Read the pages and make a summary of them with your own words, please. It is what you understand. In the check sheet shown in Figure 15.13, five types ofmiswire defects, covering all types experienced, are coded bysymbols, and these symbols are the only raw data enteredon the chart. Sums of all defect categories are shown at thebottom of each column, and the weekly total for each opera-tor is shown at the end of each row. A quick glance at thecheck sheet points to operators 28 and 33 as the sources ofthe problem. We don't know the cause at this point, but weknow where to start looking.In past times, these two people might very well havebeen summarily fired. In a total quality setting, that deci-sion would be considered the last resort. Most employeeswant to do a good job and will if they are provided withthe necessary resources and training. In a case like this, itis not unusual to find that the fault lies with management.The employees were not adequately trained for the job, orsome environmental factor (noise, temperature, lighting, orsomething else) is at fault, or the operators may simply notbe equipped for the task (because of vision impairment, im-paired motor skills, or some other problem). In any of thosescenarios, management is at fault and, therefore, should dothe morally right thing to correct the problem.Check sheets can be valuable tools for converting data intouseful and easy-to-use information. The key is teaching opera-tors how to employ them and empowering them to do so.11/211/311/411/511/6OperatorNo.BenchNo.Week.AM PM AM | PM | AM PM | AM PM | AM PMАМTotalsQUALITY TIPo- 1•-1 3O-1A3An Important Distinction to Remember•.2Attributes Data..- Has or has not- Good or bad10A2O-13o-2. Pass or fail. Accept or reject- Conforming or nonconforming11BIO-1A-14O-213AlVariables Data0 •O A o oA ooo oO AD Oooo.O- 1728C223A-3Measured values (dimension, weight, voltage, surface, etc.)o- 28.-2•* 36A-4A OO Aooo o33C3A OD-2O-1Figure 15.14 would tell us what we wanted to know ifwe were interested only in the number of shafts acceptedversus the number rejected. Looking at the shaft process inthis way, we are using attributes data: either they passed orthey failed the screening. This reveals only that we are scrap-ping between 3 and 4% of all the shafts made, It does nottell us anything about the process adjustment that may becontributing to the scrap rate. Nor does it tell us anythingabout how robust the process is-might some slight changepush the process over the edge? For that kind of insight, weneed variables data.40B2+-1Half-day totals 10 79 6 4 7 6 6 10 9 3935Full-day totals1715 11121974LEGEND: 0= Hand wrapSolder point to pointA= Harness+= RibbonO= OtherFIGURE 15.13 Check Sheet: Defect Factors-Miswires.One can gain much more information about a pro-QUALITY TIPHISTOGRAMScess when variables data are available. The check sheet ofmeasured data. That example used shaft length measured inthousandths of an inch, but any scale of measurement can beused, as appropriate for the process under scrutiny. A pro-cess used in making electrical resistors would use the scaleof electrical resistance in ohms, another process might useweight scale, and so on. Variables data are something thatresults from measurement.Figure 15.12 shows that both of the rejects (out-of-limitsshafts) were on the low side of the specified tolerance. Thepeak of the histogram seems to occur between 1.123 and1.124 in. If the machine were adjusted to bring the peak upto 1.125 in., some of the low-end rejects might be elimi-nated without causing any new rejects at the top end. Thefrequency distribution also suggests that the process as itstands now will always have occasional rejects-probablyin the 2 to 3% range at best.Statistics Expertise Not RequiredWe recognize that although much of the following discussionof histograms and control charts is related to statistics, manyusers of this book will not be expert statisticians. Unfortunately,the scope of this text does not allow for a treatise on statistics,so we have attempted to present the material and mathemati-Histograms are used to chart frequency of occurrence. Howoften does something happen? Any discussion of histogramsmust begin with an understanding of the two kinds of datacommonly associated with processes: attributes and variablesdata. Although they were not introduced as such, both kindsof data have been used in the illustrations of this chapter. Ancal processes in a way that can be followed by the uninitiatedwho are willing to stay with us. In doing this, we have sacri-ficed nothing in the accuracy of the information presented orthe techniques applied. Our objective is that both the statisticsnovice and the expert will be rewarded with a good understand-ing of these tools, their applications, and the methodology andsignificance of the math. For those interested in delving deeperinto the tools or statistics, many books are dedicated to eachof them.Using the shaft example again, an all-too-common sce-nario in manufacturing plants would have been to place aGo-No Go screen at the end of the process, accepting allshafts between the specification limits of 1.120 and 1.130 in.and discarding the rest. Data might have been recorded tokeep track of the number of shafts that had to be scrapped.Such a record might have looked like Figure 15.14, based onthe original data.attribute is something that the output product of thecr pro-cess either has or does not have. From one of the examples(Figure 15.6), either an electronic assembly had wiring er-rors or it did not. Another example (see Figure 15.30) showsthat either an assembly had broken screws or it did not.These are attributes. The example of making shafts of a spec-ified length (Figures 15.11 and 15.12) was concerned withShaft Acceptance: Week of 7/11 (Spec: 1.120-1.130")DateAccepted11Rejected1.112.13.121114.1215.12Totals:58FIGURE 15.14 Summary Data: Weekly Shaft Acceptance. The data in Figure 15.11 are the same as the data inFigure 15.12. Figure 15.11 shows columns of sterile data that,before meaning can be extracted, must be subjected to hardwork at someone's desk. Assuming it does get translated intomeaningful information, it will probably still remain invis-ible to the people who could make the best use of it-theoperators. That can, of course, be overcome by more hardwork, but in most cases, the data will languish. On the otherhand, Figure 15.12 provides a simple check sheet into whichthe data are entered more easily and, once entered, providea graphic presentation of performance. If the check sheet re-veals that the machine is creeping away from the center ofthe range or if the histogram shape distorts, the operator canreact immediately. No additional work is required to trans-late the data to useful information, and no additional work isCHECK SHEETSShaft Length: Week ofLength7/11(Spec: 1.120-1.130)Check SheetDateDateLengthDateRemThe check sheet (see Figure 15.12) is introduced here as thethird of the seven tools. The fuel that powers the total qualitytools is data. In many companies, elaborate systems of people,machines, and procedures exist for the sole purpose of col-lecting data. At times, this quest for data has become zealousto the point of obscuring the reason for data collection in thefirst place. Many organizations are literally drowning in theirown data, while at the same time not knowing what is actu-ally going on; they are "data rich and information poor." Withthe advent of powerful desktop computers, information col-lection has become an end unto itself in many instances.Having access to data is essential. However, problemsarise when trivial data cannot be winnowed from the impor-tant and when there is so much of it that it cannot be easilytranslated into useful information. Check sheets help dealwith this problem.LengthShaft Length: Week of 7/11 (Spec: 1.120-1.130")111.124111.128111.123.....111.126111.128111.1251.118* 13111.119111.123111.1221.123111.120111.122111,119** 11** Out of Limits121.124121.126121.1251.12011 13121.125121.127121.1251.12112 13 15 15121.1211.126121.124121.12512121.124121.1271.12211 11 13 14 14 15 15 151.1231.12013131.125131.1211.12311 11 11 13 13 13 14 15 15 1513131.122131.118131.124131.123131.1251.12411 12 12 12 13 13 14 14 14 15 15 15131.126131.123131.1241.12511 12 12 12 12 13 13 14 14141.125141.127141.124141.126141.129141.1251.12611 12 12 13 14 14141.126141.123141.124required to broadcast the information to all who can use it.To set up a check sheet, you must think about your objec-tive. In this example, we were making shafts to a specification.We wanted data indicating how well the machine was per-forming, a graphic warning whenever the machine started todeviate, and information about defects. Setting up the checksheet as a histogram provided all the information needed.This is called a Process Distribution Check Sheet because it is1.12712 12 14 15141.122141.124141.122151.124151.121151.1231.12811 11151.124151.127151.123Enter day of month fordata point.1.129141.1241.12315151.122151.12215151.122151.1211.130The check sheet can be a valuable tool in a wide varietyof applications. Its utility is restricted only by the imaginationof the person seeking information. The check sheet can takeany form. The only rules are that data collection must be theequivalent of entering a check mark and that the displayeddata must be easily translated into useful information. For ex-ample, it may take the form of a drawing of a product with the situations, that would be appropriate. After all, computers Results.check marks entered at appropriate places on the drawing to are good at digesting raw data and formatting the result forillustrate the location and type of defect. An accounts receiv- human consumption. But before the computer can do that,able department might set up a check sheet to record the types some human must tell it exactly what it must do, how to for-and numbers of mistakes on invoices prepared. Check sheets mat the information, what to discard, what to use, and soapply to any work environment-not just to the factory floor. on. If we can't first figure out what to do with the data, noThe purpose of the check sheet is to make it easy to amount of computer power will help. On the suppositioncollect data for specific purposes and to present it in a way that we do know what to do with the data, it is possible thatthat facilitates conversion from data to useful information. we could preformat the data so that it will be instantly usefulexample, suppose we are manufacturing parts that have as it is being collected. This is one of the powerful capabilitiesa specified dimensional tolerance of 1.120 to 1.130 inches of the check sheet.(in.). During the week, each part is measured and the dataare recorded. Figure 15.11 is a summary of the week's results. porting how the shafts being produced relate to the shaftThis figure contains all the data on shaft length for the length specification. The machine has been set up to pro-week of July 11. Without a lot of additional work, it will be duce shafts in the center of the range so that normal varia-difficult to glean much useful information from this list of tion would not spill outside the specified limits of 1.120 anddata, Imagine how much more difficult it would be if, instead 1.130 in, and thereby create waste. If the raw data could giveof a table, you were presented with a stack of computer runs us a feel for this as it is being collected, that would be very sheet in Figure 15.12 only requires noting the date (day of performance. The check sheet will list each operator's numberseveral inches thick. That is frequently the case in the infor- helpful. We would also like to know when the limits are ex- month) opposite the appropriate shaft dimension. The day- and bench location within the factory. To determine whethermation age. (The information age should be called the data ceeded. The check sheet in Figure 15.12 has been designed to of-month notation serves as a check mark, while at the same the day of the week or the time of day has anything to do withage, in our opinion, reflecting the abundance of raw, often facilitate both data collection and conversion to information. time keeping track of the day the reading was taken.meaningless data and the real paucity of useful information.)The computer could be programmed to do some- data very easily and at the same time display useful informa- floor, with entries handwritten. That will make the perfor- but using a tool that produces faulty connections is somethingthing with the data to make them more useful, and in some tion. The check sheet actually produces a histogram as the mance of the machine continually visible to all-operators, the operator must guard against. In other words, we will not1.131**FIGURE 15.11 Weekly Summary of Shaft DimensionalTolerance Results.1.132**Note: This is not a check sheet.concerned with the variability of a process. Other commonlyused check sheets include Defective Item Check Sheets (de-tailing the variety of defects), Defect Location Check Sheets(showing where on the subject product defects occur), andDefect Factor Check Sheets (illustrating the factors-time,temperature, machine, operator-possibly influencing defectgeneration).If we wanted to better understand what factors might beFIGURE 15.12 Check Sheet of Shaft Dimensional ToleranceQUALITY TIPGive the Operator Some ResponsibilityThe taking of measurements and the logging of data on thecheck sheet should ideally be done by the operator who runsthe machine, not a quality control inspector. In a total qualitysystem, the operators are responsible for the quality of theiroutput-for checking it, taking data, responding to the data,and so on. The quality control department is there to audit theprocesses to make sure that they are under control and thatprocedures are followed.contributing to excessive defects on the shop floor, we couldset up a Defect Factors Check Sheet. As an example, go back tothe section on Pareto charts and look at Figure 15.6. The topchart there revealed that miswires were the most significantTation.ForThe importance of the data in Figure 15.11 rests in re-defect in terms of cost. To collect some data about the factorsthat might be contributing to the miswire defects, a reasonableapproach would be to set up a Defect Factors Check Sheet andcollect data for a week. We are primarily concerned with theoperators themselves and the factors that may influence theirThe check sheet of Figure 15.12 is set up to accept theThis check sheet should be set up on an easel on the shop performance, the data will be recorded by day and by morningdata are entered. (See the following section for informationabout histograms.) Data are taken by measuring the shafts, supervisors, engineers, or anyone else in the work area.just as was done for Figure 15.11. But rather than loggingthe measured data by date, as in Figure 15.11, the checkconsider a tool to be at fault-only the operator if he or shecontinues to use a defective tool.

on't know the cause at this point, but we t looking. these two people might very well have red. In a total quality setting, that deci- sidered the last resort. Most employees job and will if they are provided with urces and training. In a case like this, that the fault lies with management. re not adequately trained for the job, or al factor (noise, temperature, lighting, or at fault, or the operators may simply not e task (because of vision impairment, im- or some other problem). In any of those ment is at fault and, therefore, should do ming to correct the problem. an be valuable tools for converting data into ase information. The key is teaching opera- them and empowering them to do so. O-2 Good or bad 11 B1 Pass or fail Accept or reject o-1 13 Al . Conforming or nonconforming A.1 0-2 4 Variables Data 28 C2 23 . Measured values (dimension, weight, voltage, surface, etc.) --- --- - ::::: 33 C3 36 Figure 15.14 would tell us what we wanted to know if we were interested only in the number of shafts accepted versus the number rejected. Looking at the shaft process in this way, we are using attributes data: either they passed or they failed the screening. This reveals only that we are scrap- ping between 3 and 4% of all the shafts made. It does not tell us anything about the process adjustment that may be contributing to the scrap rate. Nor does it tell us anything about how robust the process is-might some slight change push the process over the edge? For that kind of insight, we need variables data. One can gain much more information about a pro- cess when variables data are available. The check sheet of O.1 .1 3 +-1 40 82 Half-day totals 10 79 6 4 76 6 10 9 39 35 17 15 11 Full-day totals 12 19 74 LEGEND: Hand wrap . Solder point to point A= Harness Ribbon Other FIGURE 15.13 Check Sheet: Defect Factors-Miswires. HISTOGRAMS measured data. That example used shaft length measured in thousandths of an inch, but any scale of measurement can be Figure 15.12 shows that both of the rejects (out-of-limits used, as appropriate for the process under scrutiny. A pro- tise Not Required shafts) were on the low side of the specified tolerance. The although much of the following discussion control charts is related to statistics, many will not be expert statisticians, Unfortunately, ext does not allow for a treatise on statistics, oted to present the material and mathemati- way that can be followed by the uninitiated stay with us. In doing this, we have sacri- e accuracy of the information presented or plied. Our objective is that both the statistics ert will be rewarded with a good understand- their applications, and the methodology and math. For those interested in delving deeper tatistics, many books are dedicated to each Histograms are used to chart frequency of occurrence. How often does something happen? Any discussion of histograms must begin with an understanding of the two kinds of data cess used in making electrical resistors would use the scale peak of the histogram seems to occur between 1.123 and commonly associated with processes: attributes and variables a weight scale, and so on. Variables data are something that data. Although they were not introduced as such, both kinds of data have been used in the illustrations of this chapter. An attribute is something that the output product of the pro- cess either has or does not have. From one of the examples Go-No Go screen at the end of the process, accepting all (Figure 15.6), either an electronic assembly had wiring er- rors or it did not. Another example (see Figure 15.30) shows that either an assembly had broken screws or it did not. These are attributes. The example of making shafts of a spec- ified length (Figures 15.11 and 15.12) was concerned with 1.124 in. If the machine were adjusted to bring the peak up to 1.125 in., some of the low-end rejects might be elimi- nated without causing any new rejects at the top end. The frequency distribution also suggests that the process as it stands now will always have occasional rejects-probably in the 2 to 3% range at best. of electrical resistance in ohms, another process might use results from measurement. Using the shaft example again, an all-too-common sce- nario in manufacturing plants would have been to place a shafts between the specification limits of 1.120 and 1.130 in. and discarding the rest. Data might have been i keep track of the number of shafts that had to be scrapped. Such a record might have looked like Figure 15.14, based on the original data. to Shaft Acceptance: Week of 7/11 (Spec: 1.120-1.130) Date Accepted Rejected 11. 12. 13. 14. 15. 11 12 11 12 12 Totals 58