Product & Process Assessment

1. Product & Process Assessment Six Sigma Foundations Continuous Improvement Training Six Sigma Simplicity

2. Key Learning Points • Traditional Metrics vs. CI • DPU vs. DPMO • RTY vs. Hidden Factory

3. Agenda Objectives of This Module • Introduction to defects • First time yield (FTY) vs. rolled throughput yield (RTY) • COPQ vs. yield • The hidden factory • Defects per opportunity metric • Complexity explained • Defects per unit metric • The basic Binomial model • The basic Poisson model • The application of defect data in process improvement efforts • Project metrics

4. First Time (End of Line) Yield by Week 100 98 96 Weekly Yield (%) 94 92 90 Wk 1 Wk 2 Wk 3 Wk 4 Wk 5 Wk 6 Wk 7 Wk 8 Wk 9 Wk 10 Wk 11 Wk 12 Wk 13 Wk 14 Wk 15 Traditional Method of Project Selection How does your organization identify poor quality products? P FTY = * 100% U Where: FTY = First Time Yield (test yield) P = Number of units that pass test U = Number of units tested Will the first time yield be correlated to other major business metrics? What will the test yield be next week?

5. The Traditional Methodof Project Selection • Will traditional yield (end-of-line test yield) calculations correlate to business metrics? • End-of-line test yield has traditionally been considered a good predictor of profit margins and scrap rates. However, it rarely does a good job at either. Why not? What is missing? What’s the problem with classical yield calculations? • As managers and Black Belts, we shouldn’t select projects based on the FTY of a product. Expected Relationships

6. Defects vs. Defectives • Defects: • Countable failures that are associated with a single unit. A single unit can be found to be defective, but it may have more than one defect. • Defectives: • Completed units that are classified as bad. The whole unit is said to be defective regardless of the number of defects it has. FTY = the number of non-defectives/the number of total units.

7. Hidden Factory Re-Work or Scrap Re-Work or Scrap Failure Analysis Failure Analysis Test Test Product Operation 1 Operation 2 The Hidden Factory • To analyze, re-work, and/or scrap potential product requires: • More manpower • Extra floor space • Longer cycle time • More raw material • More $$$$ • How big is your “hidden factory”? • What happens to cost as defects increase?

8. Defect-Based Cost Model DPU vs. COPQ • When we track individual defects rather than percent defective, we end up with a much better predictor of costs. • What constitutes a defect? 600 High 500 400 300 COPQ ($) 200 100 Low 0 1 2 3 4 0 DPU Hint: Go collect defect data!

9. Defects? Proportion = DefectsUnit Produced DPU = DPU vs. DPMO Collect Defect Data • Which metric do you need? DPU or DPMO? • In most cases, we end up converting defect data to a proportion as follows: Black Belt Project Metric What do I need to know? Management Metric Compare the quality level of non-identical parts, processes, or products. Model process efficiency or estimate the probability of producing defect-free parts. DPMO DPU • When we use defect data, we need to determine what number to put into the denominator of the equation above. ? = The number of units produced (or processed through the operation) ? = A measure of complexity (i.e., opportunities) DefectsTotal Opportunities DPMO = x 1M Project Selection Metric Benchmarking Metric Black Belt Project Metric

10. A Metric to Expose the Hidden Factory Rolled Throughput Yield (RTY) • The product of total throughput at each step in the process Definition • A measurement of yield which exposes the extent and location of scrap and rework • The proportion processed with no defects

11. 1000 units Scrap 4% - 40 units Y1=0.92 Rework 4% 920 units 960 units Scrap 9% - 86 units Y2=0.82 Rework 9% 754 units 874 units Scrap 8% - 70 units Y3=0.84 Rework 8% 633 units 804 units RTY=633/1000=0.633 First Pass Yield=804/1000=0.804 RTY=0.92x0.82x0.84=0.633 Rolled Throughput Yield (RTY) Hidden Factory

12. Cost of Hidden Factory • To analyze, to rework, and to scrap requires: • More raw material (Scrap / re-order) • Manpower (Unproductive Hours) • Floor space (Capacity) • Longer cycle time (DSO)

13. Internal Performance Internal View Customer View Identifies Result of Final Inspection Identifies Process Yield (Scrap) Identifies Extent & Location of COPQ (Our Opportunity) Test Yield = 84% First Pass Yield = 80% Rolled Throughput Yield = 63% Yield

14. The Greater Hidden Factory • Beyond the direct costs associated with finding and fixing defects, “Cost of Poor Quality” also includes: • The hidden cost of failing to meet customer expectations the first time • The hidden opportunity for increased efficiency • The hidden potential for higher profits • The hidden loss in market share • The hidden increase in total cycle times • For an average company, the cost of poor quality can be as high as 25% of annual sales • COPQ can exceed the profit margin • COPQ is our Opportunity!

15. Use DPMO Here Use DPU Here Remember: Strategic Black Belt Overview Measure Management Management Measure Review progress and modify Analyze Select a new project Analyze Goal: Y=f(x) Management Improve Control Management Control Improve Management

16. defects = = DPU Defects per unit units (Opps/unit) = 1,000,000 DPMO DPU * / = Defects / Total Opps * 1,000,000 DPU and DPMO Calculations • Black Belts should use the DPU (or PPM) metric to track their project performance. • Management should use the DPMO metric to select projects and conduct benchmark studies for dissimilar goods and services.

17. DPMO, Measures of Complexity • Product complexity • Number of parts • Number of functions • Process complexity • Number of attachments • Number of welds • Transactional complexity • Number of entries • Software complexity • Lines of code

18. DPMO, Measures of Complexity- continued • “Complexity” is a measure of how complicated a particular good or service is. Theoretically, it’s doubtful that we will ever be able to quantify complexity in an exacting manner. • If we assume that all characteristics are independent and mutually exclusive, we may say that “complexity” can be reasonably estimated by a simple count. This count is referred to as an “Opportunity Count”. • In terms of quality, each product and/or process characteristic represents a unique “opportunity” to either add or subtract value. • Remember, we only need to count opportunities if we want to estimate a sigma level for comparisons of goods and services that are not necessarily similar.

19. DPMO, Counting Opportunities • Non-value-added rules:No opportunity count should be applied to any operation that does not add value. • Transportation and storage of materials provide no opportunities. Deburring operations can also be considered. • Testing, inspection, gauging, etc., do not count. The product, in most cases, remains unchanged. An exception: An electrical tester where the tester is also used to program an EPROM. The product was altered and value was added. • Supplied components rules:Each supplied part provides one opportunity. • Supplied materials such as solder, machine oil, coolants, etc., do not count as supplied components. x 1M DefectsOpportunity DPMO =

20. DPMO, Counting Opportunities -cont. • Connections rules:Each “attachment” or “connection” counts as one. • If a device requires four bolts, there is an opportunity of fourone for each bolt connected. • A 60-pin integrated circuit, surface mount device, soldered to a printed circuit board counts as 60 connections. • A 16-pin dual in-line package with through-hole mounting counts as 16 joints. There is no double counting of joints  one for the top side and one for the bottom side is not correct. • Once you define an “opportunity,” you must institutionalize that definition to maintain consistency. x 1M DefectsOpportunity DPMO =

21. DPMO, Counting Opportunities -cont. • Machine shop equipment rules: • There is one opportunity count for each machined surface. • If one tool makes five separate cuts, the count is five opportunities. • When a hole is drilled and counter-bored, the count is two because there are two separate operations. • A hole that is drilled and honed because the drilling operation is not trusted to hit the dimension is only a count of one. The honing operation is re-work of the drilling operation. x 1M DefectsOpportunity DPMO =

22. DPMO, Counting Opportunities -cont. • Transactional process rules: • Filling out a form provides one opportunity per data-entry field, not one opportunity for each character. • One line of assembly equivalent code counts as one opportunity for software programs. • Sanity check rule: • “Will applying counts in these operations take my business in the direction it is intended to go?” • If counting each dimension adds no value, and increases cycle time then this type of count would be contrary to the company objective and would not provide an opportunity. • Once you define an “opportunity”, you must institutionalize that definition to maintain consistency. x 1M DefectsOpportunity DPMO =

23. DPMO Examples Application to a Measured Quantitative Parameter Part Specification: 1.240 ± .003 Spec Limit Probability of producing a “good” part = 0.97725 Probability of producing a “bad” part = 0.02275 Xbar = 1.241 S = 0.001 Measurement of a Product Characteristic If This Part Were A Supplied Part, It Would Count As One Opportunity. As Supplied Parts, At Least 2.275% Of Them Had Defects. Therefore, The DPMO = 0.02275*1,000,000 = 22750.

24. DPMO Examples -cont. Application to an Inspected Parameter • A circuit board has 800 solder joints and 200 components. • How many opportunities do we have? • Six defective joints and two defective components were found in this unit. • What is the DPMO? DPMO = dpu/opportunities/unit * 1,000,000 = (8/1)/(1,000/1) * 1,000,000 = 8,000

25. The DPU Metric • Suppose we have a unit with 10 components. Each component within the unit is a chance (or opportunity) for a defect to occur. Thus, each unit can contain up to 10 defects. • Note: This means you need to be able to track more than one defect per unit through your data-collection system. Defect Missing Part Opportunity for a Defect Fundamental Question:What Is the Likelihood Of Producing A Unit With Zero Defects?

26. DPUs from the Process Fundamental Question: Given these facts, what is the likelihood of producing a unit with zero defects? Production Run • Tally the number of defects within each unit. Based on this sample, calculate the probability of producing a zero-defect unit. How many units had: _____ Zero defects _____ One defect _____ Two defects _____ Three defects _____ Four defects _____ Five or more defects Question 1: How many total defects are observed? Question 2: What is the number of DPUs?

27. 60 60*10 = 0.1 or 10% 60 60*10 = 0.9 or 90% 1 - DPU Modeling Fundamental Question: Given these facts, what is the likelihood of producing a unit with zero defects? • Given: • 60 defects observed • 60 units processed • 10 opps per unit • The probability that any given opportunity will be a defect is: • The probability that any given opportunity will NOT be a defect is: • The probability that all 10 opportunities on a single unit will be defect free is: If we extend the concept to an infinite number of opportunities, all at a DPU of 1.0, we will approach the value of 0.368. 0.90 (10) = 0.3487 => 34.87%

28. Operation 3 RTY = 98% DPU = 0.02 Two Major Uses of Defect Data • Prediction of true factory yield (RTY) • Defect data can be used in an analysis to predict factory (or line) yield, as shown below. Data analysis for project scoping • Defect data is most commonly used to perform Pareto analysis on information to determine priorities for action items within the project planning cycle. The following slides demonstrate this idea. Operation 1 RTY = 99% DPU = 0.01 Operation 2 RTY = 96% DPU = 0.04 Final RTY = (0.99)*(0.96)*(0.98) = 0.93*100% = 93%

29. 12000 Ecoat Defects 9000 PPM 6000 3000 0 Scratches Cracked Light Other 12000 Scratch Defects Process 9000 PPM 6000 Supplier Other 3000 0 Operator Dropped Material Cuts Other Operator Dropped 8000 6000 4000 PPM 2000 0 Part Sticks to Rack Packaging Other Analysis of Defect Data The Three-Level Pareto Principle

30. Product & Process Assessment Six Sigma Foundations Continuous Improvement Training