iSixSigma

MSA Query

Viewing 45 posts - 1 through 45 (of 45 total)
  • Author
    Posts
  • #53973

    MBBinWI
    Participant

    Please, if you’re a “well seasoned” belt, hold off for awhile and let some of the newbies express themselves.

    I had a query recently where a question came up on whether to conduct an MSA or not. The spec lists the requirement to 2 decimal places. The measurement device reads to whole digits. What would you do? And why?

    0
    #192376

    Gary Cone
    Participant

    I know I am violating your request but I say you never worry about MSA until you verify you have respected the most basic rule with respect to test engineering.

    Does anyone know what the basic rule is?

    MBBinWI, you can answer but you have to wait 15 hours like I did.

    0
    #192388

    Ravi Prakash B
    Participant

    Hi I am not that experienced or Seasoned belt.
    The best possible solution would be replacing the measuring device. This involves costs.
    I would look for the tolerance within the spec. If it is large enough to use the Measuring device reading up to the whole numbers i would continue with the same measuring device, else change the measuring devices.

    Regarding MSA, I would go for MSA. It will give the confidence on the data. MSA would give me not only the capability of the device, it will also give me the operators capability. Both Repeat ability and reproduciblity are required for such small measurements.

    Gary- I am looking for the basic rule, it may that one solution which may save a lot of time involved in MSA

    0
    #192391

    Darth
    Participant

    Obviously, you would change the spec to a whole number. Just because the spec has two decimal places doesn’t mean that it is reasonable. Let’s first verify whether we really need the 2 decimal places. Changing the requirement is always easier and less costly than messing around with that darn MSA stuff anyway.

    0
    #192396

    MBBinWI
    Participant

    @Darth – Dear Brother Darth: While your response may be appropriate in some circumstances, let’s pretend that the folks who developed the spec actually had data that demonstrated that the measurement was appropriate ( @garyacone) to the decisions necessary and that the two decimal level of measurement granularity was necessary to achieve the desired output at the desired consistency.

    0
    #192529

    MBBinWI
    Participant

    bump

    I’m disappointed in the derth of responses. C’mon folks, this isn’t that hard.

    0
    #192534

    Stan Mikel
    Member

    I think we should pose the question to Janice.

    0
    #192540

    MBBinWI
    Participant

    @stanmikel – probably too busy submitting the charge card vouchers from those “free” certs everyone’s getting. ISSISSIPPI is a much better deal!

    0
    #192634

    MBBinWI
    Participant

    bump

    0
    #192723

    Chris Seider
    Participant

    Wouldn’t you want to add the width of the typical process and specifications? Or were you trying to be extra clever with your post?

    Anyways…..I thought we ALWAYS questioned the validity of an MS until proven otherwise with data. :)

    0
    #192761

    gomezadams
    Member

    Is the MSA required?
    Is that what we are measuring even critical?

    0
    #192764

    MBBinWI
    Participant

    @spazwhatsup – Again, let’s pretend that whatever is being measured is necessary/critical. I’m looking for the issue with conducting an MSA under the conditions identified, to whit: A measurement spec’d at 2 decimal places and a tool that reads to single whole digits. What would you do in this situation?

    0
    #192765

    MBBinWI
    Participant

    @cseider – I really wasn’t trying to be “extra” clever, and I didn’t really think that this was all that hard. However, it doesn’t seem like many folks are willing to put it out there and answer. @garyacone was trying to be a bit clever with his query (such as yours) about the “most basic rule”. Of course that is that the measurement is appropriate to making the decision needed to be made. This is just a bit deeper than that.

    0
    #192768

    gomezadams
    Member

    Discrimination is out the window,Repeatability and reproducibility may be an issue. To overcome,the measurement equipment (calibrated and qualified against a master) need have a min. of 3 decimal places.

    0
    #192772

    MBBinWI
    Participant

    @spazwhatsup – what would you do, and why?

    0
    #192782

    gomezadams
    Member

    Rather self explanatory.

    0
    #192783

    MBBinWI
    Participant

    @spazwhatsup – so ‘splain yourself.

    0
    #192785

    gomezadams
    Member

    Find the right tool.

    0
    #193074

    Jeff Barnhill
    Participant

    Hmmm,

    Resolution and the bucket of 10 rule maybe?

    0
    #193076

    MBBinWI
    Participant

    @EddieBarnhill – Eddie: If I had a prize, you’d get it.

    0
    #193088

    How many whole numbers are there between the specs? If the specs are -10.25 and +10.25 and the process exhibits sufficiently little variation and is reasonably centered, then I don’t see a problem. In other words, with those specs if you test 100 items and all end up with measurements of -2, -1, 0, 1, or 2 then I think you have a good idea of your capability.

    Also, as a statistician I feel compelled to say “bootstrapping” as it can be done at virtually no cost (less than an hour of time by one person on a computer) or process changes which may or may not be necessary. But I know advanced methods can cause eyes to glaze over so I’ll stop there.

    0
    #193100

    gomezadams
    Guest

    [email protected]
    Go for it Joel!
    I am always up to learn new stuff!

    0
    #193112

    MBBinWI
    Participant

    @joelatminitab – Sorry, Joel, but if I assume that the spec is correct (as somewhere in the thread I said we needed to do), then a measurement tool that only reads whole digits would be incapable of sufficient precision to demonstrate measurement system capability – regardless of spread of specs or process variation. And I’m cool with bootstrapping.

    0
    #193113

    Chris Seider
    Participant

    @MBBinWi

    You mean you care about %process or %study variation? :)

    0
    #193114

    @MBBinWI – Run these commands in Minitab:

    rand 1000 c1
    Let c2 = ROUND(C1,0)
    Capa C2 1;
    Lspec -10.25;
    Uspec 10.25;
    Pooled;
    AMR;
    UnBiased;
    OBiased;
    Toler 6;
    Within;
    Overall;
    CStat.

    This takes standard normal data and rounds it to the nearest whole number and then does a capability analysis with specs of -10.25 and 10.25. You have 0 PPM expected defective. Why would you waste time correcting the measurement system or even continuing on a project with that data?

    I’m not saying this will always be the situation, just that we should first decide whether or not there is really a problem before we try to fix it.

    0
    #193117

    Chris Seider
    Participant

    @joelatminitab
    You should remember that if you don’t have enough precision within the process variability, then you can’t have a chance to optimize the results (e.g. find relationship between the X’s and the Y) since you can’t tell with much precision where the Y actually is.

    All you know is you are giving in spec product but you can’t optimize the output.

    0
    #193118

    @cseider Definitely…although with that kind of capability I’m not sure why you’d be putting effort towards further optimization! But in any event this would be a small subset of the situations described so I don’t want to get too far from the original discussion…

    @mbbinwi does the “bucket of 10 rule” answer what you were hoping to hear or are you looking for something further?

    0
    #193122

    MBBinWI
    Participant

    @joelatminitab – one of my favorite BB “murder board” (or interview candidate) questions is to give the scenario: You have a process with a Cpk of 1.7 and a GRR of 45% study variation, what do you do? I won’t answer that here, as you never know who might be reading, but any BB worth their salt should be able to reason through this and provide a business answer, not a knee-jerk textbook answer.

    That said, yes, what I was looking to uncover is the fundamental understanding of precision. The problem with your analysis is that it focuses on capability, but before I accept any of the data, I want to have confidence in the data, and for that, I need an adequate level of precision (as well as stability). The scenario presented was a tool with whole digits resolution, and a spec with 2 decimal precision. This tool is totally inadequate to the task, regardless of how the actual performance is of the measured criterion. This points out that the spec level of precision can have profound implications to the production system. And also points out that the knee-jerk reaction to manufacturing process variation – tightening up the specs – often has unintended and unwanted consequences. Thats about as much as a forum post should have, so I’m going to say that this one is closed.

    0
    #193126

    @mbbinwi – Agreed, but you should post your BB question to see what kind of discussion it gets (although a good portion of the answer is already in this post). Would be good for GB’s and BB’s to read those considerations, unless you’re not wanting to give a hand to anyone who is applying with you…

    0
    #193144

    gomezadams
    Guest

    Not trying to be facetious,but it does depend on what question is asked and how it is asked. Some times questions are loaded with assumptions.

    I once had an engineer ask me to help him fit a regression line to a point.
    I then asked is that the only data you have? He conveyed yes. Just one x-y pair.
    My response was from which direction do you want me to come from,better yet,assuming an 11-dimensional space from string theory , which dimension and which of the infinite possible approach paths to dou want me to take,or do you want me to come at it from all dimensions. He grinned and replied,I need more data huh? I also need to explain the application variables? Yep!

    0
    #193161

    Mike Carnell
    Participant

    @cseider Is this the MSA question the guy with the “Since when was it acceptable ……..?” refering to?

    0
    #193162

    MBBinWI
    Participant

    @Mike-Carnell – missed you, Mike. You must have had real work to do!

    0
    #193165

    Chris Seider
    Participant

    @Mike-Carnell

    Wasn’t sure.

    0
    #193182

    Mike Carnell
    Participant

    @MBBinWI Getting a little hectic. The work is about the same but Consuelo’s twins are about to graduate HS. How do you keep “young adults” from trashing 12 years of irrelevant education in the last 4 weeks? I am actually understanding why some species eat their young.

    @cseider and @Darth Me either but it appears to have traumatized him. It looks like the Whitney MSA was only commented on by 2 so I assumed this one was it. I do like Darth’s answer – very pragmatic.

    0
    #193186

    MBBinWI
    Participant

    @Mike-Carnell – How are the trashing their education? My youngest also graduates HS this year. She will join her sister at Badger U. Couldn’t get any of the three to commit to a “real” education, like their old man. Woulda been a lot cheaper too. Oh, well.

    0
    #193187

    MBBinWI
    Participant

    @Mike-Carnell – by the way, how do you know you’re in Canada? The hometown baseball team is on the TV, the small one, and two non-local hockey games are on the big TV’s, eh!

    0
    #193195

    Mike Carnell
    Participant

    @MBBinWI Hey! Hey! Hey! Let’s be a little cautious there. My mother is Canadian and all the Cubans at the wedding now refer to Kenny as my Canadian brother.

    0
    #193226

    MBBinWI
    Participant

    @Mike-Carnell – well, I won’t comment on being linked to Kenny familialy, that’s your problem. However, the comment about the hockey games was more of a compliment. Any culture that recognizes hockey as superior to baseball is OK in my book. Remember, I was weaned on the North Stars (before they moved south to Dallas).

    0
    #193228

    Chris Seider
    Participant

    @MBBinWI
    Where did the Stars come from? I’m not a hockey fan…..I can’t see the puck on TV! I still remember when the Olympic coverage tried to show a blue streak of where the puck was traveling.

    I live within walking distance (a bit too long) to the American Airlines Center where they play.

    0
    #193232

    MBBinWI
    Participant

    @katiebarry – what happened to the report inappropriate posts button? See post above. Language indeed. (and doesn’t iSS have a naughty words filter?)

    And to you, hisco, you sniveling coward, we’ll talk about whatever we please. Kindly moderator will either censure us (or censor as the case may be) or not. You may choose to read or not.

    0
    #193235

    Mike Carnell
    Participant

    @Cseider It is a good thing there is some place interesting within walking distance because the interaction of your new hot rod and your leadfoot isn’t going to have you driving very long.

    Just to satisfy Hisco – I see this as a resolution issue. The gage isn’t capable of measuring to the resolution of the spec. Maybe an issue maybe not. If you have an extremely wide spec even though it is toleranced to 2 decimal points and the capability is still pretty good then leave it alone (whitney’s question).

    Just my opinion.

    0
    #193236

    Katie Barry
    Member

    @MBBinWI – I edited the post. This system doesn’t have a bad language filter, unfortunately.

    0
    #193239

    MBBinWI
    Participant

    @KatieBarry – seems like that might be a good enhancement, hint, hint.

    0
    #193240

    Mike Carnell
    Participant

    @MBBinWI I have never really understood that whole concept of bad language. They are words how bad can they be? You can say exactly the same thing in words that are considered acceptable but if you combine certain letters or sounds all of a sudden they become bad. Obviously it isn’t the meaning that is offensive because if we use the alternative language it is ok. It seems like a very odd concept.

    0
    #193242

    Chris Seider
    Participant

    Wow…..didn’t know a public forum with an occasional off topic comment would bother someone so much. Maybe they just had a super bad day–I’ve had those sometimes also.

    I get grumpy sometimes and I know I can measure it with some precision–see, I got back to an MSA. :)

    Cheers from down under.

    @MBBinWI Guess I’ll have to research about the Stars’ history via my fave crutch…Google.

    PS…..Ignore User exists for us all. I will say that a strong mind can withstand the view point of others.

    0
Viewing 45 posts - 1 through 45 (of 45 total)

You must be logged in to reply to this topic.