S@ACE SPC STATISTICAL ROC Ess CONTROL _ i bebiee ee SgACE Contents Section 1:- SPC Introduction and Concept of Control Chart SPC & Voico of Process Variation : Common and Special Cause Basie Staisies Location, Spread, Shape Normal Distbution Contol chat fundomentals Type of data and control chart Selection of Control chads Contol erat preparatory stops Section bar 8 R chart Data colton Limit calculation and vatisaton Process capabity - Mothad of calculaton Cp, Cpk, Pp & Pok Control chart interpretation rules Vv Variable Common Control charts Ty Section inp Chant eu chan Attribute Control Charts Introduction <¥ACE Section — 1 SPC Introduction and Concept of Control chart SéACE What is Statistical Process Control By Definition ‘- Controlling the process variations by using statistics is called Statistical Process Control (SPC). STATISTICS :- Deals with designing for collection of data, data analysis, interpretation and initiating action based on the analysis . or The set of information derived from the Sample data to estimate the process (Population) are called Statistics. PROCESS :- Converting an input into an output by using Man, Machine, Material, Method and Environment. CONTROL :- Ensuring to make a variable to be with in the stated limit { ee -S@ACE TRADITIONAL PROCESS CONTROL Inspect PROCESS Input | 7. Output esas cease (Product or Service) Coy Based on defect detection Correct ] Detect, Little or no reference to the G process A goal post mentality {Reject SdACE Traditional Philosophy Anything outside the specification limits represents quality losses Goalpost MentalitySgACE PROCESS CONTROL: A BETTER APPROACH THE PROCESS Method Environment (Product or Service) Input ——} r—— Output People Equipment Material Tt I. Listen Voice of Process ¢ Collect 5 Act [Record Y@ACE Prevention vs. Detection > In the past, manufacturing depended on inspection to screen out nonconforming product. -This process produces rework and scrap, in other words, waste or lots of MUDA. -Inspection does not increase quality, it only affects customer annoyance. -Detection tolerates waste » Strategies for prevention are required in today’s markets | First, mistake- proof the process -If mistake- proofing is impractical or impossible, then control the inputs | to prevent nonconforming outputs -Prevention avoids waste sY@ACE VOICE OF a PROCESS THROUGH SPC XéACE Bits i: SPC HISTORY. Developed By Dr.Walter A. Shewhart During 1920’s in Bell Lab Since then SPC has evolved to cover different processes | —YdACE VARIATION: | The basic principle of SPC | YACE | What is this aR cre mammmmmameee | * One problem with mother nature, every thing is different + No two things can be produced same in this world. By all efforts we can only reduce the gap between the two This gap is known as variation. * Dissimilarity between two products for the same characteristic is called variation. + The inevitable differences among individual outputs of a process is called variation. + No two things are exactly alike |+ No two people are same... + Temperature changes continuowsly... [+ The products we produce change continuously ee 12S@ACE Why do you want to reduce the variation? + Because variation is the main source of wastage, undesired reworks, rejections, customer dissatisfaction and many hidden, unrecoverable costs. Common Causes * Common to all individual readings in time periods. A process operating under common cause is called Under statistical controlS@ACE Special / Assignable Causes * Sudden in nature + Usually attract the attention of local people associated with the process. They are not common to all time periods, but they can cause process fluctuations which are large in magnitude 655 a.m. +/- 5 minutes. This man wants to reach his work place by 6.55 a.m.. But he can not do so, exactly at 6.55 a.m. daily. Sometimes he reaches earlier (but almost never before 6.50 a.m.). Sometimes he reaches later (but almost never after 7.00 a.m.). WHY ?THIS IS BECAUSE... OF CERTAIN FACTORS WHICH + Affect the time he takes + He cannot control + Vary randomly — e.g. The traffic you encounter under normal course of travel THE VARIATION THAT OCCURS DUE TO THESE KIND OF FACTORS IS CALLED INHERENT | VARIATION OR COMMON CAUSE VARIATION OR WHITE NOISE. - e.g.. mic vibration,tool wear etc. XZACE | UNDER NORMAL SCHEME OF OPERATION Maximum deviation Inherent ' Aimed value Variability | (white noise) | Minimum deviationPROBABLY BECAUSE: * His watch was running fast. * He gota lift. * His Bus driver took a shortcut. * He stayed over in the colony. | * He had some important work to be finished before 7:30. | These causes are characteristic of | a specific circumstance and do not occur in the normal scheme of actions. Variation due to these types of reasons is called assignable or special cause variation or black noise 19 TODAY HE IS LATE EY? a other place. PROBABLY BECAUSE : * He overslept. * He missed his Bus. * The Bus driver was new * He took some other route * He stayed over at some These causes are characteristic of a specific circumstance and do not occur in the normal scheme of actions. Variation due to these types of ‘reasons is called assignable or special cause variation or black noise 20QéACE GRAPHICAL DISPLAY OF VARIABILITIES CASE I Assignable Assignable (Black noise) Variabili Variabilit Inherent Variability Aimed Value Assignable Variability Assignable Variability | | CASE II S@ACE mpHo4d KH mmo < How to estimate process behavior ? Shape | Location SpreadY@ACE ae | Process can differ in ; Location Spread Or any combination of these 23 ACE Location : MEAN — Centre of Gravity The Mean of ‘n’ values is the total of the values divided by ‘n” yemth tat. n In Standard Mathematical Notation it is inn) = x= ve estimates i. y Xx, 1 ma N The Symbol “2” means ‘Sum of’ 24SdACE [ MEAN : Example Caloulate the mean of the following data set [Par esate —— = 101 6.733 if 15 25 X@ACE Spread - Range The difference between the largest and the smallest of a set of numbers. It is designated by a capital “R” Examples Data Set 45 47 49 51 : 46 47 «449° («52 47 47 50 53 47 48 St 54 Rangeis : 54- 45=9 Range is not a very powerful statistie to measure dispersionAACE Spread — Standard Deviation The average distance between the individual numbers and the mean. It is designated by “o” oe [RRP HRP OX? N estimates a7 NACE Standard Deviation : Example | Determine the standard deviation for the-following sample data set: 1, 2,3 S= |(-1)2+0+(1)? " =1 Nireiennrn 28 a —_=_—=—_ = aY@ACE Shape: Histogram Histograms give a graphical view of the distribution of the values It reveals the amount of variation that any process has within it Histogram. 29 S@ACE | WHAT IS_ HISTOGRAM : + Adiagram that graphically depicts the variability in a population. 30S@ACE WHAT IS HISTOGRAM : + The frequency data obtained from measurements display a peak around a certain value. The variation of quality characteristics is called distribution. * The figure that illustrates frequency in the form a pole is referred to as a Histogram. a SgACE L POPULATION AND SAMPLE * The entire set of items is called the Population. * The small number of items taken from the population to make a judgment of the population is called a Sample. * The numbers of samples taken to make this ‘judgment is called Sample size. oat eee ™ SAMPLE SIZE: THREE 32ACE Histogram — steps 1.Obtain a set of 50 ~ 100 observations as shown below: Sample Results of Measurement Number 140 |2510|2.517 |2.622|2.522 [2510 | 2.611 | 2.519 | 2.532 |2543 [2.625 1120 | 2527/2536 [2506/2541 [2.512 | 2.521 | 2.521 | 2.536 |2.529 | 2.524 21-30 | 2.529] 2.523 |2.523|2.623 |2s19 | 2.538 [2.543 | 2938 | 2518 | 2.694 e40 | 2.520 | 2.514 |2.612|2.634 | 2.626 |2.632 | 2.632 | 2626 | 2.523 | 2.520 41-60 [2.636 | 2.523 | 2.626 | 2.525 | 2.532 | 2.530 |2.502 | 2.530 |2522 | 2514 s1-60 [2.633 | 2.510 | 2.642 |2.624 | 2.530 | 2.638 | 2.522 | 2.535 |2.540 | 2.528 61-70 |2.625 | 2.515 |2.520| 2.819 | 2.526 | 2.542 | 2.522 | 2.542 |2.540 | 2.528 71-80 | 2.634 | 2.645 | 2.524|2.5622 [2.620 [2.619 |2.519 |2529 | 2.522 [2.513 a1-90 2618 | 2.527 [2511 |2519 | 2.531 | 2.827 | 2529 | 2528 |2519 | 2-521 lot ACE Histogram — steps 2. Obtain the maximum value and minimum value: Sampo Results of Measurement Waximam —] Minimum valve Number value ofthe | ofthe ine ine v0 [aae [aan [asa [asa [ano [aan [aso [ame [oud [zms [zoe 750 11-20 2sa7 | 26x6 | 2506 | 2641 | 2612 | 2521 | 2521 | 2536 | 2020 | 2520 | 25a 2.506 | 21-30 2sze | 2523 | 2523 [2823 | 2s19 | 2sse | 20 | 2538 | 2518 | 2538 | 2503 2518 srw [zea [ase [am |aam [nem [ame [owe [ase [zea [zum [ame ia at E can wim [Par [zen [ame [eae [aan [ame [aaer [aa [aaa [aw [ome ame ‘srs | 2ssa | 2st | 2sez | 2eze | 2830 | 2595 | 2822 | 28s | 2600 | 2520 | 25a2 2510 ci70 | tom [zero | tame | zat | tame [aaar [oan [zea |aew [ame [zee Bate wa | zw [aan [aan |e | aa [aay [oem [asm [zou [am aan Lesa pe an 2 wahena | Pune | | 7Y@ACE Table Number of data (N) Number of classes (K) Under 50 6 = 7 50 - 100 6 - 10 100 - 250 T= 12 Over 250 10 - 20 35 YZACE Histogram- Steps je the number of classes: . Generally to decide the number of classes, divide the range by 1,2 or 5 (their fractions of multiples). In this case, the range is 0.043 and if it divided by 0.002 or 0.005 or 0.010, we will get - 0.043/0.002 - 0.043/0.005 -, 0.043/0.010=4.3~4 Therefore, number of interval of classes be taken as 9. 4. Determine class width rounded off to a convenient figure. So that it covers maximum and minimum value both C=M Mi 0.043 /9~ 0.005 : - 36ACE Histogram- Steps 5. Calculate the class boundaries so that it covers minimum and maximum value both. In the case, let the boundary be 2.5055. Note: First class boundary should contains smallest value and boundary value falls on half of the unit of measurement . 6. Calculate the mid point of first class by sum of upper and lower boundaries of first class i.e. (2.5005 + 2.5055)/2 = 2.503. Mid point of second class shall be (2.5055+2.5105)/2= 2.508 and so on. 37 YZACE Histogram —Steps 7. Make a frequency table as given below: Class Mid-Point | Frequency Marks (Tally) | Freque [| of Class x ney f 1| 2.5005 - 2.5055 | 2.503 |/ 1 2|2.5055-2.5105| 2.508 |//// 4 3|2.5105- 2.5155) 2.513) | /4// /// 8 4|2.5155-2.5205| 2.518 | /AL/ SEL //// 14 5|2.5205-2.5255| 2.523 [AAS SH SHI THI | 28 6 2.5255- 2.5305, 2.528 | /4L/ (HS IHL I 17 7) 2,5305-2.5355| 2.533) | //4/ /TH // 12 8|2.5355-2.5405| 2.538 | //4// 6 9|2.5405-2.5455| 2.543 |/A/ // 7 Total : | 90 Notes: 1, There would be an error in frequency marking (tallying) if the sum of the| frequency f, (3), did not equal the total nuifber (n) of observed value. 38SéACE Histogram-Steps — 8. Mark the horizontal axis with the class boundary values. 9. Mark the vertical axis with a frequency scale. 10. Erect the rectangles over the class interval having area proportion to the frequencies. 11 Draw a line on the Histogram to represent Mean, number of data points and standard deviation. Frequency 39 Types of Histogram Histogram of shaft Dia 25247 2s N=90 a Mean=2.5247 S.D=0.00906ACE Normal Distribution USL 95.4% ——____+ jut of Out of Spec. Spec. w3o- U-20- wo- tr —tg —U20— 0 , Q@ACE © Histogram for grade wise distribution in a class No. of studentsWaACE TYPES OF HISTOGRAMS + Normal Ll [Th — Bell shaped and natural. * Comb like Ll A Loa — Regular ups and downs, — indicates possible measurement error or rounding problem * Positive or Negatively skewed Lal [hy — Possibly due to a limiting process param: ie 43 YZACE p Precipice type — Indicate filtering out through inspection TYPES OF HISTOGRAMS — From multiple sources with small differences in averages — Look for stratification Bimodal Ly (iia. — Two peaks — coming from two different ‘sources/populations Isolated peak type — Outliers indicate mistake in sampling, data ot [Ha measurement — Possibly process shift during data collection — Or incorrect representation | + Plateau type tie L if 44.Normal Distribution Curve— Relation between spread & sigma 99-73% -36 |< 2 +30 95.46% -20 |= } +] +20 68.26% (-tsigma, 45 SZACE . By collecting sample data from the process and computing their = Mean = Standard deviation and * Shape | Prediction can be made about the process 46S@ACE For routine process control, we need Simple computation ] Easy to use by operators for ongoing process control Help the process perform consistently, predictably for quality and cost - Achieve Less variation in output Lower unit cost Increase effective capacity Provide a common language for discussing process performance Distinguish special from common causes of variation, as a guide to local or management action 47 Control Charts = Transformation of a normal distribution curve in the form of 3 parallel lines, where - + The middle line indicates mean and called central line (C.L.) The upper line indicates mean Mean +3 Sigma and called upper control limit (UCL) The lower line indicates Mean — 3 Sigma and called lower control limit (LCL) 48&éACE Types of Data Attribute Variable Anything that can be classified | | Anything that can be measured | Either / or Height Pass / fail Dollars Good / Bad Distance Go/ No Go Speed | Count Continuous, infinite | Discrete Stimulates Prevention Detection Oriented SgACE 49 Common Control Charts Variable = Average and Range (X-R ) = Individual and Moving Range (X —MR/1—MR ) Attribute * p/npChart : Unit Nonconforming (Defectives) * ‘c/uChart — : Number of Nonconformities (Defects) 50. Complete preparatory steps . Data Collection . Making Trial Control Limits Validation of Control limits Process Capability Study On going control Improvement NOMRONS Steps for Control Charts | Control Chart: Preparatory Steps > Create a suitable (conducive) environment > Select characteristics > Verify Measurement System capability > Select suitable control chart 62 ——— ————SPC — Charts For All Occasions < Data > N Defective & Defects ? % Defective No Subgroup 24 EMR Chart SS ‘Sample Size ? a Chant ‘Yes ‘Subgroup mi ~ size >8 Compute ot Sigma inp-Chart & p-chart Yes Li} X-bar & S Chart -Chart & u-chart Section -2 Variable Common Control ChartsCONTROL CHARTS (VARIABLE) Average — Range (X — R ) Chart » When ? y Measurement must be variable. ¥ Situation must be practically feasible to have at least 2 measurements in short span. Mass Production ¥ Suitable for Product (Output) Characteristics. ¥ Suitable for both Normal & Non- Normal Data. Hee eee eee eee pees SgACE CONTROL CHARTS (VARIABLE) Average — Range (X — R’) Chart > Data Collection V Decide the Subgroup Size * Rational Subgroup: Variability within subgroup : should be small ach eeeriane OEES@ACE CONTROL CHARTS (VARIABLE) Average — Range (X — R ) Chart » Data Collection Decide Subgroup Frequency = Detect change in the process over span of time. *For initial study, may be consecutive or a very short interval CONTROL CHARTS (VARIABLE) Average — Range (X — R ) Chart > Data Collection Decide no. of subgroups (For initial study: To Define the control limits) * To incorporate Major Source of Variation (Generally 25 subgroups or more Containing about 100 individual Measurements)S@ACE AVERAGE RANGE CHART ‘CUSTOMER : PART NAME: PART NO.: PARAMETER: ‘SPEC.(NOMI): 0.70 MACHINE: TOLERANCE: #/-0.2 02 [03 |o4 Jos [os |o7|os Jos [10 [41 [42 [13 [14 [45 | 16 J 17 | 18 | 19 20 075 [075 [080 [oro | o«o [o7 | oso | cas | 060 | 080 | oxs | a70 | ons | 090 | 076 | ars | ove | ows ‘ oss [oso [oro [ors [ors [av | o70 |oao [ean | 070 | 070 | oss | oso | oss | cso | acs | aco | avo 3 076 | 080 | 070 [oss [avs | 06 | oso | ass'| oso | 090 | eas | ave | oss | oso | 076 | oss | oc | oss rr vas | 070 | 07s [oas | oas | os | oss | avs | oss | 070 | emo | onw | ose | oss | 020 | aso |oas [aro 065 | 075 [085 [oso [070 | 02 | o7s | ars | ors | oas | a70 | a70 | oso | oas | eae | oa foe [aro ° a[= =] al 59 SgACE AVERAGE — RANGE CONTROL CHARTS (X— R ) | Calculate Average of each Subgroup. ste tty »Caloulate the Range of Each Subgroup. R=X, ‘max Xrnin 4X, Xp ,.... X, are individual values within the subgroup +n is the Subgroup Sample Size™ | | |gACE AVERAGE RANGE CHART ‘CUSTOMER PART NAME: PART NO.: PARAMETER: ‘SPEC.(NOMI) : 0.70 ‘MACHINE: TOLERANCE: +02 o1 [oz 03 04 [05 06 o7 08 03, 10 [11 3 14 45 7 19 07s | oe amo [oo 076 rs 080 | 080 o70 rao | ons | exo o70 [07s ors 070 oF 020 | 070 ons 07 rr = 08s oro ‘086 [070 080 08 08s 5 | 088 | oss | 076 070 oF 076 | oss 070 070 eas [o on or 07 ae | 078 ons 026 6 AVERAGE — RANGE CONTROL CHARTS (X—R ) >Calculate Average of each Subgroup. Xi aha Xe k > Calculate the Average Range “*k= No. of subgroups R +R, +R, +.S¢ACE AVERAGE — RANGE CONTROL CHARTS (X —R ) >Calculate Trial Control Limits for Range Chart. UCL, = D;R D, , D, & A, are Constant varying as per sample size (n) QACE Table of Constants for Control Charts Use the table below where n is the number of samples in a subgroup mo |2 \8 ja |b. |6 jf 6 18 ito |) D 3267 [2575 [2262 [2118 | 2004 | 1024 | 1.864 | 1816 | 1.777 4 D 0 0 0 ° 0 oo7e [0136 [ots |o223 3 A 1880 [1.023 [o729 [os77 |o«ss [oato |o373 [0337 | 0308 =. d 1128 | 1603 [2059 [2926 [essa [270s |2ea7 [2970 | 3078 2 |S@ACE AVERAGE — RANGE CONTROL CHARTS (X - R ) = OH For R Chart UCL=D,XR =0.433 LCL =D;XR = 0.000 For X Bar Chart UCL=K+A,XR =0.837 LOL=x -A,XR =0.5991 SéACE DRAW AVERAGE — RANGE CHART _| AVERAGE CHART 0.837 o7s ag 0718 0.599 ta Bi Se 8 8 1 1 12°13 14 18 16 17 18 19 20 RANGE CHART oso ———— 9 33@ACE A Few Words on Control Limits » Basis for control limits : “Control limits are also referred to as natural process limits. Control limits are based on the mean and standard deviation of the process as it is, not how we wish it to be. “There is no connection to specification limits. > Purposes of control limits “To determine when local action is necessary and to act on it (i.e., when special cause variation exists) *To display the natural variation of the process, that is, the common cause variation. “To show any obvious, unnatural patterns in the data Specification limits do not belong on control charts. S@ACE Process Capability Study : Compare Voice of Process with Voice of customer (Specification) Voice of Process ' tst ust ; -30 +30 i Process With Design Width 68Y@ACE Process Capability : LSL = Lower Specification Limit = Tolerance / 60 Calculate Process Standard De Soaaarce na a 1128 3 71693 4, is a constant varying as per sample size (n) 4 2.058 5 2.326 6 2534 > Ps Capability ( C; Calculate Process Capability (Cp ) ‘ = 8 2647 ® 2970 Cp=(USL-LSL)/ 60 USL = Upper Specification Limit 69 S@ACE A Problem With Cp : 43 2401234 a3 24012 Y Cp considers only spread, not the location, ¥ Fora truly capable process = Process spread must be smaller to specification and falls well within the specification = It should be located in a manner that its spread on both the sides "Capability index that considers both location and spread is called Cpk 70Y@ACE Process Capability : Calculate Process Capability (Cpk) Cpu = (USL=X)/30 = Zyg,/3 Or CpL = (X-LSL)/30 =Z,5./3 Whichever is minimum will be Cpk ‘Subgroup size(n) 2 1.128 1.693 2.059 2.326 2.534 2.704 2.847 olal~rlolalalo 2.970 AACE Process Capability : Standard Deviation (0) = R/d, 205 / 2.326 = 0.088 Process Capability Cp = (USL-LSL)/ 60 = (0.900-0.500) /6 X 0.088 Cpu = (USL-X)/30 = 0.689 Cpl, = (R=LSL)/3o = (0.718 - 0.500) /3 XO. Cpk (0.900-0.718) / 3 X 0.088 088 0.738 ISL=0.900 | Us LSL = 0.500 =. — ee eeSgACE Process Capability : cp Cpk | Remarks + Process Capable VO + Continue Charting + Bring Cpk closer to Cp + Process has potential Capability + Process lacks basic Capability we X___|+ Improve Cpk by Local action xX X + Improve process by Management action SdACE What about PROCESS PERFORMANCE ? + Process Capability (Cp, Cpk) indicates the ability of the process to meet the specification (Voice of customer) when Process operates under the common causes. +In practical situation, a process shows variation due to both common ag well as assignable causes. +One must analyze process behaviour due to combined effect of Common and Assignable causes. The index is known as Process Performance Index (Pp, Ppk)S@ACE Process Capability vs Performance : ] | Cpk Ppk Remarks + Process Capable and performing eet Continue Charting + Process has Capability but not a x performing due to assignable causes X + Remove assignable causes by Local action + Process neither capable nor X | performing + May require Management action (Cm, Cmk) A process variation is affected | G2 Long-Term = 0? Short-Term by many factors like ‘ aa + 0? Machines + Raw material variation + 0? Day to Day + Tools | + Operators + O2Operators | + Measurement System + O? Batches + 0? Seasonal + Time z | + Environmental change ete..._| Machine capability is an index which is calculated on the basis of variation contributed by Machirfe onlySgACE Machine Capability (Cm, Cmk) + Take 50-100 consecutive samples/ measurements in short span + Ensure the following do not change during sampling Raw material batch Operator Tooling Method of process Measurement system Environment etc. Calculate Cm, Cmk using the same formulae used for Cp, Cpk _| 7 @ACE Control Charts Ongoing Process Control + Collect the data at the frequency as established. + Plot on control chart. + Perform instant analysis and interpretation. + Give immediate feedback to the process for action if any indication of process behaviour change. | | Shift change, breakdown etc. “This helps identifying assignable causes. | 76ACE Interpreting Control Charts - Ongoing The interpretation of control charts is based on the statistical | probability of a particular pattern occurring by complete chance (or being caused by random variation). + All of the tests identify events that have a less than 0.3% chance of occurring by random chance (outside of 3s probability of being caused by random variation), + Control charts are divided into sigma zones above and below the average line. + Zone Cis <1 from the mean. + Zone Bis between1 and 2 + Zone Ais between 2 and 3 + Beyond Zone A is > 3 79 &éACE Control Chart Interpretation Test:1 One point beyond zone A x UCL cL LcLt-- Test 1 The basic test © Caused by a large change in the process. 0 Requires Immediate actionSdACE Control Chart Interpretation ‘Test:2 Nine points in a row on one side of center line. UCL ae LCL ‘0 Caused by a process mean shift Y@ACE Control Chart Interpretation Test:3 Seven points in a row steadily increasing or decreasing UCL A : == c alee cL c : B ia 0 Caused by Mechanical wear Chemical depletion Increasing contamination > etc a2ACE Control Chart Interpretation Test:4 Alternating Patterns ia *— 0 Caused by > Over adjustment > Shift-to-shift variation >Machine-to-machine variation Q@ACE Control Chart Interpretation Test:5 Two out of three points in a row in the same zone A or beyond UCL cL LCL ‘Test 5 The second basic test © High variation without exceeding the 3 sigma limit © Major special cause variationControl Chart Interpretation Test:6 Four out of five points in a row in the same zone B and beyond UCL cL LCL o Another test for shift oTest 1,5,6 are related and show conditions of high special cause variability eS S@ACE Control Chart Interpretation Test:7 Fifteen points in a row in zone C (Above & below center line) Test 7- The Whitespace test r Occurs when within subgroup variation is large compared to between group variation or Old or incorrectly calculated limitsMACE [ Control Chart Interpretation ucL CL LCL Mixtures Over control > Two different processes on the same chart. Test 8 Alternating Means Test 8 Eight points in a row on both sides of center line with none in zone C ACE 7 Interpretation for Control Charts X Bar Chart R Chart Conclusion Under Control | Under Control Enjoy Under Control Out of Control Spread Changed Out of Control Under Control Location Changed Out of Control Out of Control Both spread and location changed 88individuals (| & MR) Charts | | When to use : > Measurement is variable >There is no rational basis for sub grouping or >The measurements are expensive and / or destructive or J | > Production rate is slow or > Population is homogeneous | » Suitable for both process and product parameters 89 Moving Range Control Charts | 90— f Y@ACE Moving Range Control Charts » Average of Individual values: ee eee es k ~_ 8.04+8.5+4..11.1 y= Se 24 x¥ = 10.57 » Moving Range : MR, +MR, + MR, +.....MR, MR= k-1 MR aoe tll +04 24-1 MR=13 S@ACE Table of Constants for (I-MR) Charts [scoop] 4%) = | OO | & Size (n) 2 128 : aae? [2.660 ial 1.693 - 2.574 1,772 | 4 | 2050 | - 2.282 1.457 5 za | - | aia | 1200 6 2534 | - 2.004 1.184 7 area [007s | tea [ 1.108 | 8 2.847 0.136 1.864 1.054 far zero | oe | 1816 | 1010 92 a a eeMoving Range Control Charts > Upper Control Limits UCL, = X + E,R’ UCL, = 10.57 + 2.66*1.3 UCL,= 14.4 > Lower Control Limits LCL, = X — E,R LCL, = 10.57 - 2.66*1.3 LCL, = 7.4 SgACE 93 Moving Range Control Charts » Upper Control Limits UCL, = D,R UCL, = 3.267*1.3 UCL = 4.25 > Lower Control Limit LCLz = D3R LCL, = 0°1.3 LCLp =0 eyMoving Range Control Charts » Upper Control Limits UCL, = 14.1 » Lower Control Limits LCLy = 7.1 > Upper Control Limits UCL, = 4.3 > Lower Control Limit LCL, = 0 Estimate of Standard Deviation of X : Oc=Rid2 = 1.3/1.128 What is Sigma Level (Z- Score) + Itis a metric, used to quantify how well a process is performing as compared to specification. + Itis a measure of the gap between-the specification boundary and Process mean using o as a unit of measurement. + Number of standard deviations that fit between the average & (upper or lower ) Specification limit. SL us. T ‘Soo cs A three-sigma (3G) process. + A three — sigma level (3 0)process 96What is Sigma Level (Z- Score) LSL B U A three-sigma (30) process [A six — sigma level (60) process Sigma Level (USL) = ea oO L A six-sigma (60) process 7 S@ACE How to calculate Sigma Level (Z- Score) Sigma Level (USL) = USE = Mean o or Sigma Level (LSI) ease o Whichever is minimum will be Sigma level 8QZACE Example + Calculate sigma level / Z score / Zigma for the following data. LSL=275, USL=325, process sigma=5 and process mean = 305 Solution. =[ Usi—Mean 32 |SigmaLevel, = 30 sigma SigmaLe Mean-LSL _ 305-275 a i Cea Whichever is less Sigma Level = 4 ACE CENTRAL LIMIT THEOREM + Central Limit Theorem (CLT) + Theorem states that, regardless of the shape of the parent population, the sampling distribution of means/ averages approaches a normal distribution: * The sample means (X bars) will be more normally distributed around than individual reading X’s. + This is the underlying reason that Xbar-R control charts work. » 100CENTRAL LIMIT THEOREM + Central Limit Theorem (CLT) + Standard Deviation:- + For normal distributions, the spread in sample means (X bar) is less than X’s with the standard deviation means, equal to the standard deviation of the population divided by the square root of the sample size and is referred to as the standard deviation of the mean, 7; + The arithmatic mean of the sample mean distribution will exactly equal the mean // of the population from which the samples were drawn, 101 ACE CENTRAL LIMIT THEOREM- Mathematically o Population ie = H Population 102Section -3 ATTRIBUTE CONTROL CHART MACE | Control charts for Attribute Data > Defective Data [> Defect Data v Package does/ does not leak ¥ Bubbles in a windshield ¥ Lamp does/ does not light v Paint flaws on a casing | Y Gol no-go gauge data ¥ Errors onan invoice Y Correct/ Incorrect Shipment. v Bad die on a wafer | >'p- and np-charts > c- and u-charts ¥_ p-charts for proportioning non- Y _u-chart for the number of defects conforming units, sample size not per inspection unit, sample size not necessarily constant necessarily constant ¥ np- charts for number of non- ¥ c- chart for the number of defects, conforming units, sample size is sample size constant | |__ constantp - Chart >» When to use a p-chart + When it is difficult or uneconomical to make a numerical measurement. + When it is desired to combine different types of defects into an overall proportion + When the available data are for attributes + When the data come from a binomial process > Management Summaries : - + Many management summaries are attribute forms and could benefit from control chart analysis. + Examples : scrap rates, quality audits, first-run yields etc. 105 MACE p— Charts (Attribute) > Select the Subgroup Size Y For p- chart, sufficiently large so that np>5 » Select Subgroup Frequency ¥ Detect change in the Process over span of time > Select No. of Subgroups Y Major source of variation incorporate (25 or more)SdACE a Calculating Control limits for a p-Chart i > For each subgroup + Record :n --- the number of units inspected np — the number of nonconforming units + Calculate p - the proportion non-conforming : p=np/n * Calculate the control limits : p= n1p1 + n2p2 + ------ +nkpk / n1+n2+—-—--- +nk. UCLp = p+ 3/p (1-p) iv LCLp =p - 3p (1-p ) Jn 107 ACE Interpreting p-Charts Attribute Charts have fewer tests for special causes due to lower information content of attribute data. + A point outside the 3c limit is assumed to be special cause variation. + When np<5, the likelihood of runs below p-bar increases. , Take care when interpreting runs in these cases + Unless the subgroups are equal sizes, only the first test is used 108Process Capability for attribute Capability (Attribute) = 1- p bar Or Using Z Score 109 S@ACE np Charts * When to use an np chart? ¥ Use the same criteria as a p- chart: + When itis difficult or uneconomical to make a numeriacal measurement +When it is desired to combine different types of defective into a single value *When the available data is for attribute. *When the data comes from the binomial process ¥ And the subgroup are all the same size % How is it different from p charts? ¥ The actual number of non confirming parts are plotted, rather than the proportion defective 10YéACE Calculating Control limits fora np-Chart__| | > For each subgroup + Record : np — the number of nonconforming units + Calculate np bar - the mean of non-conforming units : + Caiculate the control limits : np UCLp= np +3 |np (1-“2) = np +3 np (1-8 a n N np -3/np (1-5) iF nk 3 LCLp = np - 3Jnp (1- Tt S@ACE np-Chart - Exercise ee) Q Construct an np chart from the data set given below Q Use subgroup size= 100 Q Interpret the result of the chart. Bottle leakage test was recorded and sample size is 100 numbers. Number of leaked bottles were recorded for , each subgroup size of 100 nos. Data for leaked bottles is given below. 1) 2, 2)5, 3)4, 4)3, 5)3, 6)6, 7)5, 8)0, 9)7, 10)5 11) 4, 12) 1, 13) 2, 14)3, 15)6, 16) 3, 17)8, 18) 4, 19) 4, 20) 4, 21)6, 22) 4, 23)2, 24)3, 25)7 | Total number of leaked bottles = 101 nzS@ACE np-Chart - Exercise Total number of subgroup = 25 np bar - the mean of non-conforming units np bar = total number of non conforming units / number of subgroup = 101 / 25 = 4.04 p bar = np bar / sample size = 4.04 / 100 = 0.0404 UCL = mp +3/np (— p) = 4.04 +3.f4.04 0 — 0.0404) 0445.91=9.95 UCL = np + 3\/np(1- p) = 4.04 +3,/4.04(1— 0.0404) =4.04-(3X1.97) =4.04-5.9 1=-1.87=0 113 MACE 7 np- Chart : Minitab Output NP Chart of Sample Count |: -3d0. uci=9.95 8 € 6. 8 2 |B 4 NP=4.04 8 2 0. tcL-0SdACE | Charts for defects- c-charts & u-charts Q When to use c or u- charts y When the data is attribute data of defects, not defectives. v When data comes from poisson distribution ¥ C-chart are charts constructed for number of occurences for constant exposure. ¥ U- charts are constructed for rate of occurrence for either constant or varying exposure. 115 SdACE Charts For Defects - c-Charts and u- Charts > Examples vy Each month, 100 invoices are audited and the total number of mistakes is recorded. yv In a molding process, there is a problem with pinholes in , Plastic bottles. Each day, a number of bottles are examined and the number of pinholes are recorded. yv Each month, the number of accidents that occur in an organization is recorded 116S¢ACE _Gathering c- or u- Chart data > Condition for Poisson distribution the no. of occurrence in the other subgroups. Y The rate of occurrences remain constant for all subgroups. v It is likely to have 2 occurrences in a very small area of exposure. Y The value recorded in any subgroup is the no. of occurrences in the given exposure. > Planning for c & U- charts v the event being studied can usually occur over some continuous time or space, called area of opportunity or exposure v The area of opportunity can consist of some unit of time (8 hours running time), an amount of material(2 sq.ft) or a number of units(100 bottles). a 44 S@ACE v The number of occurrences in one subgroup is independent of [ Calculating Control limits for C-Chart >» For each subgroup + Record : c— the number of defects found. + Calculate c bar - the mean number of defects : _ Git Got Gh + Calculate the control limitsX@ACE C- Control Charts DATE Lees | Aer ‘Aer perl orl ons aon apa? aed peas orl 0 Ped tt Ape 12 Ape ape 1a Apel as Apris Api rd ‘apis Aol 20 Aer. ‘nl Aes eae Aas = c =349/25 =13.96 A UCL, = b +32 = UCI=1396+3V1396=1396+1121=25.17 esse + LCL=13.96-3V13.96 =13.96-11.21=2.75 Ganeewonnnwannanansannanan 5 119 S@ACE C- Control Charts C Chart of C1 va=25.17 Ea13.96 Let=2.75, aYgACE U-Charts >» When ¥ When the data are from Poisson process. v When the data are rates of occurrence with different sample sizes, OG. 7 « 7 nytnz+ nse + Calculate the control limits : UCL, = 6 +3 i= ii _ Note: u-charts are c-charts variable subgroups &éACE 124 Charts for Defects - Summary > Charting defects is one step closer to the ideal, but these charts still focus on the outputs, not inputs. » C & U-charts are also charts for attribute data & suffers the same shortcomings as p & np Charts. It is easier to calculate c charts than u- charts due to constant sample size. When the sample sizes are large enough & the no. of defects is large, attribute data behaves like variable data from a normal distribution. Note: C & U charts are useful to show improvement in defects rates. Vv Vv Note: u-charts are c-charts with variable subgroups 422eee | S@ACE Normal Distribution ren atove 2 AREA ABOVE: Z| 00 coz] ccs | oo | cos [aw | ca | oo | ow “aj osoooo00 0 xsz02 loess] oasiusso;| Gaeno6r 27 o-7er7774) 0 rzo9576 04801185] oe 3 at [oaeoT7 0 aezze Tsao aezesTwaetsomsr| oad sazane 0 ecatezi| 0 so50%067| 0 42887ezoi[0 aateeeea] 2 [Oag0705% ese 0 AOEOHSO 6 HOSS 6.5776] 0 7a TSK] O aeeT Tee 0 HG7SEHTa|O eHG0818| 310 sexes) 0 reanars] 0 soTeco 03 -esteatto uaz o ePSTSOI[ 0 S1G7Z7OL 0 wa a [osu af 03742 Ta 0 SHG TEED 0 SAREBSED 0 TOEISSRT| 0 SZTEG Tos 0 2117 0 EBTSTON|O 3zOSEDU] 3 (0 0885753) 0 s1es177i[ o zaeossoe 0 Zeus] 0.291 teoee5) 02877350820 0 zanesraed 0 27758526 10 Tse -aTezaBl 0 et 73 0 6 TO8SI5| 0257ouDaz| 0 5aee| 0 2a aeecd 0 2ezez"= 0 2ST ‘31 10 Z¥iSesa] esrezezi 0 zene 0 2zmexoi| 029862780] 0.203672] 0 OeaGeTe 0-2 E359e9| OTA TESET| ‘3a [021165] fo aoerorosd 0 acsaeenss|o zones 0 torezeo2[ 0 soesearsl 0 Yosoise 0 1Szan%e 135 [0.0m S| oan] uest| 0 TRTE TOME HEPEOT OREO SSCOET| OTST] “(0 assez] fo. sasazed 0 iveneaze) 0 ea e557 01455083] 0182528 0. avocado 140077123 0137886610] To. :1stsoaar oTzseret| 2 rv4sco] 0125071 01250ase[ 0, rzYoosai| 01ToooTea| 01 a2 42-0 S16007 180i. va3eabT| 0-1raBPPEH 6 rossi] 0105440. soaOsef 010007534] 0.850530 13" pesean0su9 0 B31 757A 5 Oa ec 0 OS0127 se Dass 0 Ons EAP SETHE O DSTEN57a[ 0 DRAB] Ta GGUS [0a 0:07 ase o OTE ees OT4GS37<] 0.079500] DOT B0TS|0.OTOTSONTS 0 Osseo O81 TZTel ‘Tas esanrazo[ oeno 7570 oase6r0o oseo83E 0 6176019 006057077 0.053 easo| 0 Osc07se2| OosTossasT|ODss=rT—Im| — * 1s [Osu OREI. a5 6F2] 0 O5resors[ UOsDSO =e] ODET est asesr700 Oe eseese Dest TESTO DEST a7 101 a azrisie a osrereosa a pager Ta O OSTIRG 0 CSRNEDEDD 0 OSTESTSST OO? ao cases 00351esH 0 foosrse7iafoosragVoo|ocscrevss ocscomserd[o ceewrerte] Yao maT OND foazrazas] sea 7a] 0 oaeSo7e|ocTUA ToT 0 coseSa OT 2 Te Tonzarenoee|o zzz een octets c 0217204 apanteosel 0 orscsaz0a| 0 orezz elo o1e7e270%] 00 eer reetas oor 7228 ef oo Yoozee| 0 OTSR6T aT] OOret7 Teal a ove77 Feo oTsseez0{0.Te0Nsso| 0 OTASIeaL 0 OTeRS0=e 22° oF 350590 661365268 001053 0.0187 3578 0 abUseze | OFzzAD| 601 TS TOs] 001 eT OUT isOs | COHOTOSN?| 2 Toorarzies [oc toxcso] 001017 o.o0003059 odes eH] 0 Oonseans| 0 GOSTST#52 0 OOeBGHaa| 0 CORREA 0 cORADT A] 0.005000] 3.0TaTeDee 0 7 renee 0 UTenoe | G07] OOO DsTe] 000546367 0.06 see 0 Oee60TZ0] 00538718 25 {0 oos2oeso| .coadseereo oeas77en 0 06703t47| 0 oaes4z6s| 0005365170] 0.006259595| 0 oeosssee oessotat 0473882) + [ae [eeoteetzztocousz arf ooze e450 Opto 000307076 0 00372607 0 Doar ee 0 W554 [Er oconsraasocosazti o ooeeerae oon teereal o0507209| 0 ooze7T 0 000299025 000200872 0002 Ks] O.DIASEAET 2 [000758591 0.002477 ae 0 OoanTzae 0 oa rens 0 zzs6r<0 0 ox B59 00071870 0coz05224) 000 Sbxaa| o.oo snare] oo aaseH0 Doo reOrZT [0 OOT7SOzz|o DITeateat ooo Tot] UD Baane 0 DOTEsEz6t|o 0 aaooee| Oias STi] oT saMees ea3455670 ov 306a 0.c012543] 009122238] 0.001 129600001 ate] 00110675) oc0r07036 000703507 «COTES SgACE z [om | om [cm | cos | oo [00s | om 008 3.1 [atocerero0853555% 000823) oneterro terest or 25) "3:2 [ Woon 0 TONS 0 doo] 0.0051 ra Bancser zl oenogse "3: [ oobeest,opoOreSsa| 0 ooONsU Tao coOes8| O0 [ooobdoeTTs|OCoCSECTET| 6 GaAs 0 ONIBETE] ‘3 [comes Boas srs ooroa0re ‘ownzeo3 0 coo77 1350 [.sp0as7sa 0 o002 55] 35 | .ooaz673 ooazaoe7]| -ooztss%e|o.o0a0TeZ| ear esee 0 crores 600173 0 00168977] 35 | 000016 0.000535 0000147557] 0. 0zeTTe| Songrsr rsa oo0orze 140 cooT2 5 obo aS 0601253] 7 [wooo Toes0 oon Tosso 0. oooeaest| 0 o00257ee Banooreuts|o coco) 38 | 0ooeOrITE ooonesr ao mone sangeet b OEE" 0Cooobee| © ODODEZILC ONDOEOTO| Sogoas 16 0 omoeste7| 0 ooaeezas| on oo0e26 aodasecer o ooeas? <0 conr=sesao orca 0 OODo53052 ToreooosTReo:D500507# ovo01 3 One oa2|Ocacaze7 sf 0 o05N3552 0 Goen2€.0| 0000735 ef 000002253 005002581 0, oaatee Dan 57a o oo0Tesseo ogo; azo coco1737e| oooon Tas] 000015822] a0 zs 0.0; esae o.oo0130%8 o,oootsaee 0.900277 [0 oo0tzzz3[0 oto esa w coco Ts] conor Tode| oon 0228] 9 coo00 reo 0 oooouse| 0. on0sam) {[[oovoseaso aaonsreo oonaran|o toroorae o coccc7 1] cooooasT 2] o.ooouesoe| 9 coN0K="7| .cooosa30 00008672 DRGEETT © eoRDEST aC oRNONAST O UODAOAT neon fp onconsr DISA cOCOOSEA CONDOS 6 OHO OOOO EO DONNER ONDIZEEN O COOP oaxoz¥T5 0 eocoozvs[v como TS o wooo] 0 Coo ae 0 DoNOes| DoDRD0 ees oaccoo coe o Cocco vaso ooND013] 'ootoors02 © cocoor2|ctooooT Teo onbooT Tz 6 Cocco TE] oDoDoDTore| oooRMITeEDD oCoenez2 0 eRe Te o ONOOORS] entocore@ ecenenrs{ceogoo07 ao exooocc 3 ceconso| a amannsts| ooonoTeee 0 conenese coccoces 0 20005] DS ooooae nots 0 omoaoae 2 eexaccse fo oanaNT2| 2 dons! 0 coooeNss4 0.A7o00G sf 009000952 000000287 0 e-opnoaess 0 moaon2t| peacccora] «Capea 500002 o| 7. SSON 90 0. oooeTES OSORNO aonaor Too aaeren ee tooeoor sto oooooe 4b oaoos0 | 0 aaxaa 30 0 osod za] oononTt 7 0 osoaoT FT| aononotcl 3 ooo0oTO0-o ecceooE a aDHEOEB COEOCCON| 0 SOSSGNTE| 0 OSNNNNTE| 3 SOs aan 9 oononoseo oaoeouss|c pouooner o omooodw| 0 cocooce| oCosoabxa vcooonode| 9 20000 ooonan3| oooooass|o Ooavannzz| oonoaTaN| o ooooao7e| DooRGoo0z|vonoooou WooeDazA| SoU feos rapnoo0T#_o ooounore| « Ooxna0t7|o Dman00Te Deaccoco |v SooooDT aU voapaDT4| 0.900000" roosoo0nTt ‘eannaoor-WoncocodfvooosoapTol cmmosocafvcooawNe (Gootosos08 0 aren coos] [[-[reooooone ooocooons cooonons cconsons|sceootone| pancocose ¢eoaoscobi 0 acenanc0 [Ci [0 eooo0000 006006003 0cooanons| ¢cooanans|sonoeocons| eoceeaee|ceowooeon|oooena0c0| 6 conoc00a 8" oBdoNooOr 0 coos ooosoNOcA| cooNsoNNN cooNNNNN | ODO OANA oaooToO cca [ooaoeoonoeoecoree| o nxtoenoT|o oxorooao «eben oocoeacen| o.sooanoNT | sooo pono Hite 128QgACE ____ Glossary of Terms Process capability is determined by the variation that comes from common causes. It generally | represents the best performance of the process itself. This is demonstrated when the process is being operated in a state of statistical control regardless of the specifications. Process Capability — The 6 o range of inherent process variation, for statistically stable processes only, where o is usually estimated by o, = Rid, Customers, internal or external, are however more typically concerned with the process performance ; that is, the overall output of the process and how it relates to their requirements (defined by specifications), irrespective of the process variation. Process Performance - The 6 range of total process variation, where o is usually estimated by s, the total process standard deviation. coe flet | n-| 126Y@ACE Range : A measure of process spread. The difference between the highest and lowest values in a subgroup, a sample, or a population. Variable Data : Quantitative data, where measurements are used for analysis. Examples include the diameter of a bearing in millimeters, | | the closing effort of a door in Newtons, torque of a fastener in Newton-meters. Attributes Data : Qualitative data that can be categorized for recording and analysis. Examples : where the results are recorded in a simple yes/no fashion, such as acceptability of a shaft diameter when checked on a go/no go gage. 27