Quality Control Calculators: Process Capability, Sigma Level, Control Charts and Gage R&R

Match the question to the tool

Process capability: the voice of the process versus the voice of the customer

Process capability puts a number on whether a process can reliably make product inside its specification limits. It compares the natural spread of the process, the voice of the process, against the tolerance the customer allows, the voice of the customer. Cp measures whether the spread simply fits inside the tolerance, ignoring where it is centered, while Cpk adds centering by measuring the distance from the process mean to the nearer specification limit in units of three standard deviations. A high Cp with a low Cpk is the classic signal of a capable but off-center process. Because capability assumes the process is stable, it is only trustworthy once a control chart has confirmed statistical control, which is why capability and control charts are always taught together.

Sigma level and DPMO: a common scale for quality

Defects per million opportunities, DPMO, and the sigma level built on it give quality a single comparable scale across processes of very different complexity. DPMO counts defects, divides by the units inspected times the opportunities for a defect on each, and scales the result to a million, so a simple part and a complex assembly can be compared fairly. Sigma level then expresses that defect rate on the normal-distribution scale, where each step up means far fewer defects. The convention adds a 1.5-sigma shift to account for long-term drift, which is why a Six Sigma process is quoted at 3.4 DPMO rather than the 2 per billion a perfectly centered calculation would give.

Control charts: separating signal from noise

A control chart is the tool that decides whether a change in a process is real or just ordinary variation. It plots a statistic, an average, a range, or an individual reading, against a center line and control limits set three standard deviations away, computed from the process itself rather than from the specification. Points inside the limits are common-cause noise to be left alone; points outside, or non-random patterns within, signal a special cause worth investigating. Getting the limits right matters because tampering with a stable process is as costly as ignoring a real shift, and the correct limit formula depends on the chart type and subgroup size.

Sampling and measurement: inspecting and trusting the numbers

The last cluster covers how much to inspect and whether the measurements can be trusted at all. Acceptance sampling, standardized in ANSI/ASQ Z1.4, sets the sample size and the accept or reject numbers for an incoming lot at a chosen acceptable quality level, so you inspect a fraction rather than everything. Gage R&R, part of measurement system analysis, checks the measurement system before you trust any of the numbers above it: it separates the variation caused by repeatability and reproducibility from real part-to-part variation. If the gauge eats too much of the tolerance, every capability index and control chart built on those readings is suspect, which is why measurement analysis underpins the whole discipline.

Quality control calculator FAQs