Accuracy analysis of workpiece and tool in the hobbing process using Taguchi Method
Improvement of vehicle interior quietness in recent years has made it necessary to manufacture low-noise gears at low cost for use in automotive transmissions. In order to meet this requirement, at JATCO we adopted a gear honing process as the finishing method following heat treatment in order to deal with the sources of gear noise.
In the past, gear manufacturing generally proceeded in turn through the processes of "hobbing -> shaving -> heat treatment -> honing". However, for the purpose of reducing manufacturing costs, some JATCO's gear processes have eliminated the gear shaving process in recent years and now manufacture gears in a series of processes consisting of "hobbing -> heat treatment -> honing". This simplified approach is referred to here as "hobbing + honing" method.
However, this "hobbing + honing" method has certain tendencies. First, it increases the amount of stock that must be removed from the tooth surface by honing. Second, because less corrective force is applied in honing than in grinding, it is more difficult to eliminate the pitch error and tooth space run-out that occur in the hobbing process. Accordingly, an essential requirement for the “hobbing + honing” method is to improve the machining accuracy of the hobbing process.
Against this backdrop, in this study, the newly developed hobbing simulation was combined with the Taguchi Method to make clear the necessary accuracy required of the hobbing process in gear manufacturing. The results clarified key control factors in the hobbing process for the mass production of transmission gears.
In addition to the tooth profiles and tooth traces, the developed hobbing simulation can also simulate pitch error and tooth space run-out for the purpose of identifying the accuracy required of the hobbing process in the “hobbing + honing” method. It can also take into account the hob shift due to the axial shifting of the hob for machining each workpiece in mass production such that machining is done with a new and separate cutting blade.
Investigation of function in the hobbing process
In the "hobbing + honing" method, the machining accuracy of the hobbing process has a large effect on gear accuracy in the honing process. One aspect in particular is that an insufficient amount of stock is removed from the tooth surface by honing, so residual stock remains on the hobbed tooth surface. Such residual stock on the tooth surface causes gear accuracy defects following the honing process, which become the cause of gear noise and abnormal noise in vehicles. Therefore, as the evaluation parameters of function in the hobbing process, we selected fHβ variation (i.e., variation in helix angle error of all the teeth) and fFβ Avg. (i.e., average tooth trace undulation of all the teeth), which have a large impact on gear surface stock among the factors affecting gear accuracy.
The machining conditions, hob specifications, workpiece run-out, and hob run-out were investigated in order to make clear the effect of each of the influential factors of the hobbing process on gear accuracy. Specifically, (A) the feed direction (conv./climb) and (B) the feedrate (mm/rev) were examined as machining conditions, (C) the number of hob threads and (D) the number of gashes as hob specifications, (E) the outer diameter run-out and (F) the end face run-out as workpiece run-out, and (G) the right end run run-out and (H) the left end run-out as hob run-out. Each level of these factors was determined in a range where the hobbing process was viable.
In addition, three factors were selected as being uncontrollable in mass production. These were (a) the phase difference (2 levels) between workpiece outer diameter run-out and end face run-out, (b) the initial phase (8 levels) of hob run-out, and (c) the hob shift position (3 levels).
L18 orthogonal array
An L18 orthogonal array was determined from the selected control factors and their levels. Based on the L18 array and the error factors and their levels, the total number of analyses to be conducted became 864 (= 18 x 2 x 8 x 3).
Analysis results obtained with the Taguchi Method
A factor analysis of the target values (smaller is better) for fHβ variation and fFβ Avg. was conducted using the hobbing simulation and the Taguchi Method. The analysis results revealed two key factors that should be controlled for suppressing fHβ variation in mass production. The first one is to pay special attention to controlling workpiece end face run-out and the other one is to control workpiece outer diameter run-out.
The results also revealed that three factors should be controlled to suppress fFβ Avg. in mass production: (1) to reduce the number of hob threads when selecting the tool; (2) to reduce the hob feedrate; (3) to pay special attention to hob run-out. Notably, it was found that the first two measures here should be selected so as to be within the planned cycle time and that priority should be given to the setting of the hob feedrate because of its large impact on gear accuracy.
The authors developed a hobbing simulation for simulating the positional relationship between the tool and the workpiece, including consideration of workpiece mounting accuracy, in the hobbing process. In the present study, the key control factors in the hobbing process of the “hobbing + honing” method of manufacturing gears were investigated by the developed hobbing simulation and subjected to a factor analysis using the Taguchi Method. At JATCO, we mass produce automatic transmission gears under process control levels determined for meeting the required quality by focusing on the control factors made clear by the results of the present study.
Authors of the article
Dipl.-Ing. Kouji Matsuo, Hardware System Development Department, JATCO Ltd, Kanagawa, Japan
Dr.-Ing. Yoshitomo Suzuki, Engineering Management Department, JATCO Ltd, Shizuoka, Japan
Kenichi Fujiki, Parts Process Engineering Department, JATCO Ltd, Shizuoka, Japan