Get started
ENDE

Part Variation (PV)

Variation between the parts themselves — should dominate the measurement-system variation in Gage R&R.

Also known as: PV, part-to-part variation, part-to-part spread

Part variation (PV, part-to-part) captures the spread produced by the actual differences between the parts selected for a Gage R&R study — i.e. the process variation, not the measurement-system variation.

A good measurement system shows a large PV share and a small GRR share — meaning: what the system reports is actually the process, not its own noise. When measurement-system variation dominates instead, the study loses meaning.

  • Parts must cover the relevant process range — not all near the mean
  • Too narrow a part range → PV small, %GRR artificially large, ndc low
  • Randomised order and blinded operators are mandatory to separate PV cleanly

See also

Used in

In the Algorithm Lab

Sources

  • AIAG — Measurement Systems Analysis Reference Manual, 4th Edition (MSA-4), Chapter 4
  • Wheeler, D. J. — EMP III: Evaluating the Measurement Process, SPC Press
  • Montgomery, D. C. — Design and Analysis of Experiments, 10th Edition, Chapter 14