The demand for high-performance bevel gears in heavy industries such as mining, metallurgy, and marine propulsion has driven significant interest in Klingelnberg cyclo-palloid gear systems. These gears are distinguished by their extended epicycloidal tooth trace and uniform tooth depth, offering advantages in terms of high efficiency, smooth operation, and superior load-carrying capacity. The traditional manufacturing method for these gears employs a sophisticated two-part cutter head system on specialized, and often expensive, imported machine tools. This two-part cutter head allows for independent adjustment of the inner and outer blade groups to correct the longitudinal curvature of the tooth flanks, facilitating localized bearing contact. This specialized gear cutting equipment, with its complex split-spindle mechanism, has presented a significant barrier to widespread domestic production.
To overcome this challenge and enable the production of Klingelnberg-style gears on more common domestic CNC spiral bevel gear cutting machines, this work focuses on the development and validation of a gear cutting methodology utilizing a monolithic (one-piece) cutter head. This approach fundamentally rethinks the flank correction strategy. While the basic kinematic principle of generating the tooth flank via a imaginary crown gear remains, the methodology for achieving a favorable contact pattern must be adapted. This paper details the underlying gear cutting principle, the design of the monolithic cutter head and its blades, the computational method for determining critical machine settings and cutter parameters, and provides experimental verification of the entire process.
1. Gear Cutting Principle with a Monolithic Cutter Head
The fundamental principle of gear cutting for cyclo-palloid bevel gears involves an imaginary generating crown gear, which is a plane gear with straight-sided teeth. The gear pair is generated indirectly: the gear and pinion are each generated in a separate process by a crown gear of opposite hand. A left-hand gear is generated by a right-hand crown gear, and vice-versa. This method theoretically produces fully conjugate tooth flanks.
The tooth trace of the crown gear is an extended epicycloid, produced by the synchronized motion of the cutter head and the workpiece. As the cutter head rotates, its blades trace out the tooth flanks of the imaginary crown gear, which in turn envelopes the tooth form onto the workpiece through the rolling motion of the generation process. The relationship is defined by the roll of a circle of radius $\rho_0$ (cutter head) on a base circle of radius $\rho$ (crown gear).
The key departure from the traditional Klingelnberg method lies in the strategy for longitudinal curvature modification. A two-part cutter head modifies the concave flank of both the gear and pinion by offsetting the axes of the inner and outer blade circles and using different nominal radii. A monolithic cutter head cannot employ this technique. Therefore, the proposed gear cutting strategy is as follows:
- The gear is finished using a duplex (double-sided) gear cutting method with a monolithic cutter head of a single nominal radius $r$.
- The pinion is roughed using a duplex method, but its finishing is performed using two separate single-sided gear cutting operations. The pinion convex flank is finished with a cutter head of reduced effective radius $(r – E_{XB})$, and the pinion concave flank is finished with a cutter head of increased effective radius $(r + E_{XB})$. The value $E_{XB}$ is the crucial radius modification amount calculated to induce the required longitudinal curvature difference for localized contact.
This method effectively transfers the task of longitudinal mismatch correction entirely to the pinion finishing operations, accommodating the physical constraints of a standard single-spindle machine tool.
2. Design of the Monolithic Cutter Head and Blades
The monolithic cutter head must be designed to fit standard machine tool interfaces while replicating the essential gear cutting geometry. A standard Klingelnberg two-part head has an angular spacing of 48° between the inner and outer blade group nodes. For a monolithic head to maintain correct tooth thickness, this angle must be 36° (180° divided by the number of blade groups, typically 5). Therefore, existing blades from two-part systems are incompatible, necessitating a new blade design.
The cutter body typically has ten slots (five for inner blades, five for outer blades) alternately and equally spaced. The blade design parameters, such as hook angle, top rake, clearance angles, tip radius, and pressure angle (equal to the gear normal pressure angle $\alpha_n$), are retained from the standard system. The critical design changes are in the blade’s orientation within the slot and its mounting dimensions.
The side flank of the blade is an Archimedean spiral to maintain constant clearance angles after re-grinding. The projection of the cutting edge onto the cutter head plane must form the specified angle relative to the slot base to achieve the 36° node spacing. The blade’s mounting base distance to its grinding center $Q$ is also specifically calculated for the monolithic head’s slot geometry. When mounted with basic shims, the cutting edge node $M$ lies at the nominal cutter radius $r$.
By referencing the standardized series of two-part cutter systems, a corresponding series of monolithic cutter heads and their dedicated inner and outer blades can be designed, minimizing the required inventory for a wide gear cutting range.
3. Calculation of Machine Settings and Cutter Parameters
The gear cutting process is modeled as the meshing between the imaginary crown gear (generated by the cutter head motion) and the workpiece. The relative positioning and motion define the machine settings.
3.1 Machine Adjustment Parameters
The basic geometric relationship between the cutter head center $O_0$, the crown gear center $O_P$, and the workpiece reference point $M$ is defined by the machine setup. The key parameters are calculated as follows:
- Cutter Tilt Angle (Blade Direction Angle) $\nu$: This is the angle between the projection of the blade cutting edge and the line $MO_0$ in the machine plane.
$$ \nu = \arcsin\left(\frac{0.5 m_n z_0}{r}\right) $$
where $m_n$ is the normal module at the mean point, and $z_0$ is the number of blade groups on the cutter head (e.g., 5). - Machine Center to Back (Cutter Radial Setting) $S$: The distance between $O_P$ and $O_0$.
$$ S = \sqrt{R_m^2 + r^2 – 2 R_m r \cos(90^\circ – \beta_m + \nu)} $$
where $R_m$ is the mean cone distance and $\beta_m$ is the mean spiral angle. - Cutter Swivel Angle $q$: The angle between the line $O_P O_0$ and $O_P M$.
$$ q = \arccos\left(\frac{R_m^2 + S^2 – r^2}{2 R_m S}\right) $$ - Workpiece Tilt Angle $\delta_M$: For uniform depth teeth, this equals the pitch cone angle of the workpiece.
- Blank Offset $E_m$ and Sliding Base $X_P$: For standard generated cyclo-palloid gears, these are typically zero ($E_m=0, X_P=0$) as the workpiece and crown gear axes intersect.
- Machine Base Position $X_B$: This corresponds to the dedendum of the workpiece, positioning the cutter tip plane relative to the crown gear pitch plane.
- Ratio of Roll $R_a$: The generating roll ratio between the cradle (crown gear) and the workpiece.
$$ R_a = \frac{1}{\sin \delta_M} $$
3.2 Calculation of Crown Gear Tooth Trace Curvature
The longitudinal normal curvature of the crown gear tooth trace at the reference point $M$ is fundamental for calculating the mismatch. Using the kinematics of the extended epicycloid generation, the radius of curvature $r_G$ of the trace at $M$ is derived from Bobillier’s construction:
$$ r_G = R_m \sin \beta_m + \frac{R_m \cos \beta_m \tan \eta}{1 + \tan \nu (\tan \beta_m + \tan \eta)} $$
where the auxiliary angle $\eta$ is given by:
$$ \eta = \arcsin\left(\frac{r \cos \nu – R_m \sin \beta_m}{S}\right) $$
The normal curvature $k_{Gv}$ along the tooth trace direction is then:
$$ k_{Gv} = \frac{\cos \alpha_n}{r_G} $$
For the pinion finishing operations, this calculation is performed twice: once with $r$ replaced by $(r + E_{XB})$ for the concave flank, and once with $r$ replaced by $(r – E_{XB})$ for the convex flank, yielding $k_{Pv}$.
3.3 Determination of Cutter Radius Modification $E_{XB}$
The amount of radius modification $E_{XB}$ is not chosen arbitrarily but is calculated to achieve a desired contact ellipse size. A contact length factor $f$ (typically 0.2 to 0.4) defines the ratio of the contact ellipse’s major axis to the face width $b$. The required relative principal curvature $\Delta k_v$ at the mean point for a given $f$ is:
$$ \Delta k_v = \mp \left( \frac{0.0508}{\cos \beta_m (f b)^2} \right) $$
where the upper sign (-) applies for the gear concave/pinion convex pair, and the lower sign (+) for the gear convex/pinion concave pair.
The relative normal curvatures and geodesic torsion between fully conjugate flanks along the tooth trace ($v$) and profile ($t$) directions are:
$$ \Delta k_{12t} = -\frac{\cos^2 \beta_m}{R_m \sin \alpha_n} \left( \frac{1}{\tan \delta_1} + \frac{1}{\tan \delta_2} \right) $$
$$ \Delta k_{12v} = \Delta k_{12t} (\sin \alpha_n \tan \beta_m)^2 $$
$$ \Delta \tau_{12v} = -\sin \alpha_n \tan \beta_m \Delta k_{12t} $$
where $\delta_1$ and $\delta_2$ are the pinion and gear pitch cone angles.
The required crown gear normal curvature modification $\Delta k_{Pv}$ for the pinion generating process is found from the relationship between relative principal curvature and the directional parameters:
$$ \Delta k_{Pv} = \Delta k_v – \Delta k_{12v} – \frac{\Delta \tau_{12v}^2}{\Delta k_v – \Delta k_{12t}} $$
An iterative numerical algorithm is then used to find the value of $E_{XB}$ such that the condition $|k_{Gv} – k_{Pv} \mp \Delta k_{Pv}| \le \epsilon$ is satisfied, where $\epsilon$ is a small tolerance. This $E_{XB}$ ensures the calculated pinion flank curvature will produce the targeted localized contact under load.
3.4 Calculation of Blade Shim Thickness
The blade shim thickness must be calculated to account for differences between the cutter’s basic module $m_0$ and the gear’s normal module $m_n$, as well as for tangential addendum modification $x_s$, hard-skiving allowance $j_{os}$, and backlash $j_{tm}$. The formulas below yield the theoretical shim thickness, which is then rounded to the nearest 0.2 mm for standardization.
Let $c_i$ and $c_a$ be the basic shim thickness for inner and outer blades, respectively. The theoretical shim thicknesses are:
- For the Gear (Duplex Cut):
$$ u_{i2} = c_i + 1.25\tan\alpha_n (m_n – m_0) – x_s m_n + j_{os} – \frac{j_{tm} \cos\beta_m}{4} $$
$$ u_{a2} = c_a – 1.25\tan\alpha_n (m_n – m_0) + x_s m_n – j_{os} + \frac{j_{tm} \cos\beta_m}{4} $$ - For the Pinion Concave Flank (Single-Sided):
$$ u_{i1}^{concave} = c_i + 1.25\tan\alpha_n (m_n – m_0) + x_s m_n + j_{os} – \frac{j_{tm} \cos\beta_m}{4} – E_{XB} $$ - For the Pinion Convex Flank (Single-Sided):
$$ u_{a1}^{convex} = c_a – 1.25\tan\alpha_n (m_n – m_0) – x_s m_n – j_{os} + \frac{j_{tm} \cos\beta_m}{4} + E_{XB} $$
(Note: For single-sided pinion finishing, only one blade type is used per operation).
After blade re-grinding, a wear compensation amount $w$ is added. The final, rounded shim thicknesses $U$ are applied to achieve the correct effective cutter radius and blade positioning for each specific gear cutting operation.
4. Gear Cutting Experiment and Validation
A practical gear cutting experiment was conducted on a domestic H1250C CNC spiral bevel gear cutting machine to validate the proposed methodology. A gear pair was designed with the basic parameters shown in Table 1.
| Parameter | Value |
|---|---|
| Gear Ratio | 19/23 |
| Mean Normal Module, $m_n$ | 11.9968 mm |
| Normal Pressure Angle, $\alpha_n$ | 20° |
| Mean Spiral Angle, $\beta_m$ | 30° |
| Face Width, $b$ | 90 mm |
| Tangential Addendum Modification, $x_s$ | 0.002 |
| Pinion Addendum Modification, $x_1$ | 0.1 |
| Backlash, $j_{tm}$ | 0.25 mm |
| Skiving Allowance, $j_{os}$ | 0.1 mm |
Using the calculation methods outlined in Sections 3, the machine settings and cutter parameters were determined. The calculated radius modification was $E_{XB} = 1.9$ mm. The key gear cutting parameters are summarized in Table 2.
| Parameter | Pinion Convex | Pinion Concave | Gear (Duplex) |
|---|---|---|---|
| Effective Cutter Radius | $r – E_{XB} = 168.1$ mm | $r + E_{XB} = 171.9$ mm | $r = 170$ mm |
| Machine Center to Back, $S$ | 217.897 mm | 219.181 mm | 218.529 mm |
| Cutter Swivel Angle, $q$ | 46.483° | 47.582° | 47.034° |
| Workpiece Tilt, $\delta_M$ | 39.56° | 39.56° | 50.44° |
| Ratio of Roll, $R_a$ | 1.57015 | 1.57015 | 1.29708 |
| Blade Shim (Inner), $U_I$ | – | 0.8 mm | -3.0 mm |
| Blade Shim (Outer), $U_A$ | -3.2 mm | – | 1.2 mm |
The gear and pinion were successfully cut according to these parameters. Tooth contact analysis (TCA) software, based on the mathematical model of the generated surfaces, predicted a well-centered and sized contact pattern for both flanks. Subsequent rolling tests of the physical gear pair on a testing machine under light load confirmed smooth and stable meshing. The observed contact patterns on the gear teeth aligned closely with the TCA predictions, exhibiting the desired localized bearing contact. The slight deviation of the contact on the gear concave flank towards the toe was attributed to minor machine alignment errors, not to the gear cutting principle itself. The experiment conclusively validated the feasibility and correctness of the monolithic cutter head gear cutting method for producing functional Klingelnberg-style cyclo-palloid bevel gears.
5. Conclusion
This work has established a complete and viable methodology for the gear cutting of Klingelnberg cyclo-palloid bevel gears using a monolithic cutter head on standard domestic CNC spiral bevel gear cutting machines. The core of the method is a revised finishing strategy for the pinion, employing single-sided cuts with calculated cutter radius modifications to introduce the necessary longitudinal flank curvature mismatch for localized contact. The design logic for the monolithic cutter head and its dedicated blades has been presented. Comprehensive formulas for calculating all essential machine adjustment settings, the crown gear curvature, the critical radius modification $E_{XB}$, and the blade shim thicknesses have been derived and systematized. The successful gear cutting experiment, yielding a gear pair with good meshing characteristics, provides practical proof of the principle. This development significantly lowers the barrier to entry for manufacturing high-quality cyclo-palloid gears, offering a practical alternative to reliance on specialized, expensive imported machinery for this important class of gear cutting.