iSixSigma

Shree Phadnis

Forum Replies Created

Forum Replies Created

Viewing 34 posts - 1 through 34 (of 34 total)
  • Author
    Posts
  • #88339

    Shree Phadnis
    Member

    Dear Sanjay ,
    Pls contact me at [email protected]
     

    0
    #87665

    Shree Phadnis
    Member

    Dear Tony,
    If the data you posted is collected in time series and is with subgroup of 5, just go ahead with X bar R chart to understand the process special causes and forget about the normal stuff as for a subgroup of 5 the central limit theorem would take care of the non normal issue.More ever if you are not making a control chart how can you calculate the Cp or Cpk?Pls get help from a good MBB having good fundamental knowlwdge.

    0
    #87152

    Shree Phadnis
    Member

    Dear Sander pls mail me the manual at [email protected]

    0
    #86882

    Shree Phadnis
    Member

    Dear Vidyut,
    Thanks a lot for the input.I want to share an important thing .In one of my consulting assignments I actually stopped the prctice that was used of dart charts in the manufacturing process primarily because it provides no information on understanding variation, which is paramount in anderstanding any process.There also exists a ststistical flaw in calculating the CPK of the process based on the dart chart , that is primarily based on how the standard deveation is measured.I have seen this mistake beeg made by far too many practitions who calculate the Cp or Cpk based on sample standard dev, without using control charts.to sum, you cannot calculate cpk of a process if it is plauged by special causes and to undrestand the type of variation you will have to primarily use control charts.

    0
    #84154

    Shree Phadnis
    Member

    pls contact me with details, If possible send drawing with current 15-30 readings on all the critical dimensions.I will help you with the tolerancing.my mail id is [email protected]

    0
    #77198

    Shree Phadnis
    Member

    Dear Sridhar,
    I tend to agree with Bob comments.However if you cannot have rational subgroups then contemplate using EWMA Chart.This chart is robust to the normality assumption.
    Shree Phadnis

    0
    #76416

    Shree Phadnis
    Member

    Dear Venkat,
    While your calculations are correct, I would like to draw your attention to application of a P chart or a U chart which are more powerful then hypothesis test as an Hypothesis is a post martem while a control chart would be a analytic type of study.I would like to take you the comment made by Dr Deming about Chisquare testing and Control Charts.
    Shree Phadnis

    0
    #76178

    Shree Phadnis
    Member

    KPMG in India does Training and Certifications for Black Belts.
     
    Shree Phadnis

    0
    #75015

    Shree Phadnis
    Member

    Dear Mike,
    Thanks a lot for your insight.I am fully with you on this issue.
     
    Shree Phadnis

    0
    #74928

    Shree Phadnis
    Member

    To all,A lot of people keep on wondering the reason for the 1.5 Sigma shift and debate it time to time.My personal opinion on the reason for this 1.5 sigma is primarily driven from the fact that understanding variation is the key to success.To understand the variation the most commonly used tools are control charts.While control charts are excellent at detecting drifts and shifts in processes that are more than 1.5 Sigma they have poor sensitivity in detecting shifts less than 1.5.In such cases the EWMA or CUSUM are used but are rarely used on line due to complexity of application.I have a strong belif that the 1.5 Shift is primarily motivated from this back ground.My be Mike Carnel and others from Motrola may be able to throw some light on the same. Shree Phadnis

    0
    #74528

    Shree Phadnis
    Member

    Dear PS.,
    I do not intend to use this forum as a marketing tool, but to answer you KPMG in India does offer comprehensive Black Belt courses.Also I would like to know that if you have done BB training , How many days was the training, What topics were extensively covered and what are you looking at by saying advanced Statistical tools.An understanding of these will help me to help you better
     
    Shree Phadnis

    0
    #74267

    Shree Phadnis
    Member

    Dear Aries,
    You will find plenty of confusing guidance on this issue.To enable you to understand let us define Long Term and Short term
    1. Long term variation of a process is variation estimated of a process using sampling conditions such that the process has been able to exibit both type of variation (Special Cause and common cause).Typically the ssampling may run for week to week, Month to month.
    2. Short term variation of a process is termed as that where the process has not had an ability to exibit the two types of variation .Typiclly this occurs in controlled condition trials, or sampling over only one day.
    One thing you must remember that the time period indicated by me are as examples.
    Based on this when we estimate the Z value of the process We term as Zst or Zlt.
    Generally in most of the cases of the data we collect is assumed to be long term and then based on the Z value obtained we add1.5 to obtain Zst.
    Normally the process baseline in Six Sigma is expressed as Zst and the target is 6.
    For Z bench we calculate the Z value for USL and LSL of the process.Add the PPM defectes beyond the tails of the distribution and then enter the Z value table using this value to obtain a Z value.This is called a Z bench.
    I hope I have clarified your doubts.In case I have not then I have managed to confuse you at an even higher level, well that is what Mikel did
    Shree Phadnis

    0
    #72978

    Shree Phadnis
    Member

    Andy ,
    I agree with Bob and Andy however in real life situations I prefer to use the following method which is based on the poissons distribution.The calculations are much easier to do.All you have to do is at each process step keep track of the defects and the total units produced.Calculate the DPU at each stage. Calculate the thoughput yield using e-dpu.
    Calculate RTY as a product of the above yields.
    In your example 1. Total units 4 and defects 3 Hence DPU= 0.75
    In step two Total units 3 defects 1  Hence DPU=0.33
    In step three Total units 3 defects 1 (2 came ok hence I am considering the 1 left as defect) DPU=0.33
    Using the poissons model
    RTY= 0.472*0.7189*0.7189 = 24.5%
    I find this method more easy to implement.I hope this helps.
    Shree Phadnis

    0
    #72911

    Shree Phadnis
    Member

    There was an excellent article in Quality Progress in some last years issue dealing with similarities in QS9000 Six Sigma and CQE.
     
    Shree Phadnis

    0
    #72811

    Shree Phadnis
    Member

    Dear Joanne,
    When the manaufacturing sequence is known , you can have the data collected in Sub Groups or no subgroups.In either case depending on how the data is collected you can estimate the population standard deveation from the sub groups or the Range and use R bar/ d2  or Sbar/C4 depending on the no of subgroups for the estimation of standard deveation free from special causes, But when sequence of manufcturing is not known you use the conventional standard deveation formula (some times even multiplied by an unbiased estimator).In this method you estimate the population standard deveation which includes both common causes and special causes.For further understanding there are numerous posts on this site for PP,PPKCP&CPK.To put in short Cp and Cpk cannot be calculated without time sequence being known however Pp and PPk can be calculated.
    Shree Phadnis

    0
    #72764

    Shree Phadnis
    Member

    Dear Friends,
    I am attaching a PDF file of the entire Paper helicopter experiment.I am e mailing the same to isix sigma for the benifit of all.
     
    Shree Phadnis

    Download: Teaching Design of Experiment (DOE) – Helicopter Example
    Viewing Tip: Usually, you can click on a link to view the document — it may open within your browser using the application (in this case Adobe Acrobat). If you are having difficulty, try right clicking the link and selecting “Save Target As…” or “Save As…” to save it to your computer harddrive.Virus Note: All files are scanned prior to uploading to iSixSigma. No prevention program is entirely safe. FOR YOUR OWN SAFETY, PLEASE:1) Re-scan downloaded files using your personal virus checker before using it.2) NEVER, EVER run compiled files (.exe’s, .ocx’s, .dll’s etc.). If you don’t have a virus scanner, you can get one at many places on the net including McAfee.com.

    0
    #71751

    Shree Phadnis
    Member

    Dear Mike,
    Use Z charts.
    Shree Phadnis

    0
    #71750

    Shree Phadnis
    Member

    Dear Ratan,
    I can help you out, but not possible to explain on mail.I have done a deployment planning of Six Sigma for a Fortune 500 company who has an extension arm in India, Integrating, MBNQA, Balanced score card and Six Sigma.
    You cannot implement Six Sigma without such strong deployment Planning
    Shree Phadnis

    0
    #71562

    Shree Phadnis
    Member

    Dear Gabor,
    pls read the article on handling Non Normal data in the latest isixsigma news letter
     
    Shree Phadnis

    0
    #71465

    Shree Phadnis
    Member

    Dear Ovi,
    you are welcome.If you need any further help do not hesitate.
    Shree Phadnia

    0
    #71457

    Shree Phadnis
    Member

    Dear Ovidu,
    You said that your response is gas detection right?.Some questions to probe.
    1. Objective is to detect the gas leak right
    2. Objective is to detect it fast  before a disaster occurs right
    3. If this is the primary objective of your device the response should be the time in detection of gas leak by the instrument.
    Hope this helps.
    Shree Phadnis
     

    0
    #71029

    Shree Phadnis
    Member

    Dear Ravi,
    TRIZ is the short form for theory of inventive problem solving.The best book I have come across is Engineering of Creativity by Savransky.The web site http://www.triz-journal.com is also an excellent resource.I am sure once into triz you will get facscinated by it.TRIZ combined with Six Sigma makes your knowlege in the field of improvement untouchable.Any problem of any kind specially in the Engineering side looks very easy and sucess in improvement is practically gauranteed.
    Shree Phadnis

    0
    #70927

    Shree Phadnis
    Member

    I fully agree with you.Use the ASQ exam only to test yourself against the Body of Knowledge.
    I also agree with you regarding the ASQ Exam, which was extermely tough and I also felt that the calculations they made me do at the exam were unnecessay as on projects I would use Minitab instead.
     
    Shree Phadnis
    Master Black Belt KPMG
    ASQ Certified Six Sigma Black Belt
    ASQ Certified Quality Engineer
    ASQ Certified Quality Manager

    0
    #70842

    Shree Phadnis
    Member

    Dear Ravi,
    In fact precontrol is very simple and to get overwhelmed by it is easy, however there are certain points very important in its application.
    1. Applicable to Normal distributions only as the theory of precontrol is based on Probability theory.
    2. Apply precontrol only when you are sure that the process capability of the process is way above 1.33 min In case you do not verify this you can run an extermely large Alpha risk and might actually tamper the process rather than improve it.The rule of 5 consequetive pieces in green means 1.33 CPK should not be used for the determination of process capability
    3. The biggest drawback of Precontrol is that it cannot help us distinguish between the Special causes and the common causes.
    4.The best area to apply precontrol is on the X of the processes after the DOE and after verification of the process capability of the X .
    5. In fact for the application mentioned in point no 4 also it would be more appropriate to apply EWMA or CUSUM Charts.
    In case you need a detailed discussion pls refer to ASQ Publication some where last year in Technometrics.

    0
    #70841

    Shree Phadnis
    Member

    Dear Hoon,
    You are absolutely right.In fact Box plot is a Non Parametric test.Non Parametric tests are those where there is no assumption made about the type of distribution.A simple way of finding the percentile value is
    ((Percentile)*No of readings) + 0.5
    The answer obtained above is the no of reading ranked.For example we wnat to find 25th Percentile of a set of nos where total obs are 40.
    (0.25*40)+0.5 = 10.5
    Which means that the avg value between the lowermost 10th reading and the 11th reading is the percentile value.You can use this formula to determine any percentile value.
    Once the Median. third and first quartile are determined you can make the box.The whiskers of the box extent to the higher most vaue in the data set or the lowermost value of the data set or Q1-1.5(Q3-Q1) on the lower side or Q3+1.5(Q3-Q1).Incase there are any points beyond these calculated values they are identified as outliers.The value of Q3-Q1 is also called the interquartile range.
    In fact in analysis of data set if Histogram is the front view of Data the Box plot is the top view of data

    0
    #70801

    Shree Phadnis
    Member

    Dear Jim,
    Thanks for your explanation, however I wish to make some comments.
    1. The elaboration of the First pass yeild is perfect.
    2. The explanation for the RTY is correct, But in the explanation you said that there were 100 units processed and 30 were defects.I wish to make a comment here that it is very important to know whether 30 are defects or defectives.What you have considered is defective.
    3. to explain simply, First pass yield is based on Defective and the product at each stage gives a understanding of the total yield of the product but
    4. The rolled through put yield is based on the defects.One product can be defective but it can contain many defects.Hence for estimating this yield we use the poisons distribution and obtain the yield by e-dpu for each process and the product of these yields is the rolled through put yield which represents the hidden factory cost.
    5. As regards the explanation for DPU what you have given is DPO that is defects per total opportunity.DPU is direct ratio of defects to units processed without consideration to opportunities.
     

    0
    #70729

    Shree Phadnis
    Member

    This is a typical case observed and we call this as the process is bounded at zero.The data infact in such a case may not even be normal.If the data is not normal various diffrent approaches will have to be used.Direct calculation of Cpk may be wrong.For further information on haow to handle such processes pls feel free to contact me at [email protected]

    0
    #70476

    Shree Phadnis
    Member

    Dear Vasantrayalu,
    ASQ as of today is the only Internationally recognized body for various certifications.As for thier Black belt certification, thier requirement is stringent where you have to give an affidavit from your organization that you have succesfully completed two Six Sigma projects,and this afidavit needs to be signed by your Champion.After which you are eligible for the exam.
    If you take this exam as a challenge and appear for it only to verify your competence then there is no problem because as a Black Belt you already must have done two projects and the exam would then test you on the ASQ Body of Knowledge.
    But if you are looking at this as a way to certification and bluff your way by producing the affidavits rather than actually doing projects you are fooling yourself.In fact I believe that for all the ASQ certification exams one should never try to study for the exam but should directly appear only to verify your knowledge against the best in the Industry.
    The approach one takes towards these certifications is more important than the certifications themselves.
    Shree Phadnis
    Master Black Belt KPMG (India)
    ASQ Certified Six Sigma Black Belt
    ASQ Certified Quality Manager
    ASQ Certified Quality Engineer

    0
    #70451

    Shree Phadnis
    Member

    HI,
    IN FACT TO GIVE A CONCEPTUAL UNDERSTANDING TO SPC I WROTE THE ARTICLE BASED ON ACTUAL CASE STUDIES EXPERIENCED IN THE INDUSTRY.THE ARTICLE WILL GIVE YOU CONCEPTUAL UNDERSTANDING REGARDING SUB GROUPS.
     
    Shree Phadnis

    0
    #70006

    Shree Phadnis
    Member

    List of good resources;
    1. Statistics for experimenters by Box ( A classic)
    2. Design of Experiments by Montgomery
    3. Experiments by Hamada and WU.(This is the latest and very good resource for advanced DOE)
    4. For Dorian Shainain Methods ,World Class Manufacturing by Keki Bhothe
    5. Taguchi Methods ( PJ ROSS)
    7.Robust Engineering by Taguchi
    shree
     

    0
    #69337

    Shree Phadnis
    Member

    Pls refer to my article on isisxsigma on Design for Six Sigma roadmap.It lists out all the tools used.One Additional tool not mentioned is TRIZ.
     
    Shree Phadnis

    0
    #69336

    Shree Phadnis
    Member

    I am a Master Black Belt at KPMG India.You could get in touch with me for details
     
    Shree Phadnis

    0
    #68536

    Shree Phadnis
    Member

    Use Analysis of variance

    0
    #67096

    Shree Phadnis
    Member

    What you have shown is mathematically possible,but I would like to tell you that you are defeating the very purpose of the control chart that is whether the process has special causes or is only a process of commoncauses.Such a use of control charts will derail the purpose of control charts.While using control charts process capability is calulated on only the process that is stable.I request you to study Dr Demings pioneering book Out of the Crisis to understand control charts.

    0
Viewing 34 posts - 1 through 34 (of 34 total)