iSixSigma

Sigma level calculation

Six Sigma – iSixSigma Forums Old Forums General Sigma level calculation

Viewing 62 posts - 1 through 62 (of 62 total)
  • Author
    Posts
  • #34741

    rathinakumar
    Participant

    We have a project of reducing errors in engineering drawings. Each drawing is unique, in terms of complexity; in that it has a different number of opportunities, ranging from, say from 16 to 95. For a sample of 42 drawings (collected in a 3 month period)having a total of 10 defects and 1609 opportunities , we have the following data, (where dpu is errors per drawing, noted at the time of review) calculated on Minitab:

    Item                   Mean         

    Dpu                  0.238         
    Opportunities 38.31         
    DPMO             3634(avg. of DPMO of each dwg.)

    Sigma Level 5.743(average of sigma level of each drawing)

    It is also possible to calculate Sigma level in two other ways:

    1000000xMean dpu/Mean Opportunities = 6215

    or
    1000000xTotal defects/Total Opportunities = 6215
    Then use Sigma calculator to arrive at Sigma = 3.99

    Sigma level corresponding to mean DPMO of 3634 = 4.18

    Can the experts advise me which approach is technically correct? My understanding is 3.99 is correct.
    Thank you in advance
    Raja

    0
    #96169

    A.S.
    Participant

    In standard books formaula recommended is as follows  (to calculate DPMO—sigma level) :
    (No of Defects / nos produced)/No of Opportunities * 10000
    But some statisticians are strongly criticising this formaula since if you consider more no of oppotunities it will dilute your basic defect rate calculation.
    Request some one to add comments  please
     
     
     

    0
    #96177

    Wayne
    Member

    No one can argue that increasing the number of opportunities would result in a reduction in the defect rate. This is why its important that people are in agreement on the definition of a “defect” as it relates to their product or process. Its important to spend the time upfront in making the determination on what constitutes a “pass” or “fail” in order to remove any ambiguity.
    Unless something drastically changes or new patterns develop, the definition of a defect for that product/process shouldn’t changes otherwise any trending can become skewed.
    Just my $0.02.
    Wayne

    0
    #96196

    grover
    Participant

    I am also confused about the concept of “opportunities” in six sigma calculations. What is the criteria of taking opportunity and how do we select it. For instance, if a product has a possibility (opportunity) of get failed due to 15 reasons, then can we say that we have 15 opportunities. Also, by increasing # of opportunities in formula of DPMO (DPMO=DPU/OPPORTUNITY * 1000000) we gonna get less DPMO which means comparitively larger sigma value. Is not it a very strange concept, in a way that here we are actually talking about the probability of DPMO in a specific situation. Now, we are saying that as # of opportunities increase, DPMO decrease, meaning sigma increase. In other words we are saying that the probability of getting wrong parts get decreased (although more opportunities means that chance of failure is now increased!!) please help!!

    0
    #96197

    Savage
    Participant

    Don’t make this harder than it needs to be.  Opportunities are those things that MUST be right – not all of the things that COULD go wrong.
     
     

    0
    #96198

    Ang
    Participant

    I don’t believe that ‘reasons for failure’ should be confused with opportunities for failure.
    Reasons describe, opportunities define. That is to say, reasons are used to identify a root cause for a failure whereas an opportunity identifies a point in a process at which a defect (as pre-defined) can occur. There may be many different reasons for a process step failing but there is still only one failure).
    If you decide that reasons for failure are defects then it follows that it may be possible to have a lousy process that ‘looks’ good (in terms of DPMO and Sigma) even when it fails more often than not (but only one reason, of 50 possible, at a time is the cause of failure).
    The example is a little over the top but I am sure you get my point.
     

    0
    #96202

    Savage
    Participant

    Don’t make this harder than it needs to be.  Opportunities are those things that MUST be right – not all of the things that COULD go wrong.  Period.

    0
    #96216

    grover
    Participant

    Thanks Matt for calrification. But there is one more confusion in my mind which i mentioned in my previous post. This is about DMPO vs OPPORTUNITY. We know that as # of opportunity increases, DPMO value decreases (according to our foumula)  which in turn gives us comparatively larger sigma value. Now here is my confusion ” Increasing Opp. (or possibility of defects) by our formula, is giving us larger sigma value which is really saying that probability of getting defects is decreased. Is not it contrary to the fact that if we are saying that possiblity of getting defective parts is increased (Opportunity), then the probability of getting defective parts should be increased, not decreased!” or i am missing some basic and very important concept.

    0
    #96217

    Manjanath Nayak
    Participant

    Hi Naveen,
    There is some basic thing we need to keep in mind while defineing the Defects, Uints and opportunities.
    Defects  : Which is not confirming to the customer CTQ or product standard
    Opportunity : Any characterstic of the product/ service as viewed by the customer.
    In the case of opportunity we need to consider only those charactersics which are percieved by Customer.
    Eg. No of resistors in the Mother Board of a TV . may not be very relevant to customer. May be Picture Resolution, Functioning of the Remote, etc. may be the opportunites for the Final Inspectyion of the TV.
    In this context as we define the opportunities we ned to develop the operational definitions for those Opportunities to measure.
    As opportunities increses and your measurement points also should increse, to be inline woth opps.
    In case of TV Eg. if you are considering all the process steps as Opps. and only measuring the Key CTQs as mentioned earlier may not be OK. In such case only you will onbserver the phenomenon as you have staed in your problem.
    Please revert back in case of any further queries
     
     
     

    0
    #96256

    grover
    Participant

    Thanks Manjanath nayak for your brief description about opportunity. In case of TV example suppose TV resolution, and remote effeciency is important to customer. In that case we have two opportunities. Based on this info. lets calculate DPMO. for the time being lets it founds to be 1350 which corresponds to 3-sigma. Now suppose we came to know that TV physical appearance is also important to customer. In that case we have three opportunities. Again calculate DPMO. This time it should come lesser than the previous (based on DPMO formula). For time being lets suppose it comes out to be 577 which corresponds to 3.25 sigma. Sigma, calculated this time is greater than the previous calculation, which clearly says that probability of getting defective part (TV) in first case (less opp.) is more than the second case (more opp.). Is not it contrary to the fact that now we have more opportunity (possibility) for defective item( second case) so probability of defective should be greater than the first case (less possibility)!!  

    0
    #96277

    grover
    Participant

    Thanks Manjanath nayak for your brief description about opportunity. In case of TV example suppose TV resolution, and remote effeciency is important to customer. In that case we have two opportunities. Based on this info. lets calculate DPMO. for the time being lets it founds to be 1350 which corresponds to 3-sigma. Now suppose we came to know that TV physical appearance is also important to customer. In that case we have three opportunities. Again calculate DPMO. This time it should come lesser than the previous (based on DPMO formula). For time being lets suppose it comes out to be 577 which corresponds to 3.25 sigma. Sigma, calculated this time is greater than the previous calculation, which clearly says that probability of getting defective part (TV) in first case (less opp.) is more than the second case (more opp.). Is not it contrary to the fact that now we have more opportunity (possibility) for defective item( second case) so probability of defective should be greater than the first case (less possibility)!!  

    0
    #96280

    A.S.
    Participant

    Even with this TV example i am not clear where we have to stop the counting of number of oppotunities.Even a non techinical customer can count no of  opportunities for a defective TV failure such as Picture tube failure,Speaker failure,Remote failure etc.Still i felt that considering the no of oppotunities will show the increased sigma level.
    Manjanath nayak —–Let me have some clarity on this please
     
     
     

    0
    #96282

    Manjanath Nayak
    Participant

    Hi Naveen and all,
    The Naveen you are right. But one fundamental mistake in our discussion thread is we are comparing two different situations. Here once you change your Opportunities you need to recalculate the Sigma Value of the previous Sigma. Here we are not capturing the defects occurred due to the new opportunity. We are keeping the Defect level at earlier level and debating the issue. If you have the defects due to new opportunity then the statement and the calculations will match.
    Eg. TV Case: When the physical appearance is added as opportunities we need to capture the defects related to it. Then you can see that defects are going high and the probability of making defect also going high.
     
    I think I have made the point clear – Please revert back

    0
    #96284

    A.S.
    Participant

    Not yet clear please.
    For example  out of 100 nos produced ,if 4 nos of TVs are not showing the picture.It will become 40000 ppm.If i consider 10 opportunities,then it will become 4000DPMO.Sigma level for this is  4.15 sigma.It is difficult to accept this 4% rejection as 4.15 sigma level.
    Please clarify
     
     
     

    0
    #96310

    Manjanath Nayak
    Participant

    Hi,
    I think this discussion is not leading any where. I think there is basic gap in understanding the yield, DPMO, and Sigma Calculation. I suggest we need to revert back to our study center to revisit the learnings what we have undergone during SIx Sigma Training and Learn the difference between DPMO and Yield and SIgma Levels.
    When the basics are clear we can clarify those parts

    0
    #96323

    grover
    Participant

    Dear Manjanath, are you dissappointed with the situation. I think in addition to look at the basic stuff we can ask someone in this forum to put light on our issue also.
    Thanks anways for your views on my problem and i dont think that i am at the same place where i was when i initially asked this question. I learned something from you and others who participated in this issue.
    Regards,
    Naveen

    0
    #96336

    Gabriel
    Participant

    Naveen,
    Of course you can sak questions in this forum. That’s why this forum is here in first place.
    However, this issue was discussed a lot of times in this forum, and by doing a search you will find tons of threads on this.
    However, most of these threads are, at the end, non conlusive since there are different opinions that never reach to an agreement. I’ve participated in several of this threads, and when I saw this thread I knew that would happen again. That’s why I didn’t participate up to now.
    Just to give you a hint: Sigma level is not a measure of quality from a customer satisfaction point of view because it is a relative measurement: Quality vs complexity. You are measuring DPMO which is defects per opportunity (not defectives per unit). Then, a complex product/process (lot of opportunities) can have a better sigma level than a simple product/process with fewer defectives but a lot fewer opportunities too. However, sigma level based on DPMO is a good measure of the quality from a “producibility” point of view, since it is more difficult to produce less defect when you have more opportunities. It’s like a measure of how hard would it be to improve the prcess any further.
    If you want to measure the quality from a customer’s point of view, then you must think of only one opportunity (and potential defect) per unit. A product is either Ok or it is not. In that way, DPMO defaults to PPM.

    0
    #96345

    grover
    Participant

    Gabriel, you said “Then, a complex product/process (lot of opportunities) can have a better sigma level than a simple product/process with fewer defectives but a lot fewer opportunities too.”
    Now in my previous threads, I explained my problem that suppose we are talking about a production line and already calculated #s of defective products. Suppose it is found to be 100 out of 10000 pieces. Suppose that this specific part has three systems in it (for instance, this part is being used in the jet air craft and the function of this part is to control the velocity, pressure and temperature of the fluid passing through it). Here this part has 3 opportunities to be declared defective by the customer. Now, lets calculate DPMO= [100 / 10000 *3] * 1000000 = 3333.33, say 3333. Sigma level for this much amount is 2.715 (approx.). Here we are saying that out of one million there is a probability that 3333 parts would be defective.
    Now Suppose that this part has only two function, instead of three. Here, we have two opportunities for this part to be declared defective by the customer. DPMO in this scenerio would be= [100/10000*2] * 1000000 = 5000. Corresponding sigma would be = 2.58 (approx.). We clearly see over here that by decreasing “Opportunity” gives us “lesser” sigma value, which also means that probability of getting defective parts is increased.
    So, according to your phrase that complex systems with lot of opportunities (in our case 3 opp.)can have a better sigma than a system with less opportuntiy (in our case 2 opp.). But Gabriel, here is my confusion. lets forget about sigma level and see that if we have more opportunity of failures in any product / process, does not it mean that we have more chances of failures as compared to a system which has less opportunity of failure!!. But our previous calculations are telling us something just reverse!!!!
    Thanks and regards,  

    0
    #96349

    DaveS
    Participant

    Naveen –
     
    In a previous post you correctly identified (DPMO=DPU/OPPORTUNITY * 1000000)
    Pleasee use it and DIVIDE DPU by opportunities.
    Your confusion will be much less.

    0
    #96351

    Gabriel
    Participant

    Naveen, you are confusing DEFECTS with DEFECTIVES. A defect is a requirement not met. A defective is a part that does not meet 1 or more requirements (i.e. has 1 or more defects). But further more and even more important, you seem to be confisung OPPORTUNITY with PART. Let’s see:
    “DPMO= [100 / 10000 *3] * 1000000 = […] say 3333. […] Here we are saying that out of one million there is a probability that 3333 parts would be defective.”
    That’s plain wrong. Here you are saying that out of 10000 there are 100 defectives (or, what’s the same ratio, out of 1 million there are 10000 defectives, thats 10000 PPM, or defectives Parts Per Million of parts). You already said that yourself: “#s of defective products […] is found to be 100 out of 10000 pieces.”
    However, it is also true that in this case (with 3 opportunities per part) “DPMO = […] say 3333”, as you said too.
    What’s then DPMO=3333? Let’s start with what it is not: It is not that “out of one million [parts] […] 3333 parts would be defective”, as you said.
    It is that “out of 1 million of opportunities there would be 3333 defects” Do you see how OPPORTUNITES replaced PARTS and DEFECTS replaced DEFECTIVES?
    SInce each part has 3 opportunities, one million opportunities are found in 1,000,000 / 3 = 333,333 parts. So in 333,333 (i.e. 1 million of opportunities) parts you have 3,333 defects (not defectives). In 333,333 parts, 3,333 defects can be found in different ways. For example, you can have 1,111 defective parts with 3 defects each (1,111 * 3 =3,333 defects), and the rest of the parts without defects. In that case you would have 1,111 defectives out of 333,333 parts, what’s 3,333 PPM. Also you can have 3,333 parts with 1 defect each (3,333 * 1 = 3,333 defects) and than you will have 3,333 defectives out of 333,333 parts, that’s 10.000 PPM (or, like in your original statement, 100 defectives out of 10,000 parts).
    So, as you see, the same DPMO can correspond to different PPMs. Since with the classical definition the “sigma level” is a function of DPMO (not PPM), that also means that the same sigma value can mean different PPMs. And that’s only changing the way the defects are distributed among the defectives. We didn’t touch the number of opportunities per part yet! Let’s do it now:
    “…here is my confusion. […] if we have more opportunity of failures in any product / process, does not it mean that we have more chances of failures as compared to a system which has less opportunity of failure!!. But our previous calculations are telling us something just reverse!!!!”
    Yes, that can happen. If you increase the number of opportunities, then you can find things that were not considered defects before, but now are. If you check a sample with a criteria and then check it again with a new criteria where new potential defects (opportunities) are taken into account, then the number of defects can only increase. If that happens, most probably the number of defectives will also increase (unless the new defects are only found in parts that were already deffective because of other defects that were detectable with the previous criteria). So it is true, as you said, “we have more chances of failures as compared to a system which has less opportunity of failure”, and then the defectives rate (PPM) will increase, you will have more defectives.
    What is wrong is what you said next: “But our previous calculations are telling us something just reverse!!!!”. No. What you saw in your previous calculations whas that DPMO and “sigma level” changed in the other way. Not the defectives rate.
    As said, that can happen: If you increase the opportunities per part then the number of defects will increase, and the number of defectives will increase too (i.e. they will worsen). But if the increase in the number of defects is proportionally smaller than the increase in the number of opportunities, then the DPMO will decrease and sigma level will increase (i.e. they will improve).
    Your fault is that you are trying to get some information from Sigma Level and DPMO that they cannot give: Sigma Level and DPMO are NOT indicators of DEFECTIVES RATE! If you want to measure the temperature, use a thermometer, not a barometer! If you want to measure the defectives rate, use PPM, not DPMO!
    Since the pressure is a function of the temperature (among other things), a barometer could be used to detect a change in temperature ONLY IF ALL OTHER THINGS (such as volume, quantity of gas, etc) REMAIN UNCHANGED. In the same way, Sigma Level and DPMO can be  used to detect a change in the defectives rate ONLY IF ALL OTHER THINGS (such as number of opportunities) REMAIN UNCHANGED.
    So phrases like “by decreasing “Opportunity” gives us “lesser” sigma value, which also means that probability of getting defective parts is increased” are wrong. The change in sigma value means nothing about the number of defective parts, specially if the number of opportunities changed.
    One final point: You say that the increase in the number of opportunities means more chances of failure (it does). But in your examples along this thread you increased and decreased the number of opportunities, but the number of defecives reamined allways unchanged. That’s not fair, but that was clearly not the source of your confusion.
    What do you think now?

    0
    #96354

    Hywel
    Participant

    The opportunities should be chosen so that they are something that is important to the customer and also they should be characteristics of the product or service where you can expect to get defects in the process before it is improved (e.g. process capability <=1)

    0
    #96360

    Mikel
    Member

    You don’t understand opportunities.

    0
    #96447

    grover
    Participant

    Gabriel, First of all sorry for replying so late. I was out of town for a day. Thanks for such a nice explaination. You are right that i was totally confusing “defects” with “defectives” and “parts” with “opportunities”. One more question though, ” what is the criteria of selecting appropriate, right and to-the-point opportunity/ies right the first time of the DMAIC process”, since it directly effects DPMO and hence the Sigma value. For instance, in the service industry where customers are served by bunch of employees of the company, how do we exactly define opportunity/ies there.
    Regards 

    0
    #96457

    Gabriel
    Participant

    About the how the opportunities count should be done, I am afraid I am not the right person to help with that. I’m sure other persons in this forum will. But right now there is an cative thread about htis subject: “DPMO Calculations”. Check there.

    0
    #96530

    Stacy
    Member

    Gabriel, I am new in six-sigma world, have been reading the threads under ” Sigma level calculation”. Can you explain me why we transform DPMO value to the corresponding six sigma value. Can’t we tranform PPM to corresponding six-sigma value??
    Will appreciate your comments. 

    0
    #96553

    Muhammad Khan
    Participant

    Stacy, To my understanding, DPMO clearly defines the # of “defects” in a process which is really our concern. PPM is used to quantify # of “Defectives”in a process which is not really telling us “how many defects are there in the defective parts”. Hence DPMO is really translate the bottom line of any process and hence used to find sigma level of the process.
    Futher comments are highly appreciated
    Regards
     

    0
    #96554

    Gabriel
    Participant

    I am not the right question to aswer this, but this is my opinion, or what I could understand about the subject:
    DPMO is used to take into account the complexity in the product/process, so the DPMO and Sigma Level are theorically comparable within an industry.
    You can calculate the Sigma Level based on PPM. This indicator would be much more realted with customer satisfaction. After all, the customer is not happier because the product that failed had a lot of opportunities!
    The sigma level calculated with DPMO is supposed to be much more related with process performance.

    0
    #96556

    Fer
    Participant

    Using data you provided I get these  results:
    DPMO = 6215
    Z LT = 2.50
     
     

    0
    #96584

    Navine
    Participant

    Gabriel, could you please explain your phrase “
    You can calculate the Sigma Level based on PPM. This indicator would be much more realted with customer satisfaction”  with the same example in which we have same # of DPMO (3333)but different #s of PPM (3333 vs 10000). How do we translate it accroding to customer satisfaction.
    Regards

    0
    #96653

    Gabriel
    Participant

    There are no “different #s of PPM”. By now you should know that, in your previous examples, there was a missunderstanding between DPMO and PPM. 10000 was the PPM (100 defectives out of 10000 parts) and 3333 was the DPMO (100 defects (one in each defective) out of 30000 opportunities (3 opportunities for each part)).
    There can be, however, different ways to count opportunities, and then different DPMO values for the same defects found on the same parts.
    From the customer point of view, it is ussially said that there is only one opportunity. If your new TV set does not work properly, you don’t care about how many resistors does the TV has inside or how many opperations were needed to solder all components or to assembly the set. There is only one opportunity: When you press “ON”, the TV must turn on and work properly.
    From the process poiont of view, however, it seems unfair to treat different complexities with the same rigor. Since a TV set has so many components and so many operations (compared for example against making a nail), it is expectable that there will be more defects too. So defects are compared to opportunities as a mean to level the scale.
    Returning to the customer point of view, since 1 part = 1 opportunity and one defect = 1 defective (because the defect is simply “the thing des not work”), the DPMO equals to PPM. So you can count PPM, take them as “customer focused” DPMO, input that value in a sigma level table or calculator, and get the “customer’s point of view” sigma level.
    Or something like that….

    0
    #96682

    Muhammad Khan
    Participant

    Gabriel, I read this thread very late and after reading all the matter I have one question from you if you please help me to clarify it. You said.
    “However, it is also true that in this case (with 3 opportunities per part) “DPMO = […] say 3333″, as you said too”.
    As we know that DPMO is basically deals with “defects” not  “defectives” and Naveen starts his problem by talking about ” defectives” not “defects”. So, technically we can not calculate DPMO as per his problem until and unless we suppose that either each of 100 defective parts (out of 10000) contains one defect or every part in those 100 parts contains different #/#s of defects which equals to 100 in total.
    What you think, am i correct? 

    0
    #96683

    Gabriel
    Participant

    Yes, you are right. Any number of defects (100 or more) can be found in 100 defectives. And that number of defcts should be used to calculate DPMO, not the number of defectives.
    Just one point: Since every defective must have at least one or more defects, the only way to have 100 defects in 100 defectives is with 1 defect per defective, so your “or every part in those 100 parts contains different #/#s of defects which equals to 100 in total” does not apply (unless you can imagine a defective with less than one defect)

    0
    #96718

    Muhammad Khan
    Participant

    Gabriel, Yes, you are right and thanks for clarification. One more issue to discuss. I really dont understand the significance of finding the sigma-level from the customer point-of-view becasue it will not really translate what is actually going on with the process and hence the bottom line.
    One suggestion from you: I am going to do a six sigma project in the service industry (in the ticketing department of local aeroplane company). The customers have different kind of complains against the poor service in the reservation department of the company. I was just thinking to take a comprehensive interview of affected customers in order to get the clear picture of what kind of complains they have. Based on the interview, I will then draw a pareto chart and see which complain has got the most weightage so that i can put my whole attention to it.
    After doing all this exercise, I will then start digging all the steps included in the reservation process and see what are really the key steps (opportunities) which can be revised and then reviewed. There is one question in my mind though. Finding the #/#s of opportunities is really a subjective issue, may be i am saying that we really have ten opportunities in the whole process but someone says that he sees more than that!!. In this situation how can we justify the significance of sigma level which is found based on DPMO which depends on # of opportunity/ies in the process.
    I will appreciate your comments .. Regards

    0
    #96799

    Muhammad Khan
    Participant

    Gabriel, Yes, you are right and thanks for clarification. One more issue to discuss. I really dont understand the significance of finding the sigma-level from the customer point-of-view becasue it will not really translate what is actually going on with the process and hence the bottom line.
    One suggestion from you: I am going to do a six sigma project in the service industry (in the ticketing department of local aeroplane company). The customers have different kind of complains against the poor service in the reservation department of the company. I was just thinking to take a comprehensive interview of affected customers in order to get the clear picture of what kind of complains they have. Based on the interview, I will then draw a pareto chart and see which complain has got the most weightage so that i can put my whole attention to it.
    After doing all this exercise, I will then start digging all the steps included in the reservation process and see what are really the key steps (opportunities) which can be revised and then reviewed. There is one question in my mind though. Finding the #/#s of opportunities is really a subjective issue, may be i am saying that we really have ten opportunities in the whole process but someone says that he sees more than that!!. In this situation how can we justify the significance of sigma level which is found based on DPMO which depends on # of opportunity/ies in the process.
    I will appreciate your comments .. Regards

    0
    #96841

    Muhammad Khan
    Participant

    Sorry Gabriel, but this is the third time i am sending this message and i am sure you have not got time to take a look at it.
    Yes, you are right and thanks for clarification in your last thread. One more issue to discuss. I really dont understand the significance of finding the sigma-level from the customer point-of-view becasue it will not really translate what is actually going on with the process and hence the bottom line.
    One suggestion from you: I am going to do a six sigma project in the service industry (in the ticketing department of local aeroplane company). The customers have different kind of complains against the poor service in the reservation department of the company. I was just thinking to take a comprehensive interview of affected customers in order to get the clear picture of what kind of complains they have. Based on the interview, I will then draw a pareto chart and see which complain has got the most weightage so that i can put my whole attention to it.
    After doing all this exercise, I will then start digging all the steps included in the reservation process and see what are really the key steps (opportunities) which can be revised and then reviewed. There is one question in my mind though. Finding the #/#s of opportunities is really a subjective issue, may be i am saying that we really have ten opportunities in the whole process but someone says that he sees more than that!!. In this situation how can we justify the significance of sigma level which is found based on DPMO which depends on # of opportunity/ies in the process.
    I will appreciate your comments .. Regards

    0
    #96871

    Gabriel
    Participant

    “I really dont understand the significance of finding the sigma-level from the customer point-of-view becasue it will not really translate what is actually going on with the process and hence the bottom line”
    It is an indication of how the customer sees you. If every product you deliver is bad, the customer will not think you are good just because you are six sigma in a way that you counted 1 millon of opportunities per product and each product delivered is bad with 3 or 4 defects each (what will lead to something between 3 and 4 DPMO, or 6 sigma)
    On the other hand, do you really see the significance of knowing the sigma level from the process point of view? Tell me how knowing the sigma level or DPMO of your process will help you to imrove it?
    I would say: Don’t bother too much couinting the exact number of opportunities and finding a precise sigma level. The real improvement comes when you eliminate the defects. So put your effort in doing what you explained in your second paragraph.

    0
    #96895

    pise
    Participant

    Gabriel,Our’s is a Machined component ,which will become a finished product after CNC machining & Turning.It has got 25 parameters.Each of them can go wrong.But if I take 25 as opportunity /unit & calculate my sigma levels I will be very close to six sigma.(We produce more than a Lack components per month).Therefore we are taking each component as an opportunity.Please clarify whether Iam right.Otherwise where is the opportunity for improvement.You might have already answered this question,but iam not getting the thread.

    0
    #96911

    Gabriel
    Participant

    Hey, guys. You are asking me a lot of questions about opportunities count, but belive me I am not the right person to answer them. I have never counted opportunities yet.
    Raj, why do you want to count opportunities and calculate your sigma level in first place? If you want to use it as a metric to monitor the improvement effort, do not bother too much for tha accuracy. Just keep the method fixed and, if the metric imprves, then the process improves.
    I would use PPM (or maybe DPMO) together with a defects parto chart. Definitvelly, not sigma level.

    0
    #96940

    Muhammad Khan
    Participant

    Finding DPMO for a process can give us a lot of essential and important information which are helpful to improve the bottom line. Knowing the No. of defects against opportunities actually tells us that the process has a some potential for improvement. Using Process Re-engineering Methodology, we can then go ahead and try to find the main and most important factor which causes the defect in the product. One may find after this digging process that the tool material, which has been used for the turning operation on the product, is really the main cause of defect.
    Regards

    0
    #96947

    Anonymous
    Guest

    Muhammad,
    I concur with Gabrial, but I found you post of interest. I would be grateful therefore if you could explain the tool material problem you mentioned a little further; or was it just an example? Thanks in anticipation.
    Andy

    0
    #96953

    Gabriel
    Participant

    “Knowing the No. of defects against opportunities actually tells us that the process has a some potential for improvement”
    I agree that knowing the defects/opportunity rate, or the sigma level, will be an indication of how difficult will be to improve the process any further.
    “…we can then go ahead and try to find the main and most important factor which causes the defect in the product.”
    You don’t need to count opportunities to do that. Just count defects and make a parteo of them. If you want to monitor the evoution of the quantity of defects and need to compare differnt sample sizes, then use DPU, not DPMO.

    0
    #97082

    AT
    Participant

    Hi Gabriel
    In our case ,we always work to reduce number of Opportunities by using concepts like mistake proofing.This will definitely reduce the denominator which inturn increases the DPMO value.Which is better ? Reducing Opportunities or reducing DMPO ?

    0
    #97084

    AT
    Participant

    I completely accept your approach Mr.Gabriel.I can give an example of a motor cycle having 10000 parts (approx) & if we consider each part as one opportunity ,100 bikes will have one million opportunities.Suppose if one critical parts fails means entire bike will fail.In this case one can not say that his product is meeting 6 sigma as DMPO is Just 1 !!!

    0
    #97085

    AAA
    Participant

    Ramesh,
    You cannot reduce an opportunity. In a certain product the agrred opportunity is constant. You can reduce your DPMO by reducing your defect through your mistake proofing.
    Blackbelt

    0
    #97088

    Gabriel
    Participant

    “Which is better ? Reducing Opportunities or reducing DMPO ?”
    None of them. The better is “Reducie number of defects!”
    Count number of units inspected. Count number of defects. The ratio defects/units, named DPU, is the number you need to improve.

    0
    #97094

    Hemant Gham
    Participant

    Though I was not following this thread, just managed out of curiosity to look at yours.
    My views and experience in a situation like yours was quite intereresting.
    With control and detection methods like “Mistake Proofing” the likelihood of the occurence of a defect can vary along with reduction in no. of opportunities. I suggest you observe the change in no. of defects per fixed sample size by varying opportunities, i.e., as you have said reducing through new concepts. You may find interesting relation y = f(x). Then extrapolating this you can come to DPMO and other calculations.
    But, if not mistake proofing, then as said in Gabriel’s thread, try reducing the defects. Probably, an FMEA can help you here.
    Regards,
    Hemant

     

    0
    #97096

    Mikel
    Member

    How does mistake proofing reduce the number of opportunities?

    0
    #97100

    Bob J
    Participant

    Hi Ramesh!
    The only way I know of reducing opportunities is to simplify/redesign the process (remove muda).  My experience with mistake proofing is that it will not remove the opportunity but will reduce/eliminate the associated defects…
    Hope this helps….
    Bob J

    0
    #97104

    SSNewby
    Member

    One of the better clarifications of “opportunities” is in the following isixsigma article:
    By Gary Cone
    What Is Value Added With Respect To Opportunity Counting?There are three simple rules:

    The product or service is changed by the process (automatically excludes test and inspection).
    The change must be of value to the customer.
    Right first time (no extra credit for redoing – this distinction is important for things like deburring operations – just because your processing throws a burr doesn’t mean you get credit for it).
    So let’s review. Focusing on what you test and inspect or focusing on what can go wrong drives counterproductive behaviors. Focusing on all value added transformations, at the detail level, provides the best framework for counting opportunities.
    Therefore, DPMO should be expressed as –
    DPMO = (total defects) / (total value-added transformations) * 106
    Where,
    defects are the consumption of resources beyond the minimum required to accomplish the work, and value-added transformations are changes to a good or service that moves it toward the expectation of the customer (internal or external)
     
    Following Gary’s crisp and succinct description, mistake proofing the process does not reduce opportunities, it reduces the errors uncured while making value added (value to the customer, in the eyes of the customer) transformations throughout your process steps.

    0
    #97121

    Gabriel
    Participant

    “How does mistake proofing reduce the number of opportunities?”
    Answer one:
    It deppends. Mistake proofing can be done in the control, in the process or in the product design itself.
    Imagine a set of screw, washer and nut, were “missing washer” is a defect. You can do a mistake proofing device that discards sets without washer (mistake-proofing the control), or you can design a process so as it is impossible to “forget” the washer (mistake-proofing the process), or you can make a washer built in the screw itself (mistake-proofing the product). In the last case, you are reducing the opportunities by two (according to your definition, one opportunity for the washer itself and one opportunity for mounting the washer).
    Answer two:
    Who cares?
    From the trhee examples above, two of them did not change the opportunities, and the other one reduced it.
    Al three prevented the mistake from reaching the customer, but only two of them prevented the mistake from actually happen in first place. In these two cases (mistake proofing the process and the product) the effect will be the same (no need to scrap or rework a part for “missing washer” and no “missing washer” reaches the customer). However, mistake proofing the process wil only reduce the defects, thus reducingnthe DPMO too, while mistake proofing the product will reduce both the defects and the opportunities (probably in a higher proportion than the defects), thus increasing the DPMO. So DPMO tells me nothing I can use to evaluate the to implemen one or the other, ot ro evaluate the before vs after.
    On the other hand, all three solution have the same “external” DPU improvement and two of them (process and product) also have the same “internal” DPU. So I would choose one of the last two depending on other factors such as cost of implementation. And I didn’t use opportunities at all.

    0
    #97123

    Mikel
    Member

    I agree with all that you say (as usual) , except your stating that mounting the washer is an opportunity. Mounting is motion, which by definition is non-value added. The transformation occurs when the screw/washer combination is tightened.
    Opportunity counting is nonsense anyway to your main point. I only try to get people to do it from the value add point of view so that we don’t mistake doing tests or inspections for improvement  or control. They are neither. We only test or inspect because we are not sure of our process. Even the adding of mistake proofing where the mistake proofing is only early detection is not improvement.

    0
    #97124

    Gabriel
    Participant

    Stan,
    I learnt from you most what I know about opportunities, which is not oo much anyway. However, I am more and more confused as time goes by.
    Mounting in not value added? It is “just” motion? So I guess that when you go to Mc Donald’s the sandwich has the same value for you wether it is “mounted” or all the ingredients are spreaded on the tray? (no tightening, glueing, or other “value added” activity is required)
    And I guess that my customers will not bother if they find some balls loosen in the box, instead of inside the bearing where they should be. After all, there is no missing ball.

    0
    #122595

    Van Loon
    Participant

    Does anyone have the calculation in excel for PPM to sigma.

    0
    #122596

    Mikel
    Member

    =-normsinv(dpmo/1000000)+1.5

    0
    #157809

    Neshat
    Participant

    Now i hav got the Formula for Sigma Level
     
    Lot Sampling Size=  N
    Number of Defects= D
     
    Defect Per Unit= DPU=D/N
    Yield=Y=e(-DPU)   
    Defective PPM Opportunities=DPMO= (-106)ln(Y)
    Process Sigma Level = 0.8406+ã(29.37-2.221x(ln[DPMO]))
     
    Best Regards
    (Neshat & Irshad)

    0
    #157810

    Van Loon
    Participant

    Oh, thank you so much for responding to my post. I have been afraid to post for almost two years now since only that idiot Stan responded to my first post.
    I’ll be posting up a storm now that I have my confidence back.

    0
    #157812

    Chris Seider
    Participant

    Gabriel,
    I enjoyed your posts and trading of ideas, philosophies, etc…..where have you gone to? 
    It has been a bit too sophomoric with some of the strings in the last half year…..
    If you’ve changed your handle, let me know but I doubt it since I would recognize your style of posts….

    0
    #157814

    Mikel
    Member

    Now that was a poster!

    0
    #157815

    Fake Gary Alert
    Participant

    Don’t  insult  Stan as  he  is  one  of  the  few Gurus in  this  site.

    0
    #157816

    Chris Seider
    Participant

    Hmmmmm!
    Christie Brinkley always had a nice set of …….eyes.
    Yes, it’s getting late over here……Racquetball?…. anytime you get the need……I’m less in shape and more out of practice so you might have a chance! :) 
    Cheers from the NW province….

    0
    #157817

    BC
    Participant

    Don’t know why you call Stan an idiot.  He answered your question and it’s correct.

    0
    #157818

    Chad Taylor
    Participant

    Farrah Fawcet Poster————#1 selling poster of all time. Just a little piece of absolutely useful information

    0
Viewing 62 posts - 1 through 62 (of 62 total)

The forum ‘General’ is closed to new topics and replies.