There are 80 or more dynamic seals in a typical automobile that can suffer possible failure. Contemporary manufacturing techniques produce smoother shafts than ever before, and measurement of shafts for smoothness is difficult using traditional means. New instrumentation that meets fabricators’ needs for measuring surfaces with unprecedented precision while simultaneously providing quantitative and repeatable information is needed.
Leakage in rotary dynamic sealing applications is prevented by preloading the lip of the elastomeric seal, making its internal diameter slightly smaller than the shaft diameter. Excessive roughness on the shaft can cause rapid wear on the seal, resulting in leakage. Moreover, the turning processes often used to fabricate shafts may leave a threading pattern (lead). Even if the surface texture of the shaft meets specifications, the presence of a lead angle can generate leakage; any threading causes increased seal wear and also acts to pump fluid across the seal.
Shaft performance is dependent on three different surface parameters: roughness, macro-lead, and micro-lead. Macro-lead refers to spiral type features typical of turned metal parts. Micro-lead is more subtle and refers to high-frequency (i.e., small) pits, scratches, and features that may appear pseudo-random but that have an overall tilt with respect to the shaft axis. These may collectively create a net lubricant pumping action during shaft rotation.
Traditionally, surface roughness was obtained by measuring a two-dimensional (straight line) profile using a contact stylus profilometer, such as the Bruker Dektak series of instruments. However, today’s higher-performance seals and shafts ideally need a three-dimensional surface view, preferably over the entire area that will form the seal.
The situation with lead is even more limiting. In the prescribed method (RMA OS-1-1 rev. 2004), both ends of a cotton quilting thread of 0.23-mm (0.9) diameter are attached to a 30-g (1.1-oz) weight. The thread loop is draped across the shaft (see photo). The shaft is then rotated in both directions to see if the thread is forced to move along the shaft by unwanted grooves produced during the turning process. This movement is measured using a traveling microscope or vernier calipers.
But engineers now require an alternative method for several reasons.
First, the string method cannot measure micro-lead and is only qualitative, not quantitative; results are operator-dependent and very difficult to reproduce with a high level of precision. Second, it measures an average value for macro-lead over a length of the shaft; it cannot measure lead angle precisely over the narrow band that forms the lip seal. But most important, some automobile manufacturers are tightening the lead angle specification on many of these shafts from the current value of ±0.05° to a nominal value of 0° as required by the ISO standard ISO 6194-1:2009. This is virtually impossible to verify with the string-and-weight method.
An alternative is the 3D optical microscope, a gauge-capable surface-metrology tool that has none of these limitations. Moreover, it is fast enough to support dense sampling and has been automated for shaft-measurement applications so that results are operator-independent—particularly when it comes to shaft balance and alignment in the test fixture.
First successfully evaluated for this application by the Dana in 2002, the 3D Optical Microscope is a special type of microscope equipped with a digital camera and computer. Based on a technique called white light interferometry, when the microscope is focused precisely on a surface under test, dark lines appear in the digitally recorded image. These form a detailed surface contour map of great precision, delivering subnanometer accuracy in the vertical direction, when needed.
To meet the growing demand for precision shaft metrology, Bruker’s Nano Surfaces Division designed a dedicated version of the 3D Optical Microscope, the NPFLEX-LA optical microscope, that simultaneously measures macro-lead—regardless of part alignment—while simultaneously measuring 3-D roughness parameters. In addition, a recently developed algorithm accurately determines micro-lead from a global analysis of all surface features.
Gage repeatability and reproducibility of the NPFLEX-LA has been extensively evaluated with a series of tests using shafts with a known surface finish created by plunge grinding. For one of the repeatability tests, shafts were mounted on the NPFLEX-LA and the measurement sequence run 30 times without removing or replacing the part. For these shafts, the one-sigma standard deviation of the lead angle results was less than 0.005° (see graph). The one-sigma standard deviation for the most important surface roughness parameter (Sa) was less than 1.4 nm (0.1 µm).
Next, system reproducibility was studied by having a single operator load and unload the shaft between 10 measurement sequences. The starting location of measurements was not controlled in any way between runs to represent an operator randomly approaching the system with a part to be tested. In this case, the one-sigma standard deviation of lead angle across the runs was 0.02°.
Javier Vera, Erik Novak, and Andrew Masters of Bruker Nano Surfaces Division wrote this article for Automotive Engineering.