OPTIMIZATION OF GEAR TOOTH SURFACES

During the evolution of gearboxes, unwanted noise has been one of the most discussed subjects. Increasing demands on weight reduction and co mpactness of gearboxes, while preserving the same torque capacity, req uire optimal design of all parameters involved. Quiet running is espec ially desirable in a modem automotive gearbox. The very complicated ge ometry of helical gears, engenders the problem of a multi-variable opt imization procedure. Errors of the microform as well as the macroform of the gears highly influence the sound activation. Effects of the mac roform have been studied by many researchers and are widely known and predictable. Normally, finishing processes, such as grinding and honin g, are used to achieve the predicted macroform of a gear tooth. As a c onsequence of applying these processes, the microform of the tooth is accomplished. This paper is concerned with the characterization of the gear tooth surfaces obtained with such processes. The characterizatio n is performed with respect to the surface functional properties. Thes e parameters which are interesting due to the noise activity, were ide ntified when the literature available in the field of interest was rev iewed. 3D measuring technique, form removal, functional filtering, and 2D spectral density analysis have been utilized in order to extract t his information as surface functional parameters. Since the main aim o f this paper is to discuss how to produce gears with optimal noise con sequences, the principles of RZP-grinding, Fassler honing and Hurth gr een-shaving are discussed. During this work, the most significant proc ess parameters have been identified utilizing two-level fractional fac torial experiments, based on 3D evaluation of surfaces. The parameters identified together with the process principles were then used to pre dict the surface characteristics. Optimal strategies concerning all pr ocesses for production of low noise gears are suggested. The validity of arguments and results regarding the RZP-process was examined, while gears from a real manufacturing shop-floor were evaluated. (C) 1998 E lsevier Science Ltd.