How many dimensional data points are enough to accurately describe a part feature? The key to answering this question is understanding the stability of the manufacturing process. In general, components should be measured only as often as required to ensure the stability of manufacturing processes. This requires identifying and monitoring part features that are critical to the part’s end-use function and developing a strategy to control the dimensions of those features.
The choice of manufacturing technique is the key factor in choosing a process control method. If, for example, the manufacturing process reliably produces a critical bore with good form, its size or position may vary. In this case, control of the size and position will be important but not necessarily roundness or cylindricity control. By contrast, if the machining process produces features with significant form variation (i.e., the variability of the form is a significant proportion of the form tolerance), then understanding where and how the form errors occur becomes important.
Some features may need to mate with other parts for the end-use product to work correctly. In many cases, the form or profile of these features is critical to the functional fit; consequently, the processes used to make these features must be precisely controlled.
Gathering data using scanning techniques is the most efficient and effective way to understand the dimensional properties of part features. In fact, scanning is the only method of efficiently gathering enough data points to accurately describe size, location and form.
Evaluation programs conducted by coordinate measuring machine manufacturers and supported by workplace experience show that the more measurements performed on machined parts, the more accurately the parts can be described dimensionally.
Machined parts have a natural dimensional frequency that is created during the machining process. When a minimum number of data points is collected it’s possible to hit high and low points in the diameter, or all high points. The result is an erroneous description of the diameter. As the following illustration shows, a more complete dimensional description of the diameter can be achieved when more data points are collected.
There’s also a relationship between uncertainty and data density. Uncertainty decreases as the number of data points increases. Therefore, the more data points that are collected to describe a part feature, the more accurately the information describes the dimensional characteristics of the feature.
The question is, "How much data are enough for the specific workpiece feature being inspected?" This is the point at which a gaging strategy should be considered.
Formulating a gaging strategy starts with analyzing inspection needs. If process control is the goal, then the more data points used to describe the workpiece is the more effective the dimensional information is in controlling the manufacturing process. How many data points are enough? The amount of data that accurately defines dimensional variables of the part feature within tolerance specifications--while maintaining acceptable throughput levels--is the right amount of data for process control applications.
Add new comment