Quality Management Tools
What to know
Separate the toolkits. The seven basic tools - flow chart, check sheet, Pareto chart, cause-and-effect diagram, histogram, control chart, scatter diagram - work with data and are the first pass on any problem. The seven management and planning tools - affinity diagram, tree diagram, matrix diagram, prioritization matrix, process decision program chart, activity network diagram, interrelationship digraph - organize ideas and ambiguous situations.
Choose charts by data type. Variable data use X-bar and R (small subgroups) or X-bar and s (larger subgroups). Attribute data use p (fraction defective, varying sample), np (count defective, constant sample), c (defects per unit, constant sample size or area), and u (defects per unit, varying sample). A histogram shows the shape of a distribution but not time order; a run chart or control chart shows time order.
Capability: Cp = (USL − LSL) / 6 sigma measures potential spread only; Cpk = min[(USL − mean), (mean − LSL)] / 3 sigma accounts for centering, so Cpk is always less than or equal to Cp.
FMEA scores severity, occurrence, and detection; risk priority number = S x O x D. Severity is attacked first because inspection cannot reduce it. QFD's house of quality converts voice of the customer into design requirements. Acceptance sampling uses AQL and LTPD to balance producer's and consumer's risk.
Worked example
A shaft has USL 10.20, LSL 9.80, mean 10.05, sigma 0.05. Cp = 0.40 / 0.30 = 1.33; Cpk = min(0.15, 0.25) / 0.15 = 1.00. The process spread is capable but it is off-center - the fix is centering, not variation reduction.