About this lesson
All measurement systems introduce an element of error into the measured value. There are techniques for determining the source of that error and thereby determining if a different measurement system needs to be used.
When to use
Whenever an inspection or test is performed, there is the introduction of a measurement error. The measurement system analysis will determine the source and magnitude of the error. This module discusses the characteristics of errors.
An observed measurement, either inspection or test, is derived from a combination of the true value of the item, the standard being used in the measurement system, and the source and magnitude of the measurement errors. There are five categories of measurement errors. The course on measurement systems analysis will explain how to determine the magnitude. At this time, I will just review the categories.
Accuracy is the difference between the actual value of the item and the average measured value of that item. Accuracy is also referred to as bias. Accuracy errors can often be corrected through calibration.
Precision is the uncertainty in the measurement due to equipment, people and procedures. It is often divided into repeatability (same person measuring multiple times with different results) and reproducibility (different people using same equipment and procedures with different results). Gage R&R studies identify the magnitude of this error.
Discrimination is the number of gradations of the measuring system within the normal range of the measurement. A minimum of ten gradations is required for adequate discrimination, although I prefer 25. If discrimination is inadequate, you need to change the measurement system.
Linearity is the degree to which the bias shifts depending upon where within the range the measurement is occurring. For instance, it could be low bias at the high end of the range and a high bias at the low end of the range. If linearity is unacceptable, you need to change the measurement system.
Stability is the degree to which the bias changes over time. This is often due to wear and tear. Also, if the measurement system has a consumable, the bias shifts as the consumable is used up. If this is excessive, you need to change the measurement system.
Hints & tips
- Precision errors can quickly change as operators and procedures change. Monitor this if you have a changing work environment.
- Conduct Gage R&R studies before relying on data in the Measure phase.
- Linearity, Stability, and Discrimination are inherent in the measurement system design, they cannot be eliminated with calibration or training.
Lesson notes are only available for subscribers.