Equipment, Setup and Quality at Source

Introduction

In Lean manufacturing and Six Sigma, equipment is not merely a production asset. It is part of the process capability system through which customer requirements are converted into physical output. A capable process requires reliable machines, repeatable setups, effective maintenance, safe work methods, competent operators, and controls that prevent defects from moving downstream.

This chapter integrates Total Productive Maintenance (TPM), Overall Equipment Effectiveness (OEE), autonomous and preventive maintenance, Single-Minute Exchange of Die (SMED), Poka-Yoke, source inspection, and first-pass yield. The central managerial principle is quality at source: the person or process creating the output should detect and prevent nonconformance before it is transferred to the next operation.

Equipment improvement should not be treated as an isolated maintenance project. Reliability affects availability; setup design affects flexibility; maintenance quality affects process capability; operator ownership affects response speed; and safety determines whether productivity gains are sustainable. A mature system therefore integrates maintenance, production, quality, engineering, industrial safety, and workforce development.

Learning Objectives

By the end of this chapter, students should be able to:

1. Explain the relationship between equipment reliability, maintainability, availability, safety, and quality.
2. Calculate and interpret OEE and its components.
3. Design autonomous and preventive maintenance activities.
4. Apply SMED to reduce setup time and distinguish internal from external work.
5. Select appropriate Poka-Yoke and source-inspection methods.
6. Calculate and interpret first-pass yield and identify hidden factory effects.
7. Develop an integrated equipment and quality-at-source improvement plan.
8. Evaluate the managerial, human, ethical, and technological implications of these systems.

1. Equipment as a Production-System Capability

A production process is capable only when it repeatedly produces conforming output under controlled conditions. Equipment contributes to this capability through five related dimensions:

• Reliability: the probability that equipment performs its required function for a specified time under stated conditions.
• Maintainability: the ease and speed with which equipment can be restored to operating condition.
• Availability: the proportion of planned time during which equipment is ready for productive operation.
• Capability: the ability of the equipment and process to meet specified performance and quality requirements.
• Safety: the ability to perform work without unacceptable risk to people, property, or the environment.

These dimensions interact. A machine that runs continuously but produces dimensional defects is not effective. A highly capable machine that frequently fails is not useful. A fast setup method that exposes workers to uncontrolled energy is unacceptable. Consequently, improvement decisions should consider the complete system rather than a single metric.

A useful distinction is between failure prevention and failure containment. Preventive maintenance, condition monitoring, robust design, and operator inspection seek to prevent failure. Poka-Yoke and source inspection limit the consequences when an abnormality nevertheless occurs. Both are necessary, but prevention is generally more economical and reliable than downstream detection.

2. Total Productive Maintenance

TPM is a company-wide approach for maximizing the productive life-cycle effectiveness of equipment through the participation of all relevant functions and, especially, the involvement of operators. It extends traditional maintenance from a specialist activity to a shared operating system.

Traditional maintenance departments often respond to breakdowns, while production departments operate equipment until problems occur. TPM changes this separation. Operators conduct routine cleaning, inspection, lubrication, tightening, and abnormality reporting; maintenance specialists handle technical restoration, planned interventions, predictive technologies, and engineering improvements; managers establish standards, priorities, and resources.

The commonly cited TPM pillars are:

1. Focused improvement: cross-functional elimination of chronic losses.
2. Autonomous maintenance: operator-led basic care and early abnormality detection.
3. Planned maintenance: preventive, predictive, and corrective maintenance based on equipment requirements.
4. Quality maintenance: control of equipment conditions that influence product quality.
5. Early equipment management: designing maintainable, safe, and capable equipment from the beginning.
6. Training and education: development of operator, technician, and managerial competence.
7. Safety, health, and environment: prevention of injury, exposure, and environmental harm.
8. TPM in administrative functions: reduction of information, approval, scheduling, and procurement losses.

TPM is not equivalent to asking operators to perform skilled repairs. It requires clear role boundaries, competence assessment, lockout/tagout controls, and escalation rules. Operator ownership means ownership of basic conditions and abnormality visibility, not unauthorized exposure to hazardous energy.

2.1 Autonomous Maintenance

Autonomous maintenance normally develops through a staged progression:

1. Initial cleaning and inspection.
2. Elimination of contamination sources and inaccessible areas.
3. Development of provisional standards for cleaning, lubrication, and tightening.
4. General inspection training.
5. Autonomous inspection standards.
6. Workplace organization and visual management.
7. Continuous self-management and improvement.

Cleaning is not merely cosmetic. It is an inspection method. Oil, dust, vibration, heat, unusual noise, loose fasteners, and leakage become visible when equipment is restored to a clean baseline. Operators should record abnormalities using tags, photographs, digital work orders, or escalation boards. Each abnormality should have an owner, due date, risk classification, and verification of closure.

A strong autonomous-maintenance standard answers four questions: what is checked, where it is checked, how it is checked, and how often it is checked. It should also define the reaction when the condition is outside the standard.

2.2 Preventive and Predictive Maintenance

Preventive maintenance is planned maintenance performed at predetermined intervals or according to prescribed criteria. Examples include replacing filters after a specified operating period, checking belt tension weekly, or calibrating a measuring instrument annually.

Predictive or condition-based maintenance uses evidence of equipment condition to determine when intervention is required. Examples include vibration analysis, thermography, oil analysis, motor-current analysis, ultrasound, and sensor-based monitoring. Predictive maintenance is most valuable when failure modes have measurable precursors and when the cost of monitoring is justified by avoided consequences.

Maintenance policies should be selected by failure mode, not by fashion. A useful decision logic is:

• If failure is age-related and consequences are significant, time-based replacement may be appropriate.
• If degradation is observable before functional failure, condition-based maintenance may be appropriate.
• If no effective preventive task exists and the consequence is tolerable, run-to-failure may be acceptable.
• If failure creates unacceptable safety, legal, or environmental risk, redesign or engineered protection is required.

Run-to-failure is therefore not automatically poor maintenance. It is a deliberate policy suitable for inexpensive, noncritical, easily replaceable items whose failure has limited consequences.

3. OEE and the Six Big Losses

Overall Equipment Effectiveness is a composite indicator of how effectively a piece of equipment produces conforming output during planned production time. The standard relationship is:

The three factors are:

Here, planned production time excludes periods when production is not scheduled; operating time is planned production time minus downtime; ideal cycle time is the theoretically fastest cycle for a conforming product; total pieces includes good and defective pieces; and good pieces are units meeting requirements without rework. All time quantities must use the same units.

These formulas assume that the measurement boundaries are defined consistently and that the ideal cycle time is realistic for the product and equipment. OEE is not the same as utilization. Utilization may include demand and scheduling effects, whereas OEE focuses on effectiveness during planned production time.

The six major losses traditionally associated with OEE are:

Worked Example: OEE

A machine is scheduled for 480 minutes. Planned breaks total 30 minutes, so planned production time is 450 minutes. Breakdowns and setup losses total 60 minutes, giving operating time of 390 minutes. The machine produces 780 pieces. The ideal cycle time is 0.45 minutes per piece, and 741 pieces are good.

Availability is:

Performance is:

Quality is:

Therefore:

The result does not identify the solution by itself. The loss tree indicates that availability, performance, and quality all require attention. Managers should avoid improving OEE by suppressing legitimate maintenance, extending shifts, or redefining quality. Such practices distort the measure and can increase risk.

OEE is best used as a diagnostic indicator at an appropriate equipment or value-stream level. Comparing unrelated machines without considering product mix, operating conditions, and measurement rules can create misleading conclusions. OEE should be paired with safety, delivery, cost, employee well-being, and process-capability measures.

4. Reliability, Maintainability, and Availability

For a repairable item under a simple steady-state model, inherent availability can be approximated by:

where MTBF is mean time between failures and MTTR is mean time to repair. Both must use the same time unit. The relationship assumes that failures and repairs are represented adequately by average values and that planned maintenance, logistics delays, and external constraints are excluded from the stated model. It should not be confused with operational availability, which may include waiting for parts, personnel, utilities, or production release.

For example, if MTBF = 100 hours and MTTR = 5 hours:

Improvement can therefore come from increasing MTBF through better design, lubrication, alignment, or operating conditions, or from reducing MTTR through accessibility, standard work, spares, training, and fault diagnosis. Reliability-centered maintenance adds a further question: what function is required, what can cause functional failure, what happens when it fails, and what task or redesign best manages the consequence?

Safety must be integrated into reliability decisions. A recurring failure that creates an entanglement hazard cannot be managed only by a faster repair. Engineering controls, guarding, interlocks, energy isolation, and verified safe states take priority. Maintenance performance must never reward unsafe speed.

5. Setup Reduction and SMED

Setup is the work required to change equipment from producing one product or condition to another. It includes preparation, removal, replacement, adjustment, trial production, inspection, and return to stable production. Long setups encourage large batches, increase inventory, reduce responsiveness, and hide quality problems.

SMED is a structured approach to reduce changeover time, often aiming for a changeover measured in single-digit minutes, although the appropriate target depends on the process. Its central distinction is:

• Internal setup: work that can be performed only when the equipment is stopped.
• External setup: work that can be performed while the equipment is producing.

The usual sequence is:

1. Observe and document the complete changeover.
2. Separate internal and external activities.
3. Convert internal work to external work wherever technically safe and feasible.
4. Streamline remaining internal work through quick clamps, standard locations, parallel work, elimination of adjustments, and visual confirmation.
5. Standardize and continuously improve the method.

A changeover study must include safety, quality verification, and restart stability. Reducing the time until the machine first produces a part is insufficient if the process requires many trial pieces before a conforming output is achieved.

Worked Setup Example

A baseline changeover consists of 12 activities totaling 74 minutes. Analysis shows that 18 minutes of material staging and tool preparation can be completed while the machine is still producing. The remaining internal work is reduced by 20% through quick clamps and parallel tasks.

The revised internal time is:

The changeover reduction is:

The percentage reduction is:

The result should be validated over several changeovers because individual observations may be affected by product mix, operator experience, tool condition, or unrecorded waiting.

6. Poka-Yoke and Quality at Source

Poka-Yoke means mistake-proofing: designing a product, process, fixture, software control, or information system so that an error is prevented or immediately detected. It does not mean blaming workers for mistakes. Human error is predictable; robust systems make the error difficult to create or easy to identify.

Poka-Yoke devices may be classified as:

• Prevention controls: make an incorrect action impossible, such as an asymmetric connector that cannot be inserted incorrectly.
• Detection controls: identify an error before the process proceeds, such as a presence sensor or torque verification.
• Warning controls: alert the operator through a light, sound, message, or physical signal. Warning controls are weaker because the process may continue despite the warning.

The strongest solution addresses the cause at the point of occurrence. A downstream inspection that detects a missing component is less desirable than a fixture that cannot cycle unless the component is present. However, Poka-Yoke must be validated. A bypassable sensor, poorly positioned alarm, or excessive nuisance-stop rate can lead to disabling, workarounds, and loss of trust.

A practical design procedure is:

1. Define the required function and critical-to-quality characteristic.
2. Identify possible errors and their effects.
3. Determine whether the error can be eliminated through product or process design.
4. Select prevention or detection at the earliest feasible point.
5. Test normal, abnormal, and boundary conditions.
6. Establish reaction, escalation, maintenance, and verification standards.

7. Source Inspection and First-Pass Yield

Source inspection verifies conditions before or at the point where value is added. It differs from end-of-line inspection because it seeks to prevent the creation or transfer of defects. Examples include checking material identity before loading, verifying tool offset before machining, confirming torque at assembly, and validating software or recipe selection before a batch begins.

Source inspection may be performed by the operator, an automated control, a sensor, or a preceding process. The inspection should be linked to an immediate reaction plan. “Stop, contain, inform, and correct” is generally more effective than recording a defect for later review while production continues.

First-pass yield (FPY) is the proportion of units that meet requirements the first time through a process, without rework, repair, or repeated processing:

For a process with 1,000 units entering and 930 units accepted without rework, FPY is:

Rolled throughput yield for multiple sequential steps can be calculated, under consistent definitions, as the product of the individual first-pass yields:

If three steps have FPYs of 0.98, 0.95, and 0.97:

This means that only about 90.31% of units pass all three steps without rework. Rework can conceal the true cost and capacity consumed by poor quality. FPY should therefore be reported alongside total yield, scrap, rework hours, and customer escapes.

8. Integrated Implementation Framework

An effective implementation follows the flow of risk and loss rather than launching disconnected initiatives.

Step 1: Establish the baseline

Define equipment boundaries, planned time, downtime categories, quality definitions, setup start and finish points, and safety constraints. Verify data quality before setting targets.

Step 2: Identify critical assets and characteristics

Rank equipment according to safety, environmental, quality, delivery, and financial consequences. Identify critical-to-quality characteristics and failure modes.

Step 3: Restore basic conditions

Clean, lubricate, tighten, align, guard, label, and correct obvious deterioration. Record temporary fixes separately from permanent corrective actions.

Step 4: Build standards and ownership

Create autonomous-maintenance check sheets, escalation rules, visual controls, standard work, and competency matrices. Operators should know what they own and when specialists must be called.

Step 5: Attack dominant losses

Use OEE loss trees, Pareto analysis, fault-tree or cause-and-effect analysis, and direct observation. Select improvement projects based on risk and economic significance, not only on ease of measurement.

Step 6: Reduce setup and stabilize quality

Apply SMED, standardize tooling, verify first-piece conditions, and install Poka-Yoke or source controls for recurring defects.

Step 7: Verify effectiveness

Use before-and-after data, control charts where appropriate, audit adherence to standards, and confirm that gains do not create safety, quality, or workload problems elsewhere.

Step 8: Sustain and learn

Review standards after product, tooling, software, or process changes. Treat recurring abnormalities as design or system problems rather than permanent operator burdens.

Illustrative Case: Precision Components Cell

Background

A fictional precision-components plant produces machined housings for industrial pumps. One machining cell has three CNC machines and serves several product families. The cell experiences frequent minor stops, long tool changes, and dimensional defects after startup.

Problem

Over one month, the cell records an OEE of 61%. The principal losses are 14% availability loss from breakdowns and changeovers, 16% performance loss from minor stops and reduced speed, and 9% quality loss from startup defects and rework. Operators report oil leakage and inconsistent chip removal, but abnormalities are not systematically escalated.

Analysis

The team first validates measurement definitions. It discovers that setup time is sometimes recorded as downtime and sometimes as minor stoppage. A time study shows that tool carts are prepared after the machine stops. Dimensional defects are concentrated in the first five pieces after a tool change. Investigation identifies worn collets, inconsistent tool-offset entry, and a measurement device stored away from the cell.

A failure-mode review ranks uncontrolled tool offset, spindle overheating, and chip accumulation as high-priority risks. The analysis also finds that operators have been cleaning around guards while machines remain energized, creating an unacceptable safety practice.

Approach and solution

The team implements the following actions:

• Establishes a safe energy-isolation procedure and prohibits cleaning inside hazardous zones during operation.
• Restores basic conditions through leakage correction, chip-management improvements, and standardized lubrication.
• Creates operator checks for coolant level, visible leakage, abnormal sound, tool condition, and guarding.
• Prepares tool carts and measurement equipment externally before the changeover.
• Introduces quick-release tooling and a standardized offset-verification checklist.
• Adds a Poka-Yoke in the setup screen requiring confirmation of product code, tool identity, and offset range.
• Requires first-piece source inspection before releasing the next batch.
• Trains operators and maintenance technicians, with competence verified through observation.
• Reviews OEE, FPY, near misses, and overdue abnormalities weekly.

Outcome

After eight weeks, average changeover time falls from 52 to 31 minutes. FPY improves from 91% to 97%, and OEE rises to 74%. The plant also records fewer minor stops and faster abnormality response. The improvement is not attributed solely to the OEE increase: safety audits confirm that the previous unsafe cleaning method has been eliminated, and maintenance backlog is reviewed to ensure that temporary repairs are not being hidden.

Learning points

The case demonstrates that OEE improvement is an outcome of system changes, not a standalone target. Setup reduction, source inspection, operator ownership, and maintenance restoration reinforce one another. It also shows why data definitions, safe work design, and standard confirmation are necessary before performance comparisons are trusted.

9. Challenges, Limitations, and Governance

Metric gaming and local optimization

Targets can encourage undesirable behavior: postponing maintenance, hiding minor stops, producing excess inventory, or classifying rework as good output. Governance requires transparent definitions, auditability, and balanced measures.

Standardization versus judgment

Standards reduce variation, but abnormal situations require trained judgment. A standard should specify escalation rather than encourage workers to continue when conditions are unsafe or uncertain.

Technology and AI

Sensors, connected maintenance systems, computer vision, and machine-learning models can improve condition monitoring and defect detection. However, digital systems require reliable labels, representative data, cybersecurity, calibration, and clear human accountability. A predictive model is not a substitute for physical inspection or engineering validation. False alarms can create alert fatigue, while missed alarms can create unjustified confidence.

Human factors and ethics

Operator involvement must not become unpaid transfer of specialist work, excessive surveillance, or individual blame for system-designed problems. Workload, ergonomics, training time, psychological safety, and consultation are part of implementation quality. Workers should be able to stop an unsafe process without retaliation.

Regulatory and safety requirements

Applicable machinery-safety, electrical-safety, occupational-health, environmental, calibration, and product-regulatory requirements take precedence over production targets. Lockout/tagout or equivalent energy-control practices must be designed into maintenance and setup work. Guards, interlocks, emergency stops, and safety devices must not be bypassed merely to improve OEE.

Best-Practice Checklist

A robust equipment and quality-at-source system generally includes:

• clearly defined OEE, FPY, downtime, and setup measurements;
• critical-asset and critical-characteristic analysis;
• safe autonomous-maintenance standards;
• preventive tasks linked to credible failure modes;
• condition monitoring where economically and technically justified;
• rapid escalation and closure of abnormalities;
• SMED separation of internal and external work;
• validated Poka-Yoke at the earliest feasible point;
• first-piece and source-inspection controls;
• visual standards and competence verification;
• balanced measures covering safety, quality, delivery, cost, and people;
• periodic review after engineering or product changes.

Chapter Summary

Equipment effectiveness is a combined result of reliability, maintainability, availability, process capability, safety, and operator involvement. TPM creates a shared system in which operators maintain basic conditions and detect abnormalities while specialists manage technical maintenance and improvement. OEE decomposes effectiveness into availability, performance, and quality, but it must be supported by consistent definitions and balanced governance.

SMED reduces the economic penalty of variety by separating external from internal setup and simplifying the remaining work. Poka-Yoke and source inspection prevent or detect errors at the point where they occur. FPY exposes defects and rework that may otherwise be hidden by final inspection. Reliability and maintainability decisions should be based on failure modes, consequences, evidence, and safety requirements.

The integrated objective is not simply higher machine utilization. It is a safe, stable, flexible, and capable process in which defects are prevented, abnormal conditions are visible, and improvement is owned by the system’s participants.

Key Terms

Autonomous maintenance: Operator-led routine care, inspection, cleaning, lubrication, and abnormality reporting.

Availability: The proportion of planned production time during which equipment is operating.

Critical-to-quality characteristic: A product or process characteristic that materially affects customer or regulatory requirements.

First-pass yield: The proportion of units accepted without rework or repeated processing.

FPY: Abbreviation for first-pass yield.

Internal setup: Setup activity possible only while equipment is stopped.

Maintainability: The ease and speed of restoring equipment to required performance.

MTBF: Mean time between failures for a repairable item.

MTTR: Mean time to repair.

OEE: Overall Equipment Effectiveness, calculated as availability × performance × quality.

Poka-Yoke: Mistake-proofing that prevents or immediately detects errors.

Planned maintenance: Maintenance scheduled according to time, usage, condition, or defined criteria.

Quality at source: Producing and verifying conforming output at the point where value is added.

Reliability: The ability of an item to perform its required function for a specified period under stated conditions.

Rolled throughput yield: The product of first-pass yields across sequential process steps.

SMED: Single-Minute Exchange of Die, a structured setup-reduction methodology.

Source inspection: Inspection performed at or before the point where a defect could be created or transferred.

TPM: Total Productive Maintenance, a company-wide system for maximizing equipment effectiveness.

Six Big Losses: Breakdown, setup, minor stoppage, reduced speed, process defects, and startup losses.

Review Questions

Short Answer Questions

1. Define autonomous maintenance and explain how it differs from specialist maintenance.
2. State the three components of OEE and give the formula for each.
3. Distinguish internal setup from external setup.
4. What is the difference between prevention and detection Poka-Yoke?
5. Define first-pass yield and explain why rework should be excluded from it.

Long Answer Questions

1. Explain how TPM integrates reliability, maintainability, safety, and operator ownership.
2. Discuss the assumptions and limitations of using OEE as a performance measure.
3. Describe a complete SMED implementation for a high-mix production cell.
4. Compare preventive, predictive, and run-to-failure maintenance policies.
5. Explain how source inspection and Poka-Yoke support quality at source.

Analytical / Application Questions

1. A machine has 480 scheduled minutes, 30 minutes of planned breaks, 50 minutes of downtime, an ideal cycle time of 0.5 minutes, 760 total pieces, and 730 good pieces. Calculate availability, performance, quality, and OEE.
2. A four-step process has FPYs of 0.96, 0.98, 0.94, and 0.99. Calculate rolled throughput yield and interpret the result for capacity and rework.
3. Design an integrated TPM–SMED–Poka-Yoke improvement plan for a bottleneck machine that has frequent breakdowns, long changeovers, startup defects, and unsafe cleaning practices. Identify measures, responsibilities, and safety controls.

Multiple Choice Questions with Answers

Question 1

Which statement best describes autonomous maintenance?

A. Operators perform all specialist repairs without maintenance support

B. Operators conduct basic care, inspection, and abnormality reporting within defined standards

C. Maintenance is performed only after equipment failure

D. Production output is measured without considering equipment condition

Correct Answer: B. Operators conduct basic care, inspection, and abnormality reporting within defined standards

Question 2

A machine has planned production time of 400 minutes and operates for 360 minutes. What is its availability?

A. 80%

B. 85%

C. 90%

D. 95%

Correct Answer: C. 90%

Question 3

Which activity is an external setup activity?

A. Removing a die while the machine is stopped

B. Adjusting the machine after installation

C. Preparing and checking tools before the machine stops

D. Running trial pieces after a changeover

Correct Answer: C. Preparing and checking tools before the machine stops

Question 4

What is the strongest form of Poka-Yoke among the following?

A. A warning sign that the operator may ignore

B. A final inspection after all processing is complete

C. A design feature that makes incorrect assembly impossible

D. A monthly report of defects

Correct Answer: C. A design feature that makes incorrect assembly impossible

Question 5

A process produces 1,200 units, of which 1,140 pass without rework. What is the FPY?

A. 90.0%

B. 92.5%

C. 95.0%

D. 97.5%

Correct Answer: C. 95.0%

Question 6

If availability is 0.90, performance is 0.80, and quality is 0.95, what is OEE?

A. 68.4%

B. 72.0%

C. 76.0%

D. 85.5%

Correct Answer: A. 68.4%

Question 7

Which measure most directly represents the average time required to restore a failed machine?

A. MTBF

B. MTTR

C. FPY

D. RTY

Correct Answer: B. MTTR

Question 8

Why is source inspection generally preferred to relying only on end-of-line inspection?

A. It postpones defect detection until customer delivery

B. It detects or prevents errors near the point where they are created

C. It eliminates the need for process standards

D. It guarantees that equipment will never fail

Correct Answer: B. It detects or prevents errors near the point where they are created

Question 9

Three sequential process steps have FPYs of 0.98, 0.95, and 0.97. What is their rolled throughput yield approximately?

A. 86.42%

B. 90.31%

C. 93.33%

D. 97.00%

Correct Answer: B. 90.31%

Question 10

Which governance practice best prevents OEE improvement from creating unsafe behavior?

A. Rewarding output regardless of safety incidents

B. Allowing operators to bypass interlocks during changeovers

C. Using OEE with safety, quality, delivery, and workforce measures

D. Excluding maintenance time from all performance reports

Correct Answer: C. Using OEE with safety, quality, delivery, and workforce measures