Gage RR – Resolution
Six Sigma – iSixSigma › Forums › Old Forums › General › Gage RR – Resolution
 This topic has 28 replies, 12 voices, and was last updated 13 years, 9 months ago by Chris Whittaker.

AuthorPosts

July 31, 2008 at 2:33 pm #50652
Hello there, let me ask someone of you about the help in matter of finding proper resolution of measurement device. If the tolerance is 0.3 mm expected resolution should be 0.03 (10%) or 0.006 (2%)? Need to know it because performed gage r&r shows number of distinct categories = 1, what should I do to change? V
0August 1, 2008 at 1:52 am #174453
Michael MeadParticipant@MichaelMead Include @MichaelMead in your post and this person will
be notified via email.You have already noted the answer. It is a little confusing to me why some people require a 10% level of precision and a 10% GR&R.
Anyway, I would guess that you do not have enough variation in your sample. You also do not have enough resolution in your gage to have different messurements for the same piece. So, either your gage and operators are too good or your process is too capable. In either case, to get a good GR&R you need to have some variation in the observed data.0August 1, 2008 at 6:46 am #174458Hi
What is the logic of the 1.41 factor in arriving at the no.of distinct categories ?
BB
0August 1, 2008 at 6:59 am #174459
Michael MeadParticipant@MichaelMead Include @MichaelMead in your post and this person will
be notified via email.Well, you got me there. I don’t know anything about the number of distinct categories. My guess is that 1.41 is the square root of 2.
OK, who really knows?0August 1, 2008 at 8:08 am #174460According to the AIAG, you need at least five distinct categories to have an adequate measuring system.
0August 1, 2008 at 12:08 pm #174465Hai Valdek.
If the resolution of the gage is not at least 10* smaller than your area of interest (Product Range or Customer Tolerance) it will be very difficult to achieve a gage r&R% result of 10% or lower.
So you have to think about:do I want to distinguish my products very well among each other (=> Product Range) OR do I want to measure my products well within the area that my customer finds agreeable (Customer Tolerance) How good does my gage r&R % result have to be (your choice)? * 10% because that’s the rule for excellent * 30% because that’s the rule for acceptable * 50% because that’s practical *100% because I can then at least measure roughly * …?
The combination of this gives you an indication how good the Resolution should be. And remember: to determine the Resolution you don’t need any measurements: it’s in the specs of your gage.#Distinct Categories is another parameter that measures the quality of your gage; there is a formula between gage r&R % and #DC (look it up in Minitab help if you want to no the function)
Michael: 1.4 = SQRT(1.96) and 1.96 is in the formula of CI of Mu (see Minitab Help)
Hope this helps0August 1, 2008 at 1:14 pm #174468
Ken FeldmanParticipant@Darth Include @Darth in your post and this person will
be notified via email.The number of distinct categories is a critical element in determining whether the measurement system has sufficient discrimination.
http://www.minitab.com/support/docs/Answers/Root2.pdf0August 1, 2008 at 1:25 pm #174469Thanks, but still the answer is not given for me, I perform the question in numbers – Range of results: 0,006 to 0,014, tolerance 00,3, number of samples 10, 2 operators, 3 trials each, cp and cpk for process data = more than 2, it looks like the resolution of device is to big, distinct number = 1, what to do to make satisfied customer? pls, do give me the answer, regards, V
0August 1, 2008 at 1:32 pm #174470It is the square root of 2 (not 1.96 as someone suggested).Read Wheeler’s work on measurement systems and go through his
derivation of Discrimination Ratio and you’ll understand where the 2
comes from.0August 1, 2008 at 2:08 pm #174473Hai Valdek:
Sorry, I can not decide for you. But I can explain how you could decide based on the info you already have. In my (maybe wrong) conclusion you don’t know exactly what all info is the tool gives you.
Assuming 0.0060.014 is the range of the measurents during the Gage r&R while the customer specs are 00.3 I conclude:
you use the default calculation of Minitab for gage r&R % you use the default product range and not the Tolerance optionyour products show small variation compared to customer wish (cp=2).So the result of the gage r&R tells you (with #DC as your decision parameter) that this gage system is not good enough to distinguish between the products! That does not necessarily mean that your customer will be unhappy.
If you use the Tolerance option you get the result of %Tol as extra column which tells you how good the system is to determine the product value within the specs. Using the Tolerance option gives an extra column in the Session Window (no recalculations!). This means also that the number of distinct categories does not change!! The #DC is always related to the range of the products used in the analysis and not to the Tolerance value (or Historic Sigma) that you filled in under options.
To estimate the gage r&R% from the tolerance option (=%Tol):the default gage r&R% * Range products/Tolerance. So I predict that you get a factor 8/300 smaller result (and the # DC will improve but I don’t know the exact formula and your #DC=1 could be a rounded up because Minitab has 1 as absolute minimum)
hint: the top value of the column with 6*SD (or 5.15*SD) represents the CI of 1 measurement (ie if I have 1 measurement I have a high confidence (99.7% or 99%) that the real product value is within an interval of that width around my measurement).
Hope this helps
0August 1, 2008 at 2:28 pm #174476Hai,
I stand corrected (again).
“Being right is more difficult than being wrong because there is only 1 right and so many wrongs’
Thanks0August 6, 2008 at 10:22 am #174605
Chris WhittakerParticipant@ChrisWhittaker Include @ChrisWhittaker in your post and this person will
be notified via email.Gage R and R is used to ensure that your measurements are correct or acceptable to measure the process , they are not a measurement of the process.
To correctly measure for Gage R and R you have 2 options …short term and long term measurement
Short term is 2 operators measuring 5 of the same products in exactly the same way twice each.
Long term is 3 operators measuring 10 of the same items in exactly the same way 10 times each
the key here is to use the SAME , measuring tools, people and set of IDENTICAL items for each measurement
This will ensure the discrimination, Linearity and stability of your measurement system is standardised.
If your process variability is under 10% the measurement is acceptable, 10 to 30 % the measurement may be acceptable over 30% and your measurements are no good.0August 6, 2008 at 10:34 am #174607Hai Chris,
I don’t understand your definition/distinction of long and short term gage r&R. The only difference I see is the amount of data used and consequently the Confidence Level of your conclusions.
I use the term Long Term Gage r&R for a gage r&R that is repeated every 3 months to check if the r&R values of the Gage System have changed (and Bias is checked via a control chart on Golden Products on day2day or week2week basis).
Also I don’t stress “The Same” like you do. I find it more important that the measurement is done in a representative way (like in real production life). Then you investigate an extra aspect of measurement influence on the data that will be used to Analyse. This aspect is seldom covered in the theory books.
Remi
0August 6, 2008 at 11:07 am #174608Chris does not know what he is talking about with his short term /
long term distinctions.Any study which represents the true measurement capability has to
include ALL who perform the measure and should be repeated
periodically including when the system is calibrated or repaired,
when there is a new operator or fixture, when an operator has not
performed the measure for some period of time.GR&R needs to be a requirement of the Quality System, not a Six
Sigma tool. A SS project should only verify that the calibration and
MSA studies are in place and adequate.By the way, the 10 to 30 % rule is crap. Is 30% adequate or not? I
know the answer, do you?0August 6, 2008 at 11:24 am #174610
Chris WhittakerParticipant@ChrisWhittaker Include @ChrisWhittaker in your post and this person will
be notified via email.It is not my definition or distinction , it is the AIAG definition /distinction of what constitutes a Gage R and R .
You are looking for reproducability and repeatability ( that is what the Rand R stands for) in order to do that you need to analyse in exactly the same way.
As the AIAG wrote the theory books on Gage R and R and you want to comply to that standard then it is likely that you want to follow their methodology.
You can do a MSA (measurement systems analysis) as often as you like but if you ignore the methodology as laid out by the AIAG then I would not call it a Gage R and rR…but if it works for you so be it.0August 6, 2008 at 12:09 pm #174615Go read the books Chris, you don’t know what you are talking about.They use the terms Short and Long. Not short and long term. They also did not write the books on the theory – that came out of
GM in 1962.0August 6, 2008 at 12:45 pm #174617
luke skywalkerParticipant@lukeskywalker Include @lukeskywalker in your post and this person will
be notified via email.Hey Darth,
Was in Indy yesterday and had a chance to see the “beautiful” Lilly building on the way back to the airport. Could have sworn I saw a big flag that showed “Lilly, a Vader Subsidiary”. You must have really done some significant work up that way.
0August 6, 2008 at 9:08 pm #174637
andrewmcParticipant@andrewmc Include @andrewmc in your post and this person will
be notified via email.August 7, 2008 at 4:12 am #174642
Chris WhittakerParticipant@ChrisWhittaker Include @ChrisWhittaker in your post and this person will
be notified via email.Thanks for the feedback Stan.
I will put it straight into my “lets pretend I GAF about what you think file”0August 7, 2008 at 11:32 am #174656
TaylorParticipant@ChadVader Include @ChadVader in your post and this person will
be notified via email.Chris with that type of closed mind attitude, you will never learn. Your answer was wrong. If your going to site books or guidelines, at least get it correct, or don’t post.
0August 7, 2008 at 1:48 pm #174661August 7, 2008 at 2:00 pm #174662
Chris WhittakerParticipant@ChrisWhittaker Include @ChrisWhittaker in your post and this person will
be notified via email.Outstanding Chad considering your first post was intercepted and deleted because it breached community guidelines.
I am interested to know what it was that you originally posted that was deemed sufficent to be removed from the site?
As a person who deems themselves worthy of being proficient to post to this site can you please enlighten all others as to what your original posting stated?
HHHHHMMM did not think so …all mouth and no trousers.0August 7, 2008 at 2:10 pm #174664
Chris WhittakerParticipant@ChrisWhittaker Include @ChrisWhittaker in your post and this person will
be notified via email.I hope that the LSS professionals like Stan and Chad that subscribe to this forum have been able to assist and clarify your question and give you the solutions to your question without letting their little ego’s get in the way so that can move on.
Good Luck!!0August 7, 2008 at 3:32 pm #174668
TaylorParticipant@ChadVader Include @ChadVader in your post and this person will
be notified via email.Chris
Your interpretation is wrong, Completely wrong. My original message simply said you fail, and I will post again.
I’ll state again, if your going to post answers, give correct ones.
A Gage R&R is conducted to determine whether excessive variability within the process exist in the measurement system. The guidelines in the AIAG manual, are simply that, guidelines, they are not hard rules. Samples to be measured should represent the process normal distribution. The original poster is having problem with the ability to determine one part from another. There is a couple reasons this happens, one is the sample of parts are not representitive of the process and/or the gage is not capable of measuring the difference in the samples. Which had already been answered, no input from you needed.
To further Stan’s point, you don’t have a clue when it comes to Short and Long term. I have never heard anyone say that. And you are also wrong about where the AIGA manual came from. It was adopted procedure developed by GM; John Deere also has a very similar manual. Again Stan is correct about how a Gage R&R should be used, so pull your head out of the sand hole its in and you might learn something
Now if you want to continue to pick fights there is other sites with individuals who will debate with you all day long, I suggest you go there.
Regards
Ps: Duct Tape is like the Force, it has a light side and dark side, and holds the universe together0August 8, 2008 at 2:12 am #174678
Michael MeadParticipant@MichaelMead Include @MichaelMead in your post and this person will
be notified via email.Chris,
Some folks on here know all the answers, or the answers are not worth knowing. But, on the other hand, when Chad criticized you for saying “Shortterm” and “Longterm” hmmm. You would do better calling them the short method and long method.
Keep up the good work, but you must watch your p’s and q’s here.0August 8, 2008 at 1:29 pm #174686
TaylorParticipant@ChadVader Include @ChadVader in your post and this person will
be notified via email.Michael
Attention to detail makes a big difference…………Agree?0August 11, 2008 at 1:09 am #174717
Michael MeadParticipant@MichaelMead Include @MichaelMead in your post and this person will
be notified via email.Yes it can sometimes. If I never made a miskate, I would agree completely. :)
0August 11, 2008 at 12:10 pm #174728Chris honey, it’s your ego that is fragile.
Your advice is wrong and your understanding of this tool is shallow. The AIAG manual is called Measurement System Analysis, not GR&R. What is your System? 2 operators, 3 operators, or ALL who perform the measurments? MSA properly applied pertains to all who make the measurements.0August 12, 2008 at 10:20 am #174777
Chris WhittakerParticipant@ChrisWhittaker Include @ChrisWhittaker in your post and this person will
be notified via email.Thanks stan for your input once again!
0 
AuthorPosts
The forum ‘General’ is closed to new topics and replies.