EXIN LSSA Black Belt Practical flashcards

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

Cpk formula

Cpk equals the smaller of (USL minus mean) or (mean minus LSL), divided by three times the within-subgroup standard deviation. It accounts for both spread and off-centering.

Cp versus Cpk

Cp measures potential capability using specification width over six sigma and ignores centering. Cpk equals Cp only when the mean sits exactly at the specification midpoint.

DPMO calculation

DPMO equals defects divided by the product of units and opportunities per unit, multiplied by one million. It normalizes defect counts across processes with different complexity.

Rolled throughput yield

RTY is the product of first-pass yields of every process step. Five steps each at 95% yield give an RTY of about 77%, far below the single-step figure.

Gauge R&R acceptance

Percent GRR below 10 is acceptable, 10 to 30 is conditionally acceptable based on cost and risk, and above 30 is unacceptable. Number of distinct categories should be five or more.

Rational subgrouping

Subgroups should contain only common-cause, short-term variation. Mixing shifts or batches inside a subgroup inflates the within sigma and hides special-cause signals on the chart.

Sample size for mean shift

Per-group n approximates 2 times the squared sum of z-alpha/2 and z-beta, times sigma squared, divided by delta squared. At 80% power and two-sided 0.05 the factor is about 15.7.

Xbar-R control limits

The upper control limit for the subgroup mean equals the grand mean plus A2 times the average range. For subgroup size five, A2 is 0.577 and D4 is 2.114.

Resolution IV design

A fractional factorial where main effects are clear of two-factor interactions, but two-factor interactions are aliased with each other. Used for screening many factors cheaply.

RPN in FMEA

Risk priority number is severity times occurrence times detection, each rated one to ten. Severity of nine or ten mandates action regardless of the computed RPN.

Multicollinearity indicator VIF

Variance inflation factor above ten signals that a predictor is strongly correlated with others, making coefficient estimates unstable. Drop, combine, or re-specify the correlated predictors.

Project charter problem statement

It states the measurable gap, where and when it occurs, and its magnitude, without naming a cause or solution. Embedding a cause biases the Analyze phase.

Champion and Black Belt roles

The champion or sponsor approves tollgates, owns the business case, and removes barriers. The Black Belt leads technical analysis and coaches Green Belts on method.

Kano model categories

Customer needs classify as must-be (dissatisfiers when absent), performance (more is better), and delighter (unexpected). Must-be needs are translated into CTQs first.

Hard versus soft savings

Hard savings change the profit and loss statement, such as labor or material reduction. Soft savings are cost avoidance or released capacity and must be labeled as such.

Tuckman team stages

Teams pass through forming, storming, norming, and performing. Conflict in the storming stage is normal and resolves faster with clear roles and decision rights.

Process cycle efficiency

Value-added time divided by total lead time, expressed as a percentage. Office and service streams commonly run below five percent, showing most time is waiting.

Little's Law

Average work in process equals throughput rate multiplied by average flow time. Capping WIP therefore shortens lead time proportionally at a fixed completion rate.

Takt time

Available production time divided by customer demand. It sets the pace a process must meet, and station cycle times are balanced at or just below it.

Poka-yoke preference order

Prevention devices that make errors impossible rank first, detection at the source second, and detection downstream last. Prevention does not rely on operator vigilance.

SMED changeover reduction

Single-minute exchange of die separates internal from external setup tasks and simplifies adjustments. Typical results cut changeover time by half or more, enabling smaller lots.

Line balance efficiency

Sum of all task times divided by the product of station count and bottleneck station time. Higher efficiency means less idle capacity across the balanced line.

SPC chart selection rule

Use Xbar-R for subgroups of two to nine, Xbar-S for ten or more, and individuals charts for single measurements. Attribute data uses p, np, c, or u charts.

Out-of-control action plan

A written response specifying who reacts to a signal, what is checked first, what containment applies, and who authorizes stopping the process. It is prepared before chart deployment.

Control phase handover

The process owner formally accepts the control plan, charts, reaction plan, and training records at a documented gate. Informal transfer is a common cause of reverted gains.

Ninety-day follow-up audit

A scheduled check after project closure comparing the key metric and standard work adherence to the control plan. Drift beyond limits triggers reopening the project.

Gauge discrimination rule

A measurement device should resolve at least one-tenth of the specification tolerance. Coarser resolution cannot support credible capability or conformance claims.

Attribute agreement kappa

Kappa adjusts agreement for chance. Values below roughly 0.7 indicate appraisers do not agree beyond chance, so operational definitions or training need revision.

Nested gauge study

Used when measurement is destructive or parts cannot be remeasured. Parts are nested within appraisers, and part-within-appraiser variation substitutes for repeatability.

Multi-vari study purpose

It separates positional, cyclical, and temporal variation before formal testing. The result shows which strata carry variation so later tests target the right sources.

Correlation limits

The coefficient r measures only linear association. A value near zero does not prove independence, and strong correlation without designed or temporal evidence does not establish causation.

Confirmation runs in DOE

Runs at the predicted optimum verify the fitted model. If the observed response falls outside the prediction interval, the model is inadequate or a lurking variable changed.