Repeatability is the variation that occurs when a person uses a measurement tool (gage) to measure the same part repeatedly. This article will cover what it means to measure and evaluate the repeatability of a gage. What is the measure of a person? What is the measure of a person measuring? Repeatability is one measure of variability when measuring. Confused? You won’t be after reading this article.
Overview: What is repeatability?
If we want to conduct a measurement system analysis, one thing to check is variability due to repeatability. Repeatability is the variation that occurs when a person uses a measurement tool (gage) to measure the same part repeatedly.
2 benefits and 1 drawback of repeatability
All data comes from some sort of measurement tool (gage). If you can’t trust your gage to have acceptable variation and bias, then you can’t trust your data. Repeatability is one of the sources of variation that we want to minimize to get reliable measures. Repeatability is a basic characteristic of an acceptable gage.
1. It can be easy to repeatably measure a part
Conceptually, to measure repeatability, one person measures one part multiple times with the same gage. Easy!
2. It can be difficult to repeatably measure a part
If the characteristic being measured degrades over time (like electrical output) or breaks the part (like bonding-strength pull tests, then the repeatability measure can be complex.
3. When the variation due to repeatability is large, there is usually trouble in the operator, method, or equipment
The repeatability test doesn’t tell us what is wrong with the gage, just that something IS wrong. The repeatability test data will give you clues as to what is wrong.
It is up to you, the gage data analyst, to work with the engineers or equipment experts to figure out what it is.
Why is repeatability important to understand?
The variation in a data sample comes from two sources:
- The part being measured (e.g., the thickness of a washer)
- The measurement tool you use to measure the part (e.g., a caliper)
We would like the contribution of variation from the measurement system to be zero so that the variation we see is attributable only to the part being measured.
This will never happen. The measurement system always contributes variation to the part measure. The best we can do is to make sure the gage variation is as small as possible, stable, and consistent.
Repeatability is the variation that occurs when a person uses a gage to measure the same part repeatedly with the same measurement tool.
When a gage is acceptable, the variation due to repeatability is small. For example, when we measure a washer over and over again, the values should group closely together.
Repeatability testing is important whether the gage is mechanical or human.
Mechanical gages can be simple (like a ruler or caliper) or they can be complex (like microscopic measurement equipment)
When a person decides whether a part passes or fails inspection, they are acting as a human gage. It is just as important to make sure the person can be repeatable in their pass/fail decisions.
Conceptually, the approach is the same. The person inspects the same part multiple times. If they are repeatable, they give the same decision (e.g., pass the part) each time.
An industry example of repeatability
Operators are measuring the weight of a bolt on a scale before it goes into an airplane chassis. The gage engineer wants to verify that the scale measure is repeatable.
Here is the data for the operator repeatability.
| Bolt ID | First Measure (g) | Second Measure (g) | Range (g) |
| 1 | 28 | 30 | 2 |
| 2 | 32 | 34 | 2 |
| 3 | 70 | 91 | 21 |
| 4 | 33 | 38 | 5 |
| 5 | 25 | 29 | 4 |
What can the repeatability data tell the gage engineer about the variability of the scale?
- The repeatability variation is much larger for bolt 3. The second measure is much higher than the first measure. Questions to ask include: Did something unusual happen on the two measures? Does the scale have trouble measuring heavier bolts?
- The repeatability of the other bolts (1, 2, 4, 5) are more tightly grouped.
The gage engineer examined bolt number 3. He noticed it was one of the heavier bolts. He theorized that the scale worked best with bolts under 50 grams.
The engineer will run a subsequent study to check his theory. If this is the case, a more robust scale should be installed.
3 best practices when thinking about repeatability
1. Make sure you watch the repeatability test
Watching the repeatability test will give you insights that go beyond the data. For example, you might pick up things the operator does that are not in the measurement procedure.
2. Make sure the gage is in good shape prior to the test
You don’t want to try to validate a gage if it is broken or uncalibrated.
3. Make sure there are operational definitions in place when the gage is a human inspector
Directions for how to inspect and examples of non-defective and defective parts are needed if the operator is to have a chance to be repeatable.
Frequently Asked Questions (FAQ) about repeatability
What are the two Rs in the Gage R&R study?
Repeatability and Reproducibility are the 2 Rs.
What are 5 measurement system analysis methods for the reduction of measurement bias and variation
Can measurement system analysis be conducted on discrete data?
Yes! Gage R&R for discrete attribute data looks at the Repeatability and Reproducibility of the pass/fail status of a sample of parts. It looks at the proportion of parts pass/fail status matched.
Final thoughts on repeatability
If you want good data, you must have a repeatable gage. Repeatability is the variation that occurs when a person uses a measurement tool (gage) to measure the same part repeatedly. You can do a repeatability test on both continuous or discrete data measures.
