Why is my GRR% 86%! One part with two appraisers and three trials

R

Robert O'Brien

#1
Hello,

I am conducting a gage R&R study on a part with two appraisers and three trials. I have done the calculations a few times now, as well as a colleague of mine. I don’t believe there are any errors, so I can’t figure out why the R&R% is so high. I have attached our work – formatted almost identical to MSA manual 3rd edition pg. 113 & 114. I’m not asking anyone to do the work for me; I’m just looking for some suggestions on where I might have gone wrong as I feel I've exhausted all my options. I find it very hard to believe my R&R% is 86%!

Your suggestions are greatly appreciated.

Sincerely,
Rob O'Brien
 

Attachments

Elsmar Forum Sponsor

Caster

An Early Cover
Trusted Information Resource
#3
ndc

Robert O'Brien said:
Hello,
I can’t figure out why the R&R% is so high...........I find it very hard to believe my R&R% is 86%!Your suggestions are greatly appreciated.Sincerely,Rob O'Brien
Rob

You are right...no way is the R&R that bad. Just had a quick look at your data. This may be a case of inadequate resolution and the dreaded "ndc".

Please get a coffee (or beer) and read through the MSA forum. We all have lived through this problem.

It looks like you have only 3 or 4 possible readings from the gage. Is the data rounded off to the same number of sig figs as the tolerance? I made this error. You need to calaculate R&R with every figure the gage will produce and then even estimate the next digit if possible.

To prove this, just throw one extra random digit on your data and rerun the study. Hopefully you will get a better result.

Good luck, please ask again and post the raw data if you don't find your answer.
 

Howard Atkins

Forum Administrator
Staff member
Admin
#4
Rob,
This was the problem that I had with GRR and the way that it is presented in the MSA book.
The idea of using tolerance method is talked about briefly, see page 116
This changes the picture completely.
I have put a tolerance of 0.01 and the picture has changed completely but the ndc is not good and would show some problem.
One basic reason is that all the parts are so similar and thus there is a problem as statistics do not like "no change".
Try to find parts that are NOK and use them.
I have noticed that people do not like to use NOK parts but for GRR this is the best suggestion
 

Attachments

R

Rob Nix

#5
Howard is indeed correct. Put another way, since the parts you chose were so consistent the only other place for variation is repeatability and reproducibility. In this case the lion's share goes to repeatability variation, meaning that your system works well against variation from operator to operator (reproducibility). That's a good thing.

To determine whether the repeatability is acceptable you either have to widen the range of possible (and typical) part outcomes, or you have to include the drawing tolerances, as Howard stated.
 
R

Robert O'Brien

#6
Hello All,

Thank you very much for your responses. I am going to rerun the study. First, I want to clarify a few things.

The dimension in question in .130 +/- .005”, so a total tolerance of .010”. Since my engineering specification was to three decimal places, I rounded my measurements to three places regardless that the caliper I was using discriminates to four places. For my new study, I will use every figure as Caster suggested.

The next part I am still a little unclear on. I read the part on tolerance in the MSA manual on page 116 – 117, “…%EV, %AV, %GRR and %PV are calculated by substituting the value of tolerance divided by six in the denominator of the calculations in place of total variation”. Thank you Howard for the excel template, but I like to be able to do the calculations by hand. I am having trouble setting up the formula, and I can’t seem to find it in the MSA manual. I understand the tolerance over six part of it (.010/6 in my case), but how do I factor in EV, AV, GRR etc? Please let me know.

Also, regarding parts which are NOK (I believe this means not ok, as acronymfinder.com told me), I think the majority of parts are going to be close to the same in dimension. This is a short run injection molding process where the dimension is fixed by the mold cavity and the shrink rates are very similar.

I will do the study again and report back the results once I know the formula for the tolerance method.

Thank you and regards,
Rob O'Brien
 

Miner

Forum Moderator
Staff member
Admin
#7
Robert O'Brien said:
Hello,

I am conducting a gage R&R study on a part with two appraisers and three trials. I have done the calculations a few times now, as well as a colleague of mine. I don’t believe there are any errors, so I can’t figure out why the R&R% is so high. I have attached our work – formatted almost identical to MSA manual 3rd edition pg. 113 & 114. I’m not asking anyone to do the work for me; I’m just looking for some suggestions on where I might have gone wrong as I feel I've exhausted all my options. I find it very hard to believe my R&R% is 86%!

Your suggestions are greatly appreciated.

Sincerely,
Rob O'Brien
I double checked your calculations, and you came up with the correct result.

The problem is that the gauge repeatability is twice the process variation. Now there are two possible explanations for this:
  • The repeatability of the gauge really is too high for use as a process control (i.e., SPC) gauge. Depending on your tolerances, it may or may not be acceptable for part inspection purposes. A gauge may be perfectly acceptable for classifying a part in or out of tolerance, but not acceptable for process control purposes.
  • The process variation is understated because the parts do not represent the full variability of the process. Make sure that the parts do represent the full spread of process variation.

A possible cause for high repeatability not caused by the gauge is within part variation. For example, an outside diameter with variation in form (e.g., lobes, roundess, taper, etc.) will show up as a repeatability problem with the gauge. You can sometimes compensate for this by fixturing the part to consistently present it to the gauge in a repeatable orientation. Another cause is part deformation during measurement.

See pages 15, 52-53 & 123 of the MSA manual for more potential causes.

Based on your additional information, the repeatability is not acceptable for product inspection either. The gauge performance curve in the attached file illustrates this. See pages 17-18, 169-174 in the MSA manual
 

Attachments

Last edited:

Caster

An Early Cover
Trusted Information Resource
#8
Costly!

Robert O'Brien said:
Also, regarding parts which are NOK (I believe this means not ok, as acronymfinder.com told me), I think the majority of parts are going to be close to the same in dimension. This is a short run injection molding process where the dimension is fixed by the mold cavity and the shrink rates are very similar. Rob O'Brien
We are a die caster so we have the same (good!) problem. 10 parts grabbed in a row are very, very similar. Unfortunately this creates a bad R&R number!

So we end up machining a spread of 10 parts from just below to just above the customers spec. If you read the MSA with this statement in mind, that is what it says to do.

This is really costly, both in time and material, but it works.

This plus no rounding off made our CMM into a good gage again!
 

bpritts

Involved - Posts
#9
Miner and Caster are absolutely on the right track. It sounds very much like your process is more repeatable than your gage.

We have also had the same problem in a different area: measuring torques. We have automoted torque controllers that are capable to +/- 1 foot pounds with a consistent joint. Then we use torque wrenches to audit the torques. Well, we couldn't get a torque wrench reading to repeat at +/- 5 foot pounds! So our measurement capability was worse than our process capability.

We "solved" this by redefining the problem.

-- Our tolerance was +/- 10 foot pounds
-- We convinced ourselves, and then our SQA, to treat the torque controller itself as the measurement device
-- We also were able to obtain some repeatability data from the torque controller supplier, who helped us demonstrate the accuracy, precision, and repeatability of the automated measurement process
-- We still use the torque wrench as an occasional audit, just in case the torque controller totally goes haywire.

Ask yourself why the GRR matters.

== #1 -- to understand your risks in making wrong inspection decisions, as
they might affect the disposition of the individual part you are checking
== #2 -- to further understand the risk of making wrong process control decisions, if the part you are checking is only a sample of the process
==#3 -- to use the measurement results for continuous improvement of the process

With a poor R & R, your gage may be adequate for the #1 or #2 purpose,
but probably not for #3. In injection molding, the tool isn't going to vary much once it's warmed up. Different lots of material might. Putting the tool on a different molding machine might. But ask yourself, too... once the tool is built and the process established, are you really going to use the gage to control the process? Or, will you use the machine (barrel temps, time, pressure, etc) to control the process?

If you are trying to improve the process, then you might consider what the
accuracy, stability, and repeatability of the process control measurement devices in the molding machine are. They are what's actually driving the part quality.

Sorry if the rant got too long...

Regards,
Brad
 

Howard Atkins

Forum Administrator
Staff member
Admin
#10
Robert O'Brien said:
Hello All,

The next part I am still a little unclear on. I read the part on tolerance in the MSA manual on page 116 – 117, “…%EV, %AV, %GRR and %PV are calculated by substituting the value of tolerance divided by six in the denominator of the calculations in place of total variation”. Thank you Howard for the excel template, but I like to be able to do the calculations by hand. I am having trouble setting up the formula, and I can’t seem to find it in the MSA manual. I understand the tolerance over six part of it (.010/6 in my case), but how do I factor in EV, AV, GRR etc? Please let me know.

Also, regarding parts which are NOK (I believe this means not ok, as acronymfinder.com told me), I think the majority of parts are going to be close to the same in dimension. This is a short run injection molding process where the dimension is fixed by the mold cavity and the shrink rates are very similar.

I will do the study again and report back the results once I know the formula for the tolerance method.

Thank you and regards,
Rob O'Brien
1 Sorry about the NOK, I thought it was universal.

2 The meaning is every where that TV appears then use the tolerance.
Hope this helps
 
Thread starter Similar threads Forum Replies Date
T GRR based on part tolerance or process variation. Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 7
J Calculating part variation from historic data for GRR study Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 1
P New Part - How to choose what needs GRR (Gage R&R) Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 5
A Cg, Cgk and GRR of a CMM Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 5
L GRR for a tolerance that has changed Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 6
K %GRR was between 10-30% so we have to have a "backup plan" per auditor IATF 16949 - Automotive Quality Systems Standard 15
V Minitab GRR crossed Xbar and r method macro needed Using Minitab Software 9
A Minitab 18: How to automate GRR Summary Table save Reliability Analysis - Predictions, Testing and Standards 1
J Question on very low NDC number with tolerable GRR ratio's Capability, Accuracy and Stability - Processes, Machines, etc. 7
C Evaluating GRR Scores %Tol vs %TV Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 9
S GRR on ID with set of pin gauges Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 4
Q What does % study variation of 20% mean in GRR? Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 17
C %GRR <10% not possible. How to get approval by auditor? Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 15
V GRR Study on a Finished Product - Batch Manufacturing Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 1
S Script of R software to conduct Nested GRR for Destructive Testing Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 0
A Hypothesis test for Bias, not for GRR. WHY ?? Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 1
B How to perform a MSA - GRR (Gage R&R) A&R/ANOVA Vs GUM/VDA Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 1
S GRR (Gage Repeatability and Reproducibility) for Material Strength Tester Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 6
K The basics of GRR (Gage R&R) Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 2
A Need MSA %GRR Interpretation Advice Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 2
M GRR approach for off line Measuring Equipment Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 2
P Does anyone have the GRR Formula? Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 1
5 Reasonable GRR result for a 0-1" micrometer with a total tolerance of .002" Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 16
T MSA - Should I compare my GRR to tolerance or to my sample? Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 16
D What the impact is of a high %GRR/TV Capability, Accuracy and Stability - Processes, Machines, etc. 6
V Sample Range for a valid GRR using ANOVA in Minitab Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 2
F GRR (Gage R&R) with Different Scales Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 10
R New to MSA - I was told to use ANOVA for the GRR's Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 2
J Type 1 GRR (Gage R&R) on a One Sided Specification Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 6
J Analysis of GRR studies, MSA 4th Ed, Chapter III section B - Rigid Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 1
S Strange PT Ratio on an Excel spreadsheet for variable GRR ANOVA Method Excel .xls Spreadsheet Templates and Tools 2
A Customer Demands (MSA=Bias + GRR/2). Can somebody explain what it means? Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 6
R ndc=0 and GRR=14% ??? Interpretation - basic question Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 9
I Minitab ANOVA Gage R&R Error (vs. EXCEL ANOVA GRR Calculation) Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 5
R MSA - GRR (Gage R&R) Requirement for CMM Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 21
Rameshwar25 What is the significance of Spread in Calculating GRR (Gage R&R)? Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 1
Rameshwar25 GRR (Gage R&R) Nested Anova Method Example Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 40
J CMM MSA - Can you use 1 GRR to satisfy requirement for all Key Features? Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 11
M Minitab Type 1 Repeatability & GRR% of Tolerance Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 1
M Cp calculation on page 193 of the MSA 3rd edition May 2003 with %GRR Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 1
T When is GRR (Gage R&R) required? AIAG MSA Manual Version 4 Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 19
H GRR (Gage R&R) Study for Digital Temperature Gauge Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 4
L %GRR/TV or /TW for GRR of One Sided Specification - Gage R&R Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 6
S Gage R&R (GRR) Study - Please comment on the results Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 7
J Attribute GRR (Gage R&R) - How to get Kappa Software Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 2
S p-Value(s) for Anova (Crossed) GRR (Gage R&R) Studies Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 7
D GRR (Gage R&R) procedure for Mitutoyo Digital Micrometer wanted Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 1
T % Gage R&R Calculation - 100x(GRR/ (Spec/6)) or 100x(GRR/TV) Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 20
O The Basis of %GRR (% Gage R&R) and other Statistics - Minitab Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 21
A Effect of %GRR (% Gage R&R) on Cpk Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 13

Similar threads

Top Bottom