Download Members: $0.00 Non‑Members: $75.00 |
Buy Now |
Publication Details
Published Date: | |
---|---|
Citation: | John Stoup and Wei Ren |
Abstract
The Dimensional Metrology Group at the National Institute of Standards and Technology (NIST) has developed its suite of ultra-high accuracy Moore M48 coordinate measuring machines (CMM) into the premier 3-axis CMM capability in the world for gauges and artifacts up to 1 meter in size. Performing CMM measurements with expanded (k=2) uncertainties as low as 50 to 75 nanometers produces a set of challenges unique to these exceptional machines, but the general concepts still pertain to all industrial applications of CMM measurement.
When defining a level of CMM competence, one must understand the limitations of the instrument, its role in the corporation, the laboratory environment, and the parts or gauges it will measure. Of these, understanding the limitations of the instrument, within its particular environment, is almost always the least understood and developed effort. CMM accuracy is usually talked about in generic terms, or as having a manufacturer’s tolerance or some level of volumetric accuracy. The machines are almost always better than that, and just require a good faith effort to extract their talents.
This presentation will describe how NIST gets the best accuracy out of its CMM’s, and how you can tailor these methods to your own laboratory. We will categorize the concepts and discuss what is appropriate depending on the accuracy required for the measurement task. We also will describe some techniques to help find your machine’s true performance limitations. Our experience shows that when an instrument’s performance is properly understood, an operator can design comparative measurement procedures that yield more accurate results while also providing the data necessary for a reasonable estimate of the process measurement uncertainty. We will also discuss the value of maintaining a small collection of diverse gauges solely for the purposes of efficient performance monitoring.