Although new technologies are adopted in the semiconductor industry within months, changes within the world of metrology take much longer. It’s safe to say, however, that these new developments promise to be well worth the wait. The new core technologies will enable measurement with a precision 100 times better than what’s currently possible.
This article takes a look at some of the advances in surface and form measurement that will revolutionize the way we measure during the next ten years. The new breed of micro-measurement products focus on measurement automation and efficiency, increased accuracy, advances in nanomeasurements and new sensor technologies. Tangible core technologies include:
Super high-fidelity sensors
Advanced structural stability
Robust estimation and filtering
As their name implies, linear motors move in a linear direction vs. that of conventional rotary-type motors. Most sur-face roughness and contouring measurement is accomplished by traversing a stylus in a linear fashion across the feature to be measured. Using conventional rotating motors to achieve a linear motion requires components such as gears, screws, belts and spindles. The parts in the transmission, gearbox or drivetrain add measurement noise and tend to wear out over time. Up to 90 percent of measurement noise comes from the dependent motion of the traverse system drive unit.
Linear motor systems, by contrast, significantly reduce measurement noise as it relates to the drive unit. They obtain more real measurement and less system noise. The signal-to-noise ratio is increased, which translates into better meas-urements and fewer errors. In addition, fewer components in the system increase reliability, resulting in less downtime.
Incorporating measurement systems with super high-fidelity sensors significantly improves the precision and accuracy of probes. In the case of inductive sensors such as linear variable displacement transducers, linearity is increasing, and real resolution has increased 100 times from 10 years ago. Precision tolerance on windings and materials, 20-bit and greater analog/digital signal processing and systematic error correction have taken sensor resolutions and accuracies down to the nanometer level. The fidelity of inductive systems is infinite, which provides an advantage over other sensing technologies.
Digital sensors are also becoming more precise. Improvements are being made in interpolation, grating accuracy (courtesy of semiconductors) and detector dimen sions. Digital sensors usually apply some grating or an etched pattern and subdivide this pattern to interpolate a higher resolution. Technically, digital systems are high-speed sensors and counters that change energy pulses (i.e., light or electrons) into ones and zeros. Typical sensor response and counting frequencies have increased from hundreds of hertz to gigahertz and greater, making digital sensors affordable, fast and accurate. They’re now used in most drive systems, where they measure contour, roundness or surface roughness.
All measurement systems inspect both the intended--usually a feature on a workpiece--and the unintended, such as ambient and system vibration or system dimensional changes. With the advent of finite element analysis, measuring systems are being developed in which advanced structural stability can be guaranteed, and system performance characteristics are characterized prior to producing the first piece.
In the case of surface roughness, in which a high system-resonant frequency is desirable, the search for an optimum manufactured measuring product is first tested in a computer using finite element analysis. For contouring and roundness systems, poor thermal characteristics can cause datums to move and a system’s component relationships to change, which can be detrimental. For example, a column, base and rotary table relationship changes due to temperature fluctuations on the production floor, causing a perfect cylindrical workpiece to be measured as a tapered cylinder. This is why most quality-conscious manufacturers use rigid air bearings and materials like granite and ceramic in their high-end nanometric roundness system bases and measuring datums.
ISO/TC 213’s dedicated work group, WG 15, devotes its time and energy to new filtering and form-removal techniques such as robust estimation and filtering. Filters and outlier-removal techniques are required to divide measured data into roughness and waviness, or to remove outliers from a geometric feature. One particular filtering technique gaining momentum is the robust spline gaussian filter. It combines the power of outlier elimination using robust estimation techniques with the edge-to-edge benefits of spline filtering. Today’s digital filters require both a startup length and ending length so that the typical measurement is 0.18 in. (4.8 mm), but parameters are evaluated on only 0.15 in. (4.0 mm) of this surface.
This isn’t the case with spline filtration because no startup or ending lengths are needed. Full evaluation of small measurement surfaces is possible. The “robust” aspect of the robust spline gaussian filter means surface extremes such as dust, pores or scratches won’t distort the filtered mean line. This virtually eliminates filter overshoot, an occurrence where valleys turn into peaks and vice versa (also known as “Gibbs phenomena”). Robust spline gaussian filters provide a more exact representation of the measured surface with virtually no filter-induced error.
This technology has increased system accuracy while potentially reducing cost. Most systems today use computer-aided accuracy (CAA) compensation to eliminate system errors. However, compensation doesn’t allow gage manufacturers to build inferior systems. For CAA-compensated equipment to survive the test of time, measuring systems must be more repeatable, which is a good reason to include a short equipment variation test in machine acceptance criteria. CAA techniques have increased roundness accuracy to the nanometer level, and roughness and/or contouring system straightness to 10 nm per millimeter or better.
Today’s micro-measuring systems are quicker, more versatile and require less operator intervention. In the past, if a gage operator had the task of measuring a crankshaft, it meant setting up the part for every measurement at each location to be measured. On an eight-cylinder crankshaft, for example, this could include, at a minimum, eight pin locations, five main bearing locations and an oil seal--14 locations in total. Setup and measurement could take one to two hours with only a single measurement per location.
With today’s measurement automation capabilities, a full CNC surface-roughness system can measure those same features in four locations (0°, 90°, 180°, 270°), or 56 locations in total, in significantly less time with far less appraiser variation and much better representation of the actual workpiece.
Measuring system modularity and multisensor capability allows systems to grow as measurement requirements change and need for throughput increases.
When purchasing a new system, consider the core technologies described above as a good place to begin your evaluation.
For surface-roughness systems:
Linear motor technology drastically reduces system noise and increases signal-to-noise ratio.
High-fidelity sensors allow for larger measuring ranges and greater precision.
A robust spline gaussian filter allows users to measure more, throw away less and eliminate outliers all at once.
In high-volume situations, consider measurement automation options.
For roundness systems:
Select a stable base made of granite or other thermally stable material. System instability can create problems in a roundness system when measuring cylindricity, where any change in the relationship of column or R-axis to the part will affect results.
Computer-aided accuracy can increase overall rotational accuracies 10 times. Generally, air spindles are superior to mechanical-bearing spindles and are much more accurate, although ball bearings are useful when measuring a part that weighs greater than 100 pounds.
For contouring systems:
Choose a system with multisensor capability.
Surface roughness and contour analysis in a single detector is efficient.
Interchangeable detectors offer greater flexibility. Make sure that the system can be easily upgraded with added detectors as they become available.
Higher data density is important when measuring contour and surface roughness in a single measurement. Look for systems that acquire data with sampling pitches capable of submicron accuracy.
Robert Wasilesky is the business manager for surface, form and geometry products at Carl Zeiss IMT Corp. He is a member of the ASME B46 Surface Roughness Committee, ISO/TC 213/WG 15 Filter and Extraction, ISO/TC 213/WG 16 Areal Methods, and the ASME B89.3.1 Roundness Committee.