NEWS CATEGORIES

 

What Is the Relationship between Accuracy and Repeatability?

Load cells and other measuring devices usually have published standards for accuracy and repeatability. In the United States, the two main bodies governing these standards are the National Institute of Standards and Technology (NIST) and National Type Evaluation Program (NTEP). Let’s begin by defining these terms, because although I’ve sometimes heard them used interchangeably, they refer to different aspects of interpreting data.
Accuracy: is the degree of closeness (proximity) of measurements of a quantity to that quantity’s actual (true) value.
Repeatability: can also be called “precision” and refers to the variation in measurements taken by a single person or instrument on the same item and under the same conditions. Or, put in other terms, the degree to which repeated measurements under the same conditions show the same results. When the variation in measurements is smaller than some agreed-upon limit, the measurement is said to be “repeatable”. According to NIST repeatability is dependent on the following: use of the same measurement procedure, the same observer, the same measuring instrument, the same conditions, the same location and that the repetition occur over a short period of time.
Let’s use one of our most popular load cells, the LFS 210 1K s-beam load cell, to illustrate how individual measurements could be accurate but not repeatable or repeatable but not accurate. According to the published specifications for this unit, the accuracy should be +/- 0.05%, which would be 0.5 pounds, the linearity should be +/- 0.03%, which would be 0.3 pounds and the repeatability should be +/- 0.01%, which would be 0.1 pounds. (In these example, I have not provided individual data points, but you can assume a data set for each scenario.)
For example: let’s say this load cell was not being loaded properly (off-axis). We could repeatedly load this 1K load cell with a known weight of 500 pounds and consistently produce a reading of between 496.0 and 496.4 pounds. Then we could say that these measurements meet the published standards for repeatability, but not for accuracy, because in order to be accurate for this range, our readings would have to be between 499.5 and 500.5 pounds. So in this example, the measurements are repeatable, but not accurate.
Now let’s say that a series of measurements with a 500 pound load produced results varying from 499.0 to 500.5. Any measurements falling between 499.5 and 500.5 pounds would be considered accurate for this load cell, but because some of the data points were lower, you could not claim to meet the published standard for repeatability.
Now imagine a series of readings between 498 and 502. This data set would be neither accurate nor repeatable as defined by the published standards.
Conversely, a data set where all points fell between 499.8 and 500.0 would meet the qualifications for accuracy and repeatability.
As always, if you have any questions related to this material, our support staff at Cooper Instruments is available to help. Contact them by calling (800) 344-3921 or emailing sales@cooperinstruments.com.
We’d love to hear your feedback regarding this or any other article we’ve posted. To leave feedback, ‘Like’ us on Facebook and then post your feedback on our wall.