Gauge Repeatability and Reproducibility (GR&R) is a statistical method used to quantify the total variation in a measurement system. A measurement system here refers to the combination of tools, operators, procedures, and environmental conditions used to take a measurement.
To break down the two core components:
In plain language, GR&R answers the question: "If I measure the same thing multiple times, how much will the numbers change?" A good measurement system will produce consistent results, meaning the variation it introduces is small compared to the variation in the parts you are measuring.
The quality of your measurement system directly impacts every stage of production, from part function to assembly and cost.
The most critical impact is on part performance. If your measurement system is unreliable, you may incorrectly reject good parts or accept bad ones. For example, consider a precision bearing raceway that requires a roundness of 0.002 mm. If your measurement system has a GR&R variation of 0.003 mm, you cannot confidently distinguish between a good part and a bad one. This can lead to premature bearing failure, increased wear, or even safety hazards in applications like aerospace or automotive.
In modern manufacturing, parts are often produced by different suppliers or on different machines and must assemble seamlessly. A poor measurement system can lead to parts that meet specifications on paper but do not fit in practice. For instance, a shaft measured on a machine with high GR&R might be accepted as within tolerance, but its actual size could be too large to fit into a mating hole. This leads to assembly line delays, rework, and scrap.
The financial implications of an unreliable measurement system are significant. Scrap costs increase when good parts are rejected. Rework costs rise when bad parts are accepted and need to be corrected later. Warranty costs and customer dissatisfaction increase when defective parts reach the market. Furthermore, investing in a high-precision machine tool is wasted if your measurement system cannot verify the parts it produces. For example, purchasing a GA-DV750 vertical machining center with a repeatability of 0.0024 mm is a poor investment if your inspection system can only measure to ±0.01 mm.
GR&R is typically measured using a designed experiment. The most common method is the crossed GR&R study, which is a practical, shop-floor friendly approach.
This method is relatively straightforward to implement:
The result is often expressed as a percentage of the total process variation or the tolerance. The common acceptance criteria are:
The accuracy of the GR&R study itself depends on the number of parts and operators used. The standard 10-part, 3-operator, 2-repetition study is a balance between statistical rigor and practicality. Using fewer parts or operators will reduce the study's accuracy, leading to less reliable conclusions about your measurement system.
The total variation measured in a GR&R study comes from several sources. Understanding these sources is key to improving your measurement system.
This is the inherent variation of the measurement tool itself. It can be caused by:
This is the variation introduced by the people and procedures involved in the measurement. It can be caused by:
This is a less common but important source of error. It occurs when the measurement system's performance changes depending on which part is being measured. For example, an operator might consistently measure small parts accurately but struggle with large, unwieldy parts. This is known as the Operator x Part interaction.
GR&R is a powerful tool, but it is not a universal solution. It has several important limitations that users must be aware of to avoid overconfidence.
A common mistake is assuming that a "passed" GR&R study means your measurement system is perfect. It only means the system is adequate for the specific application it was tested for. If you change the part, the tolerance, or the environment, the GR&R results may no longer be valid.
Another point of overconfidence is relying solely on the machine's published specifications. For example, the GA-MV856 vertical machining center has a published positioning accuracy of 0.010 mm. This is a best-case scenario achieved under ideal laboratory conditions. In a real shop floor environment, with variations in temperature, tool wear, and operator setup, the actual accuracy and repeatability of the machine's built-in measurement system will be different. A GR&R study is required to verify the system's real-world performance.
| Method | Accuracy of GR&R Assessment | Effort (Time & Resources) | Typical Use Case |
|---|---|---|---|
| Crossed GR&R (10 parts, 3 operators, 2 repetitions) | High | Medium | Most common shop-floor application; suitable for critical features and new measurement system validation. |
| Nested GR&R | Medium | Low | Used when parts are destroyed during measurement or cannot be measured multiple times. |
| Short GR&R (5 parts, 2 operators, 2 repetitions) | Medium | Low | Quick checks for non-critical features or as a preliminary assessment. |
| Attribute GR&R | Medium | Medium | For pass/fail measurement systems (e.g., visual inspection, go/no-go gauges). |
Follow this checklist to ensure your GR&R study is valid and provides actionable insights:
Q: What is the difference between GR&R and the machine's published repeatability? A: A machine's published repeatability is a measure of the machine's own ability to position a tool consistently. GR&R measures the entire measurement system, which includes the machine, the operator, the fixturing, and the environment. The machine's specification is a component of the overall GR&R. For example, the GA-DV750 has a published repeatability of 0.0024 mm, but the GR&R of its in-process measurement system might be higher due to other factors.
Q: Can I use GR&R for non-dimensional measurements, like surface roughness or hardness? A: Yes, GR&R can be applied to any measurement system. The methodology remains the same: measure the same parts multiple times with different operators to quantify the variation. The acceptance criteria may need to be adjusted based on the specific characteristic and its tolerance.
Q: My GR&R is 25%, which is in the "marginal" range. Should I accept it? A: This depends on the application. If the feature is non-critical and the cost of improving the measurement system is high, it may be acceptable. However, for critical features like those in aerospace or medical devices, you should aim for a GR&R of less than 10%. Always consider the risk of making a wrong decision.
Q: How often should I perform a GR&R study? A: There is no universal answer, but a good rule of thumb is to perform a full GR&R study annually for critical measurement systems. You should also conduct a study whenever there is a significant change, such as a new operator, a recalibration of the tool, a change in the part design, or a shift in the production environment.
Q: If my GR&R is too high, what is the first thing I should check? A: The first step is to analyze the data to see if the largest source of error is repeatability (tool) or reproducibility (operator/process). If repeatability is high, check if the tool is calibrated, worn, or appropriate for the measurement. If reproducibility is high, look at operator training, fixturing, and environmental controls. Often, a simple change in how operators are trained to use the tool can significantly improve GR&R.
Send us your part, batch size and target port. We reply with a configuration, a binding lead window and a total-cost comparison.
Request a quote →