IQI offers an extensive array of cost-effective training seminars for improving the performance of manufacturing processes. Consider scheduling an on-site presentation of one of our proven seminars to help your company launch an improvement initiative or reenergize your current efforts. For over 23 years, we have been presenting on-site seminars for our clients in North America, Europe, Australia, Asia, and the Middle East. All of our trainers are Fellows of the American Society for Quality (ASQ), ASQ Certified Quality Engineers, ASQ Certified Reliability Engineers, Certified Six Sigma Master Black Belts, and have over 20 years of industrial experience.

IQI Courses available for On-site Presentation

Six Sigma Black Belt Certification
Six Sigma Green Belt Certification
Six Sigma Management Overview
Basic SPC
Short-Run SPC
Measuring Process Capability
Statistical Problem Solving
Gage R&R Studies (MSA)
Analysis of Means (ANOM)
Verifying Process Improvements

On-site seminars eliminate the traveling costs associated with sending your personnel to one of our public seminars. In addition, class participants minimize their out-of-office time by attending an on-site seminar.

We offer to customize our training manual for on-site seminars to include only the techniques needed in your shop as well as incorporating examples of your processes and products. Such a focused approach means your people get more out of the seminar in less time, which translates into a greater return on your training investment.

The most common compliment we receive on our seminars is that they are taught on the "do-it" level. We make extensive use of the case-study method of teaching, presenting only a minimum amount of theory, and then showing numerous examples of how to apply this knowledge on your production floor. By the way, all our seminars are presented live, no videos are used.

Being aware of the need to minimize disruptions to production schedules, our on-site courses can be structured in numerous ways to best accommodate your particular situation. For example, a course can be divided into half-day sessions, rather than full day, or we can train on the second or third shifts. We can also train one group during the last half of the first shift and a second group during the first half of the second shift. We will provide training when and where you want it to achieve the results you desire.

We impose no limitation on the number of people you schedule to attend one of our on-site seminars (we charge on a daily basis, not by the number of attendees). From past experience, we recommend 15 to 20 per class, but with an appropriate training facility, we can handle many more. You are welcome to include people from other divisions of your company, from your key suppliers, or even from your customers. By encouraging class participants to ask questions at any time, we adjust the pace of presentation to best fit the individual needs of each audience.

After each day's class, the IQI instructor is available for touring your facility and offering implementation suggestions. Even after the training is completed, we encourage the class participants to call us with your questions. You can also have one of our consultants return to your plant at various time intervals to work directly with the plant improvement teams. These return visits help promote new ideas and energize the implementation process by: identifying gaps that need attention; providing recommendations on how to improve the implementation efforts; and improving interpretation skills and team effectiveness.

The fees for presenting an on-site seminar vary based on the type of seminar and the number of days required. For a detailed quote and a list of available dates, please contact IQI.

Basic SPC - One Day

A sound understanding of traditional SPC methods is the first step in the journey to continuous quality improvement. Learning how to properly collect process data, construct and interpret control charts, and then correctly determine process capability is required to complete in today's worldwide marketplace.

• Examining Process Data
• Histograms
• Process average
• Process range
• Process standard deciation

• Subgroup Statistics
• Subgroup average
• Subgroup range
• Subgroup standard deviation

• Control Charts for Variable Data
• IX & MR chart
• X-bar, R-chart

• Control Charts for Attribute Data
• c, u, np, and p charts
• Flowchart for selecting the proper chart
• Alternatives to attribute charts

• Interpretation of Control Charts
• Sampling plans (frequency and size)
• Points outside of control limits
• Time-related changes; runs, trends, and cycles
• When to recalculate control limits

• Difference between Control Limits and Specification Limits
• Introduction to Process Capability
• Cp index

Everyone involved with process-improvement activities, especially those on the shop floor.

Only a knowledge of basic algebra is needed for this course. Because several examples will be worked out in class, attendees should bring a calculator.

Short-Run SPC - Two Days

Having trouble implementing SPC because lot sizes are small? Frustrated because customers want charts, but you can't get a traditional control chart to work on a high-mix/low-volume process? There is a very good reason why you're so upset -- traditional charts are just not designed to work with short production runs!

Regular control charts require long continuous runs of a single product on a given process while job shops and companies practicing lean manufacturing often have short runs of many different products, all run on the same process. This creates three significant problems for traditional control charts.

First: Because different part numbers have different nominal dimensions (or amounts of variation due to dissimilar material or tooling), a separate chart is needed for each part number. This means hundreds, if not thousands, of charts at each operation. Such a mountain of paper is certainly not a welcomed sight in shops striving to become a paperless factory. You can easily identify companies using this one-chart-per-part-number approach when you see a filing cabinet next to every machine. Their operators spend more time filing paper than they do running the machine!

Second: Let's assume you decide to bite the bullet and keep a separate chart for each part number (your customer demands SPC documentation!). Unfortunately, the run for a given part number is often over long before control limits can be calculated. Since proper chart interpretation cannot be done without control limits, the chart provides very few benefits for the operator. With little value added, this charting approach is soon abandoned. Supervisors then throw their hands up and grumble, "We tried that SPC stuff, but it just didn't work here!"

Third: There is another problem with keeping separate charts. When the performance data for a process is divided among several charts, operators will have a difficult time detecting any time-related process changes (runs, trends, cycles) that occur over two or more part numbers. Because the goal is process control, not part-number control, traditional control charts fail us again.

The innovative methods covered in this exceptional seminar solve all of the above three problems. By utilizing special data transformation formulas, our unique short-run charts allow an operator to plot all the different part numbers run on his or her process on the same control chart. With only one chart per process, the amount of paperwork required to monitor process performance is greatly reduced.

Now that all process data is on a single chart, we are able to determine control limits faster (with some of our charts, as soon as the operator starts charting). With control limits on the chart, operators are able to interpret the chart in a real-time mode, while the parts are still being run, not after the run is over. And when all process data is plotted on the same chart, interpretation of time-related changes (runs, trend, or cycles) happening over several part numbers becomes possible. Operators are now monitoring process control as well as part-number control.

During this interesting two-day seminar (see course outline below), we introduce 48 different charts for short runs. Many of these are included in the U.S. Army document on short runs (written for defense subcontractors) as well as in our reference handbook for short-run SPC (written for all quality practitioners). To illustrate the correct application and interpretation of our short-run charts, many case studies from numerous industries are discussed in detail.

We also include; a flowchart for selecting the proper short-run chart for a given process, information on choosing the proper subgroup size, how to determine an effective sampling frequency, and when to recalculate control limits.

Session 1: Review of Traditional SPC

This session briefly reviews the theory and concepts of the traditional X-bar, R chart, derives the control limit formulas and explains the difference between population parameters and sample statistics. The distribution of subgroup averages is developed and its relationship to the process output is discussed.

• The normal distribution
• Population versus sample
• Distribution of sample averages
• Central limit theorem
• Specification versus control limits

The objective of this first session is to develop a common language and understanding of traditional SPC concepts since many of these same ideas are used in the derivation of our new short-run SPC methods explained in the remainder of the seminar.

Session 2: Why Control Charts Work

In this session, a machining example illustrates how control charts deliver more information about process changes than 100% inspection, demonstrating the benefits of control charting to shop-floor personnel.

• Inspection method
• Risks of inspection
• Control chart method
• Benefits of charts

The purpose of this session is to show that 100% inspection is not the answer to monitoring the quality of a process run in a short run or just-in-time (JIT) mode. The use of statistics improves the operator's ability to understand and control his process at minimum cost.

Session 3: Why Traditional SPC Won't Work with Short Runs

Traditional SPC works best with one long run of the same part number on a given process. However, job shops and JIT require many short runs of different part numbers to be run on a single process, with each part number usually having a very small lot size. This session gives the definition of a short production run and demonstrates why traditional charting methods cannot be used effectively for short runs. The drawbacks of using other common methods to monitor short production runs are also explained.

• Definition of a 'short run'
• Risks of first-piece inspection
• A chart for every part number
• Lack of data for control limits
• Inability to detect time-related changes

After this session, class participants understand why current methods for monitoring a process run under short runs do not work well and will realize the need for developing alternative charting methods.

Session 4: Development of the Nominal X-bar, R Chart

A special adaptation of the traditional X-bar, R chart involving the coding of piece measurements as a deviation from the nominal print specification permits the data from all part numbers run on a process to be monitored on just one chart. This method saves time, allows control limits to be determined sooner, and detects time related changes in a process.

• Derivation of chart
• Coding sample data
• Control limit formulas
• Simplifying calculations
• Chart interpretation
• Assumptions for chart

After this session, participants know how to apply this innovative chart in short run situations, code and plot data, calculate the control limits and finally, interpret the chart so process improvements can be made. A case study on bending sheet metal on a press brake in a job shop and another on forming the threads of fasteners in the aerospace industry highlight how this new technique is successfully used in conjunction with short runs.

Session 5:The Target X-bar, R Chart

A major assumption of the Nominal chart requires nominal to be "best" as the process average for all part numbers. If this is not true, even for just one part number, then the Target chart, which uses the desired alternative value "the 'target average'" to code the data, must be employed.

• When is nominal not best?
• When nominal is undefined
• Derivation of the Target chart
• Data coding methods
• Control limit formulas
• How to choose a target average
• Chart assumptions

During this session, participants learn the advantages of this intriguing short run chart and four methods of selecting the proper target average when nominal is not best. An example involving a JIT cell producing printed circuit boards is presented.

Session 6: The Short Run X-bar, R Chart

The Target chart assumes the variation (process spread) of all part numbers plotted on the chart are similar. If this condition is not met, then the Short Run chart, which codes the sample data for differences in both average and variation, must be used.

• Derivation of Short Run R chart
• Derivation of Short Run X-bar chart
• Re scaling the subgroup statistics
• Control limit formulas
• Selecting target values
• Plotting different characteristics on the same chart
• Chart follows part through plant

In this session, participants gain an appreciation for the power and versatility of this revolutionary chart through several case studies: automotive brake linings produced in a job shop environment; shafts ground in a JIT cell; holes drilled in a lean operation; parts formed in a plasma cutting operation in the defense industry.

Session 7: Traditional IX & MR Chart

This session briefly reviews the traditional Individual X and Moving Range (IX & MR) chart, which has a subgroup size of 1. In addition to the assumptions associated with this chart, its benefits and drawbacks are discussed.

• Examples of application
• Moving range calculation
• Control limit formulas
• Benefits and drawbacks
• Normality assumption

Participants learn in which situations this chart can be applied, understand its limitations and participate in two case studies from the aerospace industry.

Session 8: The Nominal (and Target) IX & MR Chart

These charts handle different part numbers with lot sizes as small as just one piece (which is the ideal lot size for JIT). The IX plot points are calculated as either a deviation from the nominal print specification or from a target average. A variation of this chart, the Nominal Moving Average and Moving Range chart, is also mentioned.

• Data coding methods
• Control limit formulas
• Review of assumptions
• Three methods for verifying normality

At the end of this session, participants know how to plot different part numbers run over a given process on the same chart, even where lot sizes are extremely small. A case study concerning vinyl extrusion explains how this special short run chart can help improve quality when the production of upstream operations must be balanced with final assembly.

Session 9: The Short Run IX & MR Chart

This chart is designed for processes where lot sizes are extremely small and the standard deviation of different part numbers varies significantly. Constant control limits can be drawn on the chart before the first sample is plotted. A useful variation of this chart (the Short Run Moving Average and Moving Range chart) is also covered.

• Determination of target values
• Derivation of plot points
• Derivation of control limits
• Chart assumptions

Seminar participants gain an appreciation in this session for the quick response of this unique chart with extremely small lot sizes. In an informative case study, participants learn how to apply this method to stamping operations for electronic control panels run in a JIT mode.

Session 10: Traditional Attribute Charts

A brief review of traditional attribute charts is given, discussing the difference between attribute and variable data as well as the two major distributions for attribute data (binomial and Poisson). The differences between nonconformities (defects) and nonconforming units (defectives) are explained.

• np chart
• p chart
• c chart
• u chart

Participants learn how to construct and interpret the various attribute charts as well as distinguish when to apply each one in a long-run situation. Case studies are given for assembling printed circuit boards.

Session 11: Short-Run Attribute Charts

In this session, the proper data-coding formulas are derived for modifying traditional attribute charts to handle short production runs. Different part numbers (with different defect/defective rates) run on the same process can now be plotted on the same chart, with the same set of control limits.

• Short Run c chart
• Short Run u chart
• Short Run np chart
• Short Run p chart
• Control limit formulas
• Determining target values

Participants learn the construction, interpretation and assumptions of these new attribute charts for short production runs. Several examples of successful high-mix/low-volume applications are studied: wiring harness assembly defects, airplane seat assembly problems, transmission housing porosity, soldering defects on printed circuit boards, final test rejects in the assembly of electronic components, paint defects.

A detailed glossary defining all the new SPC terms covered in this workshop is included at the end of the training manual.

Everyone involved with process-improvement activities in job shops or companies practicing lean manufacturing.

Attendees should know how to construct an X-bar, R chart. As several examples will be worked out in class, attendees should bring a calculator.

Measuring Process Capability - Two Days

Once a process is in control (see our Basic SPC and Short-Run SPC seminars), it becomes very important to determine how capable it is of meeting the customer's requirements. Unfortunately, there is much confusion about how these studies should be conducted, and if done correctly, how the results should be interpreted and reported.

To end all this confusion, we created a seminar based on Davis Bothe's best-selling book, Measuring Process Capability. This seminar is usually presented over two days, but can be extended into three days to include advanced topics such as determining the capability of a process where: the target for the process average is not the middle of the tolerance; autocorrelated data is generated; the feature is hole location; the output experiences tool wear; there is variation in the process average from one setup to another; the product has within-piece variation; the feature has a unilateral specification; and the feature has no specifications. An overview of several new capability indices is also included.

Session 1: Importance of Process Stability

This initial session emphasizes the necessity of first stabilizing a process output before attempting to measure its capability. Data collected from an unstable process provide incorrect assessments of process performance and misallocation of valuable resources. A review of variable data control charts is given along with the definition of process capability.

• Definition of stability
• Review of control charts (X-bar, R, X-bar, S, and IX & MR)
• Difference between specification limits and control limits
• Definition of process capability

After this session, workshop participants understand how control charts demonstrate stability, the difference between variable data and attribute data, the difference between specification and control limits and what are the minimum requirements for achieving process capability.

Session 2: Estimating Process Parameters

Process parameters (average and standard deviation) of a process output are needed to estimate process capability. This session shows how these values are correctly estimated from control chart data. The difference between "short term" (within-subgroup) and "long term" (overall) variation is clarified.

• Estimating the process average
• Estimating the short term process standard deviation
• Estimating the long term process standard deviation

At the end of this session, participants know how to estimate the process average and standard deviation from either an X-bar, R, X-bar, S, or IX & MR chart. They also understand when and why to use the short term standard deviation or the long term standard deviation. Case studies are presented for manufacturing of cardboard boxes, particle contamination of a cleaning bath for printed circuit boards, machining metal shafts, and grinding automotive exhaust valves.

Session 3: Measuring Potential Capability

Many methods exist for measuring process capability. This session introduces those metrics designed for quantifying potential capability, the ability of a process to produce conforming product assuming that its output can be centered at the middle of the tolerance. The formula for each metric is presented along with a detailed explanation of proper interpretation, critical assumptions, major benefits and drawbacks.

• Potential versus performance capability
• Short term capability measures, Cr and Cp
• Long term capability measures, Pr and Pp

Advantages, disadvantages, and assumptions of all measures are covered. Some of the examples involve: size of holes drilled into printed circuit boards, length of electrical connectors, and dimensions of compact disks.

Session 4: Measuring Performance Capability

Most manufacturing processes are not centered at the middle of the tolerance. This session describes all of the more commonly used metrics in the aerospace, automotive, electronics, defense, and medical industries for measuring performance capability, how well the process is actually doing given its current average.

• Incorporating the process average, Zlsl, Zusl, Zmin
• Short term capability indices, Cpl, Cpu, Cpk, C*pk
• Long term capability indices, Ppl, Ppu, Ppk, P*pk
• Computing yield rates with ppm (parts per million)
• Taguchi loss function and the Cpm index

Participants leave this session with an appreciation for the variety of capability metrics and the unique benefits of each. They learn how to rate a process based on capability, determine its safety margin, estimate the percentage of nonconforming product, and choose between competing suppliers based on their quality levels. Capability metrics (C*pk and P*pk) are given for cases where the target for the process average is not the middle of the tolerance. Examples demonstrating proper application of these measures include plating of silicon wafers, thickness of extruded parts, output voltage of audio amplifiers, and computer storage devices.

Session 5: Capability Measures for Non normal Distributions

Almost all capability measures assume the process output has a normal distribution. However, many processes do not. This crucial session demonstrates three methods of checking for normality and then shows two different approaches for calculating capability when the process output distribution is non normal.

• Normal probability paper and histograms
• Simulation of process changes on capability
• Goodness of fit tests
• ISO 9000 capability procedure and the ppm index

In this informative session, participants learn how to verify normality and measure process capability when this assumption is not met. The techniques are illustrated with detailed case studies concerning grinding metal shafts, milling metal blocks, surface finish of cam rollers, hardness of aluminum forgings and drilling holes for a missile's inertial guidance system.

Session 6: Capability Measures for Attribute Data

Many companies using attribute data control charts also need to ascertain the capability of a process to satisfy customer requirements. This session reviews the four most popular attribute data charts, then demonstrates how to estimate Ppk and ppm for each type. A discussion on the difference between defects per unit (dpu) vs. defects per opportunity (dpo) is included.

• Review of c, u, np and p charts
• Estimating capability from binomial charts (p and np)
• Estimating capability from Poisson charts (c and u)
• Capability based on dpu vs. capability based on dpo
• Determination of capability goals

This session teaches how to measure capability for attribute based data with examples taken from the automotive (water pumps, porous castings, cracked pistons), electronics (fiber optics cable, printed circuit boards, copying machines) and medical industries.

Session 7: Machine Capability Studies

Quite often, capability must be determined for a new piece of equipment. In general, these are labeled machine capability studies and are extremely important since they assure only equipment consistent with a company's capability goals is purchased. This outstanding session also covers how to conduct a machine capability study, set equipment capability goals, and remove the effects of gage variation from machine capability estimates.

• Isolating variation generated by the machine
• Control chart approach to measuring machine capability
• Data collection form for a machine capability study
• Sequential S test for machine capability
• Establishing machine capability goals
• Removing gage variation from capability estimates

Participants learn steps for correctly undertaking and evaluating a machine capability study using either the traditional control chart approach or the newly developed sequential S test approach which can greatly minimize the number of test pieces, test time, and overall study costs.

Session 8: Combining Capability Measures

There are numerous occasions when it is desirable to measure capability for a combination of outputs or output features. This unique session provides answers to two tough questions: first, what is the capability of a process whose output consists of a mixture of many process streams (e.g., multiple cavities of a die, multiple spindles of a boring machine, several assembly lines) where each stream has a different average and/or standard deviation? We usually measure the capability of each process stream separately, but the customer receives the combined output of all streams. What level of capability does the customer get?

The second tough question: what is the capability of a machine to produce a part which has several characteristics, where each characteristic has a different average, a different standard deviation, and different specification limits? The capability of each part feature is typically assessed separately, but customers only consider the part to be acceptable if all its features are within their respective tolerances. The concepts covered in this session can be extended to measure the capability of an entire machining (or assembly) line.

• Average Cpk for multiple process streams
• Product Cpk for multiple product characteristics
• Normalized Cpk for standardizing capability estimates

The objective of this session is to impart information concerning three new powerful measures of capability for complex processes, either multiple process streams or multiple characteristics (note that these original methods are not available in any other textbook or workshop). These innovative techniques are clarified through examples of applications for plastic injection molding, CNC machining, multiple spindle machines, manufacturing toothbrushes, and assembling automotive alternators.

Session 9: Six Sigma Philosophy

In recent years, there has been much discussion concerning the six sigma quality philosophy which now dominates the electronics industry. Unfortunately, there is also considerable confusion surrounding this complex topic. This session presents the basic tenets of this novel philosophy as well as explaining all associated capability metrics.

• Definition of six sigma quality
• Reasons for the 1.5 sigma shift
• Derivation of the 3.4 ppm requirement
• Strategies for achieving six sigma quality

Participants become aware of how this profound philosophy pertains to all aspects of new product development including product design, internal manufacturing, and external supplier relations.

Personnel charged with conducting process capability studies, those who generate capability reports, and those who analyze those reports. This is also an excellent course for supplier quality engineers who must evaluate the capability of their suppliers.

Attendees should know how to construct an X-bar, R chart. Because several examples will be worked out in class, attendees should bring a calculator.

Statistical Problem Solving - Two Days

If your process is in control but not capable of meeting the customer's quality requirements, you must improve its capability. This type of improvement involves making changes to the process in order to shift its average, reduce its variation, or a combination of both. Often, these kinds of problems are assigned to a company green belt or black belt. However, if your organization doesn't have such personnel, or if they are already overworked, you must assign other people to work on process improvement. But first, they will need the proper training.

Our trademarked DOT*STARstrategy (a structured approach for process improvement) unleashes the problem-solving creativity of shop-floor personnel. Teamed up with several basic statistical techniques (that everyone can easily learn and correctly apply), this proven strategy has been helping our clients solve tough chronic quality problems for well over 25 years (five years longer than the six sigma DMAIC strategy has been around!). So why not provide your people with the time-tested skills and tools needed to achieve continuous process improvement? Based on Davis Bothe's book, Reducing Process Variation, this course includes an abundance of case studies to demonstrate how the lessons learned are correctly applied to solving real-life manufacturing and assembly problems.

Session 1: Overview of Variation Reduction Strategies

Several philosophies for solving quality problems by reducing process variation in manufacturing companies are discussed in this opening session. Differences between problem solving for sporadic versus chronic problems are explained. The importance of team work is emphasized as well as how to properly organize variation reduction teams.

• Dr. Juran's breakthrough concept
• Deming's PDCA cycle
• Six Sigma's DMAIC approach
• IQI's DOT*STAR strategy

There are two main objectives for this first session; the first is developing a common language and understanding of problem solving concepts. The second is outlining the seven steps of the powerful DOT*STAR variation reduction strategy which are then described in the remainder of this workshop.

Session 2: Step 1 Define the Problem (Identify the Opportunity)

The first step of any variation reduction effort involves defining the scope of the problem. The importance of having a precise and specific problem definition is stressed during this session.

• How to avoid working on too big a problem
• Identify the who, what, where, when, how of the problem
• How to maximize investment in resolving problems

The purpose of this session is to show that a clear problem definition is required before attempting to solve any problem since this maximizes the chance of finding a suitable solution at minimum cost. Specific techniques for helping identify a precise problem include traditional Pareto analysis as well as nested Pareto charts, visual Pareto charts for real time shop floor control, pie charts, three dimensional Pareto diagrams and weighted Pareto analysis for incorporating multiple factors into the problem selection process.

Examples discussed involve metal painting, foundry scrap, plant injuries, electronic equipment assembly and missile assembly defects are discussed.

Session 3: Step 2 Observe the Process (Collect Background Information)

Once the problem is adequately defined, a focused background investigation can be conducted to gather specific information relating to the problem area. This session reveals several compact statistical techniques for reviewing existing process data as well as collecting and analyzing new information to help generate clues about potential causes of process variation.

• Pros and cons of using historical data
• Process flow diagrams
• Check sheets, concentration diagrams, matrix charts
• Histograms, stem and leaf plots
• Run charts for indicating time related process changes
• Multi vari charts for identifying largest family of variation
• Quantifying attribute data

The purpose of this session is to have class participants understand how the proper planning of data collection can greatly increase the information content of each observation, thus significantly reducing the overall cost (and time) of conducting a background study. The multi vari technique can quickly identify clues about which family of variation is most important (within piece, piece to piece, time to time, or process stream to process stream).

Case studies are presented for winding coils, machining bar stock, grinding gear blanks, manufacturing exhaust valves, assembling computer monitors, soldering circuit boards, defects in castings, wave soldering, remanufacturing of fuel injectors, mixing chemical compounds, surface finish of clutch plates, machining cylinder heads, turning shafts, installing bushings into castings, softness of foam seat cushions, hardness of brake linings, gaps around car doors, weight of bread, sensitivity of temperature sensors, torque of truck lug nuts, inserting bristles into toothbrushes, drilling holes in printed circuit boards.

Session 4: Step 3 Think of Potential Causes

With all the clues generated from observing the process in step 2, the variation reduction team must now think of specific potential causes of the process variation. This session explains several techniques for producing the largest number of potential causes and how to organize these for evaluation in step 4.

• 5Ws and 1H, five whys
• Brainstorming
• Cause and effect diagrams
• Fukuda's CEDAC diagram
• Process C/E diagram
• Force-field analysis

Participants learn several detailed methods for conducting a successful idea generation session and then constructing diagrams to help organize all these potential causes of process variation. Examples covered include gasket leaks, cracks in contact lenses, and reducing order-processing time for a warehouse.

Session 5: Step 4 Select the Most Likely Cause

After all potential causes are listed and thoroughly evaluated, one needs to be chosen and tested to verify it is the correct cause of the problem. This session introduces methods for increasing the likelihood of selecting the correct cause as soon as possible, thus minimizing the amount of testing required before the correct cause is found.

• Executive choice
• Simple vote
• Multi voting
• Nominal group technique
• Decision making by consensus

During this session, participants learn the advantages of the multi voting method over all others for promoting teamwork, utilizing the knowledge, experience and judgment of the participants, and minimizing the time required for choosing the correct cause.

Session 6: Step 5 Test to Verify the Chosen Cause

After a potential cause is selected in Step 4, it must be tested to verify it is the correct cause. This session reveals several easy to use techniques (requiring only minimum sample sizes) for quickly confirming a relationship between the suspected cause and the known problem.

• Construction of a scatter diagram
• Proper interpretation and curvilinear relationships
• Isograms (scatter diagrams for gage correlation)
• How to properly randomize the order of testing
• Comparison testing for piece to piece variation
• Component swapping procedures for large assemblies

No problem can be fixed until the true cause is found. In this very important session, participants gain an appreciation for the power and versatility of the presented techniques for identifying the true cause of process variation. In addition, they learn how to determine the critical value for the major cause of process variation. Also, they will understand when, where and why randomization must be done to prevent errors due to improperly taken measurements.

There are examples relating to bushing installation, monitor production, foundry casting defects, chemical batch processing, seal leaks, hardness of rubber seals, porosity in castings, calibration of fuel injectors, contaminated fuel, machining connecting rods, output of water pumps, watch assembly, tight oil pumps, seized diesel fuel injectors, leaking oil seals and sealing of plastic bags.

Session 7: Step 6 Act to Develop and Implement a Solution

Once the correct cause is identified, an acceptable solution resolving the quality problem must be determined. This session introduces various methods for generating potential solutions along with procedures for evaluating these solutions to select the most cost effective then concludes by showing how to predict any possible problems associated with its implementation and minimize their impact.

• Force field analysis
• PMI (Plus Minus Indifferent) analysis
• Solution FMEA
• Pilot study
• Gantt chart and arrow diagrams
• Poka yoke (mistake proofing)

At the conclusion of this session, participants know how a team should evaluate solutions, select the most promising and make sure its implementation goes smoothly for all concerned. Suggestions are listed for presenting the results to top management. Several poka yoke examples are given. A case study is presented for the class to work on involving the evaluation of a potential solution that selects a different supplier.

Session 8: Step 7 Review Results and Revise as Necessary

Since the manufacturing world is constantly changing, there is no guarantee variation will remain reduced after the solution is implemented. This session presents techniques for continually monitoring the process so changes are immediately identified and corrective action implemented before variation can increase and become a problem once again.

• Control charts
• Standardization
• Audits
• Total preventive maintenance
• Kaizen (continuous improvement)

After explaining the above techniques, a review of the entire DOT*STAR strategy is given with several case studies so seminar participants learn how to solve quality problems in their own companies by reducing process variation. They will understand and appreciate the need for continuous process improvement in order to have their company become, and remain, a world-class enterprise.

Everyone involved with process improvement activities in manufacturing companies: quality engineers, manufacturing engineers, green belts, quality technicians, SPC facilitators, shop-floor supervisors, quality managers, manufacturing managers, maintenance personnel, machine operators, assembly workers, and inspectors. This would also be a terrific course for supplier quality engineers who are charged with helping improve the quality of their suppliers' products.

None.

Gage R & R Studies (MSA) - One Day

Successful process improvement occurs when decisions are based on facts. These facts are most often derived from measurements made on some aspect of a process. If the measurement system is not providing reliable and accurate measurements, our decisions could be faulty and our problem-solving efforts could be ineffective.

In this one-day workshop, attendees will learn how to analyze a measurement system according to the guidelines presented in the third edition of AIAG's manual, Measurement Systems Analysis. By working several examples, attendees will also discover how to assess and interpret gage bias as well as gage repeatability and reproducibility (for both variable gaging and attribute gaging).

• Important Measurement System Definitions
• Reference value
• Accuracy
• Bias
• Stability
• Linearity

• Quantifying Measurement Variation
• Repeatability
• Consistency
• Uniformity
• Undertaking a Measurement System Study
• Assessing Gage Linearity and Uniformity
• Measuring Reproducibility and Gage R&R
• Assessing Gage R&R with the Average and Range Method
• Charting and analyzing the results
• Assessing Gage R&R with the ANOVA Method
• Estimating Gage R&R
• Interpreting the outcome
• Determining the number of distinct categories
• Using Isograms
• Checking the agreement between two gaging systems
• Attribute Measurement System Study
• Using the short form
• Benefits of the long form (the analytic method)
• Interpreting the gage performance curve
• Quantifying the defects with a rating scale

Anyone who might either conduct a gage R&R study or need to know how to properly interpret the results. This course would also be perfect for supplier quality engineers who must check the reliability of their suppliers' measurement systems.

Attendees should know how to construct an X-bar, R chart. As several examples will be worked out in class, attendees should bring a calculator.

Analysis of Means (ANOM) - One Day

The analysis of means (ANOM) is a powerful, yet easy-to-apply, graphical technique for statistically comparing the performance of several supposedly identical processes, e.g., the defect rates among three assembly lines, the averages for each of the four cavities in a die, the standard deviations for each of five machines, or the percentage nonconforming among six assembly workers. This seminar explains a step-by-step procedure for creating ANOM charts as well as how to properly interpret them to provide answers to a variety of important industrial problems. ANOM is very similar to ANOVA, but its calculations are much less complicated (they can be easily done by hand, no software required). In addition, because ANOM is a graphical method, its interpretation is also much more straightforward than a comparable ANOVA study on the same data set (even your boss will understand it!).

Several case studies demonstrate how to quickly uncover significant differences in: the percentage nonconforming between five cavities in a die; defect rates between nine workers; reject rates between five assembly lines; the average fill volumes of eight filling heads; and the standard deviations of three machines. Course participants will learn a proven technique that helps them better understand variation in the performance of their processes.

• Using ANOM with Attribute Data
• Percentage nonconforming
• Number of nonconformities
• Matrix diagrams and response diagrams
• Analyzing interactions
• Testing for Differences in Standard Deviations
• Comparing several options
• Testing for Differences in Averages
• Comparing several options
• Matrix diagrams
• Analyzing interactions

Everyone involved with process improvement activities in manufacturing companies: quality engineers, manufacturing engineers, green belts, black belts, quality technicians, quality managers, manufacturing managers, maintenance personnel, machine operators, assemblers, and inspectors. This would also be a terrific course for supplier quality engineers who are charged with helping improve the quality of their suppliers' products.

Attendees should know how to construct an X-bar, R chart. As several examples will be worked out in class, attendees should bring a calculator.

Verifying Process Improvements - Two Days

After an idea for improving a process is selected, it should be tested to confirm it really does improve the process in the manner expected. Far too often, the cause believed to be a major source of process variation turns out to be only a minor contributor. How can you efficiently determine whether or not your chosen cause is indeed a major one and avoid the embarrassment (and costs) of implementing a solution that doesn't work?

This intense two-day seminar explains numerous statistical techniques for verifying process improvements. Some are fairly basic, like control charts, while others are more complex, such as hypothesis testing. Learn how to design customized test plans, set experimental risk levels, determine the minimum sample size for conducting verification tests, compute the power of a test, calculate confidence intervals for process parameters, and correctly interpret the test results.

• Basic Verification Tests
• Planning a verification test
• Verifying a shift in the average with a control chart
• Additional rules for identifying out-of-control conditions
• Verifying a reduction in variation with a control chart
• Special run rules for range charts
• Confirming changes with attribute-data control charts
• Confirming changes with "success" testing

Statistical Hypothesis Testing

• Concept of hypothesis testing
• Stating the hypotheses
• Relationship between the a (alpha) and b (beta) risks
• Effect of sample size on the a and b risks
• Three test assumptions
• Testing for a reduction in variation
• Comparing the process standard deviation to a given value
• Computing a confidence interval
• Creating a power curve
• Filtering out minor reductions in the standard deviation
• Comparing standard deviations from two processes
• Filtering out minor reductions
• Comparing the standard deviations from three or more processes
• Testing for a shift in the process average
• Verifying an increase in the process average
• Verifying a decrease in the process average
• Filtering out minor shifts in the process average
• Comparing the averages from two processes
• Testing for a minimum difference between averages
• The paired comparison test
• Testing for a reduction in the percentage nonconforming
• Filtering out small reductions
• Comparing the percentage nonconforming between two processes
• Comparing the percentage nonconforming from three or more processes
• Derivations

Personnel who need to verify that their ideas for process-improvement will be effective: quality engineers, manufacturing engineers, black belts, product engineers, process engineers, quality managers, and manufacturing managers. This seminar would also help supplier quality engineers verify the effectiveness of the improvements promised by their suppliers.

Attendees should know how to construct an X-bar, R chart. As several examples will be worked out in class, attendees need to bring a calculator.