Contoured, free-form shapes are being used more frequently by product designers who see inherent benefits and functionality in them. Almost any surface that interacts with natural elements works better as a contour. Automobile bodies are more aerodynamic when they are streamlined, making cars more fuel efficient. Equipment and devices shaped for human use, items like telephone handsets, computer keyboards, or tennis rackets, are more functional when designed with curved surfaces.
Not only are these naturally-shaped products more practical and attractive, they are often less costly to manufacture. The ability to make parts conform to natural contours often eliminates the need to construct a product from several geometric workpiece shapes, reducing machining and assembly time.
For these reasons and others, more and more manufacturers are designing component sub-assemblies and complete end-use parts using free-form shapes. Free-form shapes and complex part geometries, however, present special design and manufacturing challenges. These shapes are difficult to machine consistently, making initial design and process control critical issues.
Capturing Data for Manufacturability
This is where the gauging system can help. It's the job of the gauging system to capture workpiece dimensions for initial machine tool path programming and, via inspection, for process control purposes. Unlike prismatic parts that require a minimum number of data points to establish dimensional information, accurate measurement of free-form shapes requires the compilation of a massive number of data points. For example, the diameter of a circle can be defined with a minimum of three data points. Form measurement of a circle could require 3,600 to 4,000 data points, depending upon the level of definition required.
Coordinate measuring machines have been used for three decades to gather dimensional data for inspection and process control purposes. CMMs collect detailed dimensional data by moving a sensing device called a probe along workpiece surfaces. Most CMMs acquire data using a touch trigger probe that contacts individual points on the workpiece. This single-point measurement technique can generally collect data at a maximum rate of about 50-60 points per minute.
The traditional method of measuring the dimensions of complex shapes on a coordinate measuring machine, the twist in a turbine engine fan blade or the curvature in a mold cavity, is to obtain data points individually and compare them to a blueprint. Variation in measuring results occurs due to the dispersion of probe points on the workpiece and the CMM's positioning accuracy. The analysis of the data is handled by algorithms whose results are dependent on factors such as noise in the data and systematic deviations of the measured surface from the ideal.
Scanning Adds a New Dimension to Data Gathering
Today, however, there are CMMs that can automatically measure the 3D shape and form of workpieces, and use that information in conjunction with CAD systems to provide insights into the manufacturing operation. These machines are called scanning CMMs, and until recently they were expensive and sometimes difficult to use. A new generation of machines, such as Brown & Sharpe's GLOBAL series CMM, offers advanced scanning capability in an affordable package that makes it possible for virtually any shop to use high speed dimensional data gathering technology.
Scanning is simply a way of automatically collecting a large number of data points to accurately define the shape of a part. Scanning provides more data points for analysis in the same amount of time, or often in less time, than conventional single-point touch trigger probe techniques. Bridge type CMMs using conventional point-to-point probing technology can gather data points at speeds of up to 1.7 seconds per point. If the minimum number of data points required to define a diameter is between 6 and 16, it takes a CMM 10.2 to 27.2 seconds to generate reproducible results. This seems fast, but when compared with the speeds possible with scanning, it is slow.
Using high speed scanning to measure the same diameter produces results calculated with 600 plus data points calculated in less than 4 seconds. The high speed data collection rate makes it possible to determine dimension, position, and form simultaneously with high repeatability and accuracy.
Scanning CMMs use two distinct types of modes to collect data, open-loop and closed-loop. Open-loop scanning is a high-speed technique used with continuous analog probes on parts whose geometry is mathematically defined, allowing the machine to be programmed to maintain contact with the surface. The probe is driven along the known path of the nominal surface and records the magnitude of the error between the actual surface and the nominal. Verifying the dimensions of a mold after machining is a typical application of open-loop scanning techniques. In a closed-loop system, the probe detects changes in surface direction of the part and automatically adjusts itself to maintain contact with the part surface. Closed-loop scanning is slightly slower than open-loop scanning, but attains similar accuracies. It is particularly useful for digitizing unknown complex shapes for reverse engineering applications.
Gathering More Data Faster
Advances in data gathering technology over the last three years have made scanning an even better way to gather data. The advances fall into three categories: controllers, data gathering sensors and software.
There is a demand for controllers that can execute more clever routines than in the past to provide advanced data gathering capability for reverse engineering applications. New controllers can react more intelligently to changes in direction of the scan, such as when the probe encounters a hole in the workpiece. When a surprise in the measuring path like this occurs, new controllers can automatically set up an action, such as a spiral search, to quickly and safely come back into contact with the part and continue with the inspection routine.
This is where the closed-loop and open-loop scanning modes overlap. In newer systems, each scan defines the scan to come, extrapolating the recently acquired information to predict what the next pass is going to look like. If contact with the part is lost, the system shifts to an unknown mode. Once contact is re-established, the system shifts again, this time back to a defined scanning path.
Controller advances are an answer to users who would like to simply place a part in the CMM measuring volume, stake out four corners of the part and walk away, letting the machine controller handle the entire measuring routine. CMM manufacturers are working toward this "set it and forget it" mode.
Advances in smart controller technology are coming rapidly now as a result of a proliferation of open architecture controllers using inexpensive PC chips. This lets developers use standard PC programming languages rather than proprietary controller languages, which, in turn, allows more motion control experiments in less time than in the past. This makes it very easy to try new controller ideas. Because these newer controllers are PC based, they can use a network to pass data directly to a computer for analysis at speeds up to 1000 points per second. The traditional RS-232 interface has a limit of about 25 points per second.
Optical Sensors in the Spotlight
In the area of data gathering sensors, the balance is shifting from contact sensors to non-contact sensors. Non-contact sensors offer the advantages of speed, plus they can be used in applications where contact with the surface of the workpiece would cause damage.
As the cost of CCD camera chips comes down, optical sensors have become a practical choice for scanning applications. Optical sensors can be affected by the surface finish of the workpiece, but for most reverse engineering applications, non-contact sensors are the best solution since they can rapidly gather large amounts of data accurately and can digitize details that may be too fine for a traditional ball probe.
The ratio between throughput and accuracy is always a concern in scanning operations. Generally speaking, the faster the data is collected, the less accurate it is. However, that is changing. Improved accuracy and throughput ratios are the result of improvements in CMM drive mechanisms. Stiffer drive mechanisms that can react more rapidly to direction changes and advanced controllers that produce optimized motion paths expressly for each machine's individual characteristics have greatly improved the accuracy and repeatability levels of scanning CMMs.
Today, users are accepting current accuracy levels and are looking for advances in throughput. They want data as quickly as they can get it. The focus is on speed with reliable accuracy.
Software Is Critical to Accuracy
Because the measurement of forms requires the collection of so many data points, there are special scanning functions that aren't found in other CMM software packages. For example, scanning software must have a filtering ability to distinguish between subtle changes in the surface direction and variability in the workpiece surface finish, such as the rough area on a turbine blade. Filters are also necessary to eliminate the effects of vibration.
Radius correction (parallel curve) is another special function of scanning software. Non-scanning CMMs use probe center coordinates for measurement in that the data generated by the machine is the location of the center point of the ball. In scanning applications, the data must be shifted by the radius of the probe, using the parallel curve function, to represent the real surface of the workpiece. During analysis, spline functions are used to remove the mismatch between the scanned points and the nominal points so that deviations from the nominal to the actual can be calculated.
Scanning opens up new horizons in manufacturing metrology. With inexpensive scanning CMMs and the capability to effectively harness the power of massive amounts of dimensional data through CAD/CAM systems, virtually every manufacturer can exercise a level of dimensional data analysis never before possible. It's the shape of things to come.
