Accuracy vs. Repeatability

What is the difference between repeatability and accuracy?

Repeatablity: Within what degree of uncertainty will multiple measurements compare to each other?

Ex: 50 microns at 68% (two-sigma)

Accuracy: Within what degree of uncertainty does the absolute measured result compare to a reference result?

Ex: 200 microns at 99.6% (six-sigma)

Perceptron Accuracy and Repeatability

Can a measurement system be repeatable without being accurate?

Yes, a dimensional gauging solution can produce extremely repeatable measurements independent of its accuracy.

Can a measurement system provide any value if it is known to be repeatable, but not considered to be accurate?

If used in the proper context, a repeatable dimensional gauging solution can provide extremely beneficial process knowledge. Once a system is producing reliable, robust, repeatable data it can be used as a very powerful tool for understanding and diagnosing problems in your manufacturing process. Is your process in control? Is your process stable? You can also proactively manage your process by immediately detecting: increased variation, mean shifts, quality trends, bi-modal processes, and much more.

Can a system be accurate without being repeatable?

No, repeatability is a prerequisite for accuracy. If your measurement system cannot product reliable, repeatable measurements – you cannot move on to the next step of verifying that the measurement results are accurate.

For example, if you own a clock that continually runs either too fast or too slow – the first order of business is to make the clock run “on time”. Once that task is accomplished you can move on to the task of setting it to the “correct” (or accurate) time.

What additional benefits are provided by an accurate measurement system?

In addition to the aforementioned process knowledge provided by a repeatable dimensional gauging solution, an accurate system will provide specific product knowledge. For example, an accurate system allows you to understand not only if your process is stable and under control, but also if that stable process is being built to nominal specifications. Are you well within the bounds of your tolerance or are you nearing the limits? Did the recent mean shift in the process push you outside your tolerance limits or are you able to confidently continue production?

How do I make the leap from repeatable to accurate?

Many options exist, but generally one of the following techniques is used:

  • Initial Calibration – During the initial installation of the measurement system, additional steps are taken to insure accurate results from the start. Using a laser tracker, or similar device, the system is tightly calibrated into the desired coordinate grid. The end result being repeatable and accurate measurement data.

  • System Correlation – After the initial installation, once system repeatability has been established, a correlation study is conducted where the measurement results are compared to those of a “reference” gauge (sometimes a manual measurement device, sometimes a CMM). The results are then offset to match the results from the “reference” gauge. The end result being repeatable and accurate measurement data.