample

Values will need to be correctly enter for each of thes 3 fields each time the program is run. You SHOULD run it in this order (and you really don't need to use more than 5 samples, really):


O1-T1-S1
O1-T1-S2
O1-T1-S3
O1-T1-S4
O1-T1-S5

O1-T2-S1
O1-T2-S2
O1-T2-S3
O1-T2-S4
O1-T2-S5

O1-T3-S1
O1-T3-S2
O1-T3-S3
O1-T3-S4
O1-T3-S5



O2-T1-S1
O2-T1-S2
O2-T1-S3
O2-T1-S4
O2-T1-S5

O2-T2-S1
O2-T2-S2
O2-T2-S3
O2-T2-S4
O2-T2-S5

O2-T3-S1
O2-T3-S2
O2-T3-S3
O2-T3-S4
O2-T3-S5



O3-T1-S1
O3-T1-S2
O3-T1-S3
O3-T1-S4
O3-T1-S5

O3-T2-S1
O3-T2-S2
O3-T2-S3
O3-T2-S4
O3-T2-S5

O3-T3-S1
O3-T3-S2
O3-T3-S3
O3-T3-S4
O3-T3-S5


Use numbers for the operators, not names (it will be easier).
ALL the trace fields MUST be correct or Datapage will NOT calculate the R&R for you. If even 1 is incorrect, it will not work. Also, GR&R is an add-on to Datapage, do you actually HAVE the GR&R package?

EDIT" The tracefields HAVE to be after the STAT command in the program or they will not show up in Datapage.

[Kaji Ryoji]

So I think that there may have been some confusion as far as what they want.

So from what I gather the trial field is basically another whole run of the Gage R&R and not just a specific measurement on the part.

Like I said this is my first time dealing with Gage R&R. We want 3 operators, 10 parts and 3 points measured (maybe 3 points they could be after just 1 for some basic data).

Would this only need to be 1 trial if I have 3 differen't point locations?

Also, I do have the GR&R package for Datapage.

Of course sorry for my lack of knowledge on the subject.

[Matthew D. Hoedeman]

So I think that there may have been some confusion as far as what they want.

So from what I gather the trial field is basically another whole run of the Gage R&R and not just a specific measurement on the part.

Like I said this is my first time dealing with Gage R&R. We want 3 operators, 10 parts and 3 points measured (maybe 3 points they could be after just 1 for some basic data).

Would this only need to be 1 trial if I have 3 differen't point locations?

Also, I do have the GR&R package for Datapage.

Of course sorry for my lack of knowledge on the subject.

A Gage R&R is 2 (or 3) operators, checking 5 to 10 parts, 3 times each. The 3-10-3 (or 10-3-3 more commonly) is 10 samples (parts), 3 operators (people), 3 trials (checking the same part, 3 times), NOT 3 check points. 3 check points is a little thin for a Gage R&R. You should be checking a minimum of 6 points (IMO) for a small part, 2 in each major vector axis (2 with mostly X vector, 2 with mostly Y vector, 2 with mostly Z vector) and these should be as far from the locators as possible. Don't do a 'Z' point above a 'Z' net, in other words.

1) Calibrate the TIPS
2) Align the fixture
3) Operator #1, checking all the parts, all the points, 3 times each, in order, 1,2,3,4,5 then 1,2,3,4,5 then 1,2,3,4,5
4) Same for operator #2
5) Same for operaotr #3

so, for a 10-3-3, you will end up having a total of 270 checks of ALL points in the study. 10 parts, 3 times each, 3 operators each. Each operator will load the fixture 90 times. Each operator will check the same part 3 times (trials).

[dtmtim]

For those that don't have the GR&R add-on to datapage there is some good freewere available here ...

http://www.symphonytech.com

t.

[cmmguy]

so, for a 10-3-3, you will end up having a total of 270 checks of ALL points in the study. 10 parts, 3 times each, 3 operators each. Each operator will load the fixture 90 times. Each operator will check the same part 3 times (trials).

Matt, check your math....

[Matthew D. Hoedeman]

Matt, check your math....

DANG! One too many X3 in that, ain't there?

[Kaji Ryoji]

Just an update.

I think I have it under control. I'm working on finishing the data for my 3 trials then I am going to get a few more operators involved to do the other trials.

For the most part the GR&R report input looks pretty straight forward. Even if that doesn't pan out I've got other GR&R software that will do what I need.


Thanks again.

Kaji