Stability is the ability of a measurement system to keep the same measurement result over time, reflecting drift. Repeatedly measuring the same master part under identical conditions at intervals, if the measurement mean or spread changes over time, the measurement system is unstable - typical causes are gauge wear and aging, environmental temperature and humidity changes, voltage fluctuation and master-part change. Poor stability makes the same product measure differently at different times and undermines the comparability of long-term quality data.
Choose a master part (or standard) that stays stable over the long term, measure it repeatedly under the same conditions on a fixed schedule (daily, weekly), taking 3-5 readings each time. Plot the data in time order on an I-MR or Xbar-R chart and watch whether the center line drifts or the range widens. More accumulated data over a longer period give a more reliable drift judgement - collect at least 20 measurement periods.
No points beyond the limits and no systematic patterns (steady runs or trends) indicate that the measurement system is stable; several points on one side of the mean or a continuing up/down trend on the mean chart suggest drift - check whether the master part changed, the gauge wore or the environment shifted. The tool flags points by run rules, hints at drift direction and magnitude, and combines the range chart to see whether spread grows over time.
The master part itself must be stable and traceable and the measurement conditions consistent. After detecting drift, correct the cause before re-establishing the baseline. Combine stability, bias and GRR studies periodically in the MSA flow to form a complete measurement-system monitoring scheme, and link the stability results with calibration records - re-evaluate the baseline after due calibration or repair.