CSSC Black Belt flashcards

32 cards. Read the term and the definition here. The deck at the top of this page is optional.

DMAIC phases

Define, Measure, Analyze, Improve, Control. A data-driven roadmap for improving an existing process by reducing variation and defects in the critical-to-quality output.

Project charter purpose

Authorizes the project and defines the business case, problem and goal statements, scope, team, milestones, and expected benefit. It contains no solution and is signed by the champion.

Y equals f of X

The output Y (CTQ) is a function of process inputs X. Six Sigma identifies and controls the vital few Xs that drive most variation in Y.

SIPOC diagram

Suppliers, Inputs, Process, Outputs, Customers. A high-level map used in Define to set scope and in Measure to locate measurement points.

TIMWOODS wastes

Transport, Inventory, Motion, Waiting, Overproduction, Overprocessing, Defects, Skills unused. These are non-value-adding activities Lean targets for elimination.

Cost of poor quality

Four categories: internal failure, external failure, appraisal, and prevention. Failure costs usually dominate, so projects target them first.

Gage R and R percent

Percent GRR equals 100 times GRR divided by total variation, where GRR combines repeatability and reproducibility. Below 10% acceptable, 10-30% conditional, above 30% unacceptable.

Repeatability versus reproducibility

Repeatability is variation when one appraiser measures the same part repeatedly with the same gage. Reproducibility is variation between different appraisers using the same gage.

Number of distinct categories

ndc equals 1.41 times part variation divided by gage variation. A value of at least five indicates the measurement system can discriminate parts usefully.

Kappa statistic

Kappa equals observed agreement minus chance agreement, divided by one minus chance agreement. Used for attribute agreement studies; 0.75 or higher is generally good.

Cp versus Cpk

Cp equals specification width divided by six sigma and ignores centering. Cpk compares the nearer specification limit to the process mean, so it penalizes off-center processes.

Ppk versus Cpk

Ppk uses the overall long-term standard deviation; Cpk uses within-subgroup short-term sigma. A large gap between them indicates shifts between subgroups over time.

Sigma level conversion

For a centered process, Z equals three times Cpk. Convert Z to DPMO using the normal table, applying the conventional 1.5 sigma long-term shift.

Type I and Type II errors

Type I (alpha) rejects a true null hypothesis, a false alarm. Type II (beta) fails to reject a false null hypothesis, a missed effect. Power equals one minus beta.

p-value interpretation

The probability of observing data at least as extreme as collected, assuming the null hypothesis is true. Reject the null when the p-value is below alpha.

Paired t test use

Compares two measurements taken on the same unit, such as before and after. Pairing removes unit-to-unit variation and increases power relative to an independent two-sample test.

ANOVA F statistic

F equals mean square between groups divided by mean square within groups. A significant F shows at least one group mean differs; post-hoc tests identify which groups.

R-squared in regression

The proportion of variation in the response explained by the model. Adjusted R-squared penalizes added predictors that do not improve fit meaningfully.

Multi-vari study

Stratifies variation into within-piece, piece-to-piece, and time-to-time components. It locates the dominant variation source before formal hypothesis testing or experiments.

FMEA risk priority number

RPN equals Severity times Occurrence times Detection. Higher values prioritize action, but high severity alone warrants action regardless of the computed RPN.

Full factorial design

A design with all combinations of k factors at two levels, requiring 2 to the k runs. It estimates all main effects and all interactions without confounding.

Design resolution

Resolution III confounds main effects with two-factor interactions; IV confounds two-factor interactions with each other; V keeps main effects and two-factor interactions clear.

Interaction effect

The effect of one factor differs depending on the level of another factor. On a plot, non-parallel lines indicate an interaction is present.

Poka-yoke

Error-proofing that prevents a mistake from being made or makes it immediately detectable. Prevention devices are preferred over detection devices such as alarms.

Five S

Sort, Set in order, Shine, Standardize, Sustain. A workplace organization method that exposes abnormal conditions and supports standardized, repeatable work.

Control chart limits

Upper and lower limits set at three standard deviations of the plotted statistic around its center line. They describe process voice, not customer specification limits.

Common versus special cause

Common-cause variation is inherent and stable, requiring process redesign to reduce. Special-cause variation is assignable and signals investigation of a specific event.

Western Electric rules

Sensitivity rules signaling special cause, such as one point beyond three sigma, two of three beyond two sigma, or eight consecutive points on one side of the center line.

p chart versus c chart

A p chart plots proportion defective and allows varying sample size. A c chart plots defect counts and requires a constant inspection unit.

Control plan contents

Process step, characteristic and specification, measurement method and frequency, sample size, control method, reaction plan and owner, plus revision history.

DMADV roadmap

Define, Measure, Analyze, Design, Verify. A Design for Six Sigma roadmap used when creating a new product or process with no existing baseline to improve.

QFD house of quality

A matrix relating customer requirements to engineering characteristics, with importance weights, relationship strengths, competitive ratings, and a correlation roof to prioritize design decisions.