Equally important to the dollar budget is the error budget, which is often overlooked or determined de facto by cost constraints. For optimum system performance, each individual system element needs an acceptable error allocation. In determining system error budgets, "repeatability" is often confused with "reproducibility".
The accepted definition of "repeatability" is the closeness of agreement among a number of consecutive measurements of outputs for the same input where this input is approached from the same direction after a transversal of input across the full-scale span.
The accepted definition of "reproducibility" is the closeness of agreement among a number of repeated measurements of outputs for the same input where this input is approached from any direction and these measurements are made over a period of time.
The subtle difference between "repeatability" and "reproducibility" is as follows;
(a) Repeatability includes neither drift errors (since consecutive measurements over a period of time are too short for drift to be a
factor) nor hysteresis,
(b) Reproducibility includes drift (repeated measurements over any length of time), hysteresis, and repeatability. Note reproducibility includes repeatability.
All measurements are made within the test unit’s allowed range and operating conditions.