In the field of precision gear manufacturing, the processing of spiral bevel gears has always been a complex and critical task due to the intricate geometry and high accuracy requirements. Spiral bevel gears are widely used in automotive, aerospace, and industrial machinery for transmitting power between intersecting shafts with high efficiency and smooth operation. Traditionally, the machining of these gears relies on specialized equipment, such as those from Gleason or Oerlikon, which are often expensive and rely on mechanical systems with change gears. To enhance flexibility and modernization, this study focuses on the numerical control (NC) transformation of a conventional mechanical spiral bevel gear milling machine using a domestic NC system, specifically the GSK 980MDa. This approach aims to leverage cost-effective domestic solutions while maintaining the robust mechanical structure of existing machines. The transformation involves two-axis NC retrofitting, hydraulic system redesign, programmable logic controller (PLC) development, and custom NC programming. Throughout this work, the emphasis is on improving the machining performance and adaptability for spiral bevel gears, ensuring that the modified machine can handle various gear parameters through software adjustments rather than hardware changes.
The core of the transformation lies in understanding the generating motion principle for spiral bevel gears. In a typical spiral bevel gear milling process, the gear tooth surface is generated by the relative motion between a cutter head (equipped with cutting tools) and the workpiece. This motion consists of two synchronized components: the rotation of the cradle (which simulates the generating gear) and the rotation of the workpiece. In mechanical machines, this synchronization is achieved through a complex gear train with change gears, limiting flexibility. For NC transformation, I replaced the mechanical drives with servo motors to enable direct digital control. Specifically, the cradle rotation axis is defined as the X-axis, driven by an AC servo motor, while the workpiece rotation axis is defined as the Y-axis, also driven by a separate AC servo motor. The existing gear transmissions for these axes are retained to preserve mechanical rigidity, but the control is shifted to the NC system. This two-axis linkage allows for precise generation of spiral bevel gear tooth profiles based on digital parameters rather than physical gears.
The generating motion can be mathematically described to ensure accuracy. Let the cradle rotation angle be denoted as $\theta_X$ (X-axis) and the workpiece rotation angle as $\theta_Y$ (Y-axis). For a given spiral bevel gear design, the ratio between these angles, known as the roll ratio or generating ratio, is constant during the cutting process. This ratio $R$ is defined as:
$$ R = \frac{\Delta \theta_Y}{\Delta \theta_X} $$
where $\Delta \theta_X$ and $\Delta \theta_Y$ are incremental rotations. In practice, the motion follows a predefined path: starting from an initial position for alignment, the cradle and workpiece move synchronously through the generating arc, and after each tooth is cut, the workpiece indexes to the next position. The angles involved can be derived from gear geometry parameters such as number of teeth, spiral angle, and pitch. For instance, if the generating arc covers an angle $\alpha$ on the cradle, the corresponding workpiece rotation $\beta$ is given by $\beta = R \cdot \alpha$. After cutting, for indexing to the next tooth, the workpiece rotates an additional angle $\gamma$ based on the number of teeth and the indexing pattern (e.g., skip indexing). These relationships are crucial for NC programming and are implemented through parametric equations.
To manage the feed motion during machining, the hydraulic system of the milling machine was redesigned. The feed motion controls the engagement and disengagement of the workpiece relative to the cutter head. In the original machine, this was mechanically operated, but for NC integration, I implemented a PLC-controlled hydraulic system. The system includes components such as solenoid valves, flow control valves, and hydraulic cylinders. The primary functions are manual control for setup and automatic control for machining cycles. In manual mode, operators can adjust the workpiece position using manual switches, with speed regulated by throttle valves. In automatic mode, the PLC triggers the hydraulic actuators based on NC program commands to perform precise feed and retract motions during each cutting cycle. This dual-mode design enhances operational flexibility and safety.
The hydraulic control logic is summarized in Table 1, which outlines the solenoid valve states and corresponding actions. The system uses three-position four-way valves for main control and two-position four-way valves for safety overrides. For example, during automatic cutting, when a tooth is completed, the NC system sends a signal to retract the workpiece, and after indexing, it signals to advance for the next cut. Limit switches ensure positional accuracy and prevent overtravel. The integration of hydraulic control with NC allows for seamless automation, critical for efficient production of spiral bevel gears.
| Solenoid Valve | State | Function | Control Mode |
|---|---|---|---|
| Valve 3 (Manual Advance) | Energized | Workpiece advances | Manual |
| Valve 4 (Manual Retract) | Energized | Workpiece retracts | Manual |
| Valve 5 (Auto Advance) | PLC Signal | Auto advance after index | Automatic |
| Valve 6 (Auto Retract) | PLC Signal | Auto retract after cut | Automatic |
| Valve 7 (Safety Override) | Limit Trigger | Emergency retract | Automatic |
A key aspect of the NC transformation is the development of custom M-codes through PLC programming. The domestic GSK 980MDa NC system supports PLC integration for auxiliary functions. I used binary decoding techniques to assign specific M-codes for hydraulic control, such as M8 for workpiece advance and M9 for retract. The PLC processes these codes by monitoring NC signals: when an M-code is issued in the program, the NC system sends a binary code and a strobe signal to the PLC. The PLC decodes this and executes the corresponding action, such as activating solenoid valves. After completion, the PLC sends a finish signal back to the NC system to resume program execution. This method ensures synchronized control between NC motions and hydraulic actions. The timing sequence for M-code execution is critical to avoid conflicts, as shown in Figure 1 (though figures are not referenced directly, the concept is described). The developed M-codes are listed in Table 2, along with their functions relevant to spiral bevel gear machining.
| M-code | PLC Output | Function | Description |
|---|---|---|---|
| M8 | R4.0 | Workpiece Advance | Moves workpiece into cutting position |
| M9 | R4.1 | Workpiece Retract | Retracts workpiece after tooth cut |
| M10 | R4.2 | Cutter Head Start | Activates cutter head motor |
| M11 | R4.3 | Cutter Head Stop | Stops cutter head motor |
| M12 | R4.4 | Coolant On | Enables coolant flow |
| M13 | R4.5 | Coolant Off | Disables coolant flow |
| M14 | R4.6 | Indexing Enable | Allows workpiece indexing |
| M15 | R4.7 | Emergency Stop | Halts all motions for safety |
The NC programming for spiral bevel gears was designed to be parametric, allowing easy adaptation to different gear specifications. Instead of hard-coding motions, I developed a macro-based program where key parameters are defined as variables. These variables include the number of teeth ($Z$), module ($m$), spiral angle ($\beta_s$), pressure angle ($\alpha$), roll ratio ($R$), and indexing skip number ($N_s$). The main program handles initialization and calls a subprogram that executes the generating and indexing cycles. The relationship between cradle and workpiece motions is computed in real-time using these variables. For instance, the generating arc angle $\alpha_g$ for each tooth is derived from the gear geometry and machine settings, and the corresponding workpiece rotation is $\beta_g = R \cdot \alpha_g$. The indexing angle $\gamma_i$ for skip indexing is given by:
$$ \gamma_i = \frac{360^\circ}{Z} \cdot N_s $$
where $Z$ is the number of teeth and $N_s$ is the skip number (e.g., 1 for consecutive teeth). This parametric approach eliminates the need for physical change gears, making the machine highly versatile for producing various spiral bevel gears.
To illustrate the programming logic, consider a sample NC code structure for cutting a spiral bevel gear. The program begins with safety checks and tool positioning, then enters a loop for each tooth. In the loop, the cradle and workpiece move synchronously along the generating path, controlled by G-code interpolations for the X and Y axes. After the cut, an M-code triggers retraction, followed by indexing via Y-axis rotation, and another M-code triggers advance for the next cut. The loop continues until all teeth are machined. The use of macros allows operators to input parameters through a user-friendly interface on the NC system, simplifying setup for different spiral bevel gears. This interface displays prompts for key variables, ensuring accurate data entry and reducing human error.
For validation, I conducted machining tests on the transformed milling machine. The test gear was a spiral bevel gear with the following parameters: number of teeth $Z = 35$, module $m = 5 \text{ mm}$, pressure angle $\alpha = 20^\circ$, spiral angle $\beta_s = 35^\circ$. The machine settings included a cutter radius of 114.3 mm, blade angles of $22^\circ$ for inner blades and $-18^\circ$ for outer blades, and a roll ratio of 1.413. The machining process involved roughing and finishing passes, with careful monitoring of cutting forces and tool wear. The results showed that the two-axis NC system accurately generated the tooth profiles, with smooth motion and consistent indexing. After machining, the gear was measured on a coordinate measuring machine (CMM) to evaluate deviations from the theoretical surface. The measurement data indicated that the maximum profile error was 0.0792 mm, occurring near the toe and root regions, which is acceptable for many industrial applications. This demonstrates the feasibility of using a domestic NC system for spiral bevel gear machining.
The transformation also involved considerations for tool wear and cutting optimization. Spiral bevel gears require precise control of cutting conditions to maintain surface quality and tool life. I analyzed the cutting forces using dynamometer measurements during test runs. The forces were found to be within acceptable limits, and the NC system’s ability to adjust feed rates dynamically helped minimize tool stress. Additionally, the hydraulic system’s automatic control reduced idle time between cuts, improving overall productivity. These aspects are crucial for the economic viability of machining spiral bevel gears in small to medium batches.
Further technical details include the servo motor specifications and feedback systems. The X-axis and Y-axis servo motors were selected based on torque and speed requirements for generating motions. The NC system uses encoder feedback for closed-loop control, ensuring positional accuracy within microns. The motion control algorithms incorporate acceleration and deceleration profiles to avoid jerks and ensure smooth transitions, which is vital for the complex kinematics of spiral bevel gear generation. The mathematical model for the generating motion can be extended to include corrections for machine deflections or thermal effects, but in this transformation, the basic model sufficed for achieving the desired accuracy.
In terms of system integration, the PLC not only handles M-codes but also manages auxiliary functions like lubrication, cooling, and safety interlocks. The ladder logic program includes routines for error handling, such as detecting hydraulic pressure drops or motor overloads. This enhances the reliability of the machine when producing spiral bevel gears continuously. The NC system’s memory stores multiple programs for different gear types, allowing quick changeovers. The user interface was customized to display real-time status of axes, hydraulic systems, and cutting parameters, making it easier for operators to monitor the machining of spiral bevel gears.
To summarize the motion parameters and their formulas, Table 3 provides a comprehensive list of key variables used in the NC programming for spiral bevel gears. This table serves as a quick reference for setting up the machine for different gear designs, emphasizing the flexibility gained through NC transformation.
| Parameter | Symbol | Formula | Description |
|---|---|---|---|
| Number of Teeth | $Z$ | Input variable | Total teeth on the gear |
| Module | $m$ | Input variable | Size factor for gear dimensions |
| Spiral Angle | $\beta_s$ | Input variable | Angle of tooth spiral |
| Pressure Angle | $\alpha$ | Input variable | Angle between tooth profile and tangent |
| Roll Ratio | $R$ | $R = \frac{Z_{\text{cradle}}}{Z_{\text{work}}}$ (theoretical) | Ratio of cradle to workpiece rotation |
| Generating Arc Angle | $\alpha_g$ | $\alpha_g = f(m, \beta_s, \text{machine settings})$ | Cradle rotation per tooth cut |
| Workpiece Rotation per Cut | $\beta_g$ | $\beta_g = R \cdot \alpha_g$ | Workpiece rotation during generating |
| Indexing Angle | $\gamma_i$ | $\gamma_i = \frac{360^\circ}{Z} \cdot N_s$ | Workpiece rotation for tooth indexing |
| Cutter Radius | $r_c$ | Input variable | Radius of the cutter head |
| Blade Angles | $\theta_i, \theta_o$ | Input variables | Inner and outer blade angles for cutting |
The success of this NC transformation opens avenues for further enhancements. For instance, adaptive control could be integrated to adjust cutting parameters in real-time based on sensor feedback, further improving the quality of spiral bevel gears. Additionally, the domestic NC system can be networked for data logging and remote monitoring, aligning with Industry 4.0 trends. The cost savings from using domestic systems, coupled with the retained mechanical robustness, make this approach attractive for manufacturers seeking to upgrade legacy machines for spiral bevel gear production.
In conclusion, the two-axis NC transformation of a spiral bevel gear milling machine using a domestic system has proven effective. By replacing mechanical drives with servo motors, redesigning the hydraulic control, developing PLC-based M-codes, and implementing parametric NC programming, I have enabled flexible and precise machining of spiral bevel gears. The test results confirm that the transformed machine meets accuracy requirements, with deviations within acceptable limits. This work demonstrates the viability of domestic NC systems for complex gear machining tasks, offering a cost-effective alternative to imported solutions. Future work could focus on optimizing cutting strategies for different materials and expanding the system to handle more complex spiral bevel gear designs, such as those with modified tooth surfaces for noise reduction or higher load capacity.