In modern drives, the requirements regarding low noise, transmittable torque and low weight are constantly increasing. As soon as a noise excitation occurs in the transmission, tooth flank ripples at the gears are frequently responsible. These ripples usually have an amplitude between 0,10 µm and 1,00 µm. Fine structures like these can only be reliably detected with high-precision gear measuring machines. However, there are no validation and traceability options for the calculated ripple characteristics.
For this reason, a ripple standard with a 69th ghost order was manufactured, allowing to check the accuracy of ripple measurement on gear measuring machines. The standard is tested regarding its practicality in comparative measurements, to be able to evaluate and improve the accuracy of the ripple measurement used today. The investigation of a large number of gear measuring machines indicates that they achieve a very low variance.
Manufacturing and analysis of gear ripple standard
The ripple standard was produced with a periodic torsional oscillation of the workpiece rotary table of frot = 69 1revwkpc, which results in a 69th ghost order. In cooperation with the machine tool manufacturer Kapp, the ripples were applied precisely to the standard.
The evaluation of the ripples is done with an analysis software that combines all measured curves around the circumference into one curve and then calculates with compensation sine functions the spectrum of circumferential ripples. The amplitude of the 69th order is between 0,41 μm – 0,47 μm.
The comparative measurements were carried out on 29 gear measuring machines from Klingelnberg in the production environment and in measuring rooms. Only Klingelnberg measuring machines were used, as the corresponding evaluation software for ripples was only available there. The measurements were done at the manufacturer, at the HAW Hamburg and in the production environment of Magna (GETRAG) and ZF. Figure 1 shows a ripple standard during a measurement.
Figure 1: Measuring of a gear ripple standard
Two objectives are pursued by carrying out comparative measurements. The first objective is the determination of reference values of the orders and their respective amplitudes being present on the standards. As to the reference value the measurement results of 12 newer measuring devices of the series P26 to P100 were used, not being older than 4 years. The second objective is to compare the amplitude acquisition accuracy for all tested measuring devices to the reference value. For this purpose, all measurement results were used for comparison, also those obtained on older measuring devices (i.e. age 4 years and older, such as PNC35 measuring devices, which are easily exceed 16 years).
In the comparative measurements the amplitudes of the applied 69th order are reliably detected by all measuring instruments. The spread around the mean value is extremely small, with a maximum of 0,010 μm only. In the further course of the evaluation, all-systematic orders are considered, being potentially present on the normal. The evaluation of the confidence intervals shows a very low variance as well as the precise registration of the orders.
In summary, a highly accurate gear ripple standard was produced for the inspection of gear measuring devices. In comparative measurements, a large number of measuring instruments were used to determine the orders present on the standard and their amplitudes. The measurement results show identical and comparable amplitudes with a very small spread around the respective mean value. The new standard makes it possible to measure noise-causing ripples on gears worldwide comparable.
The authors gratefully acknowledge financial support by the Forschungsvereinigung für Antriebstechnik e.V. (FVA) [FVA-Book 1308] for the achievement of the project results.
Left: M. Sc. Thies Kahnenbley, HAW Hamburg/Institute of Production Engineering
Right: Prof. Dr.-Ing. Günther Gravel, HAW Hamburg/Institute of Production Engineering