In the field of power transmission, miter gears, particularly large modulus straight bevel gears, are critical components for transmitting motion and torque between intersecting shafts. Their applications in heavy machinery, mining equipment, and marine propulsion demand high load-bearing capacity, reliability, smooth operation, and long service life. However, manufacturing these gears poses significant challenges due to the complexity of their spherical involute tooth profiles and the scarcity of specialized equipment. Traditional methods often rely on dedicated bevel gear generators, which are expensive and not always accessible. This paper addresses these issues by proposing an alternative machining approach using a finger-type milling cutter on a standard vertical milling machine. We delve into the principles of form machining for miter gears, detailing parameter calculations, tool path generation, and comprehensive error analysis. Furthermore, we introduce a modification strategy based on tooth profile error distribution to enhance meshing performance. The entire process, from machining to modification, is simulated in VERICUT to validate feasibility, followed by physical machining experiments and metrological verification. Our goal is to provide a practical, cost-effective solution for producing high-quality large modulus miter gears in settings with limited resources.
The core idea of machining miter gears with a finger-type cutter lies in approximating the spherical involute profile with a planar involute on the back-cone. The cutter profile is designed based on the tooth shape at the midpoint of the face width. By dynamically adjusting the depth of cut along the gear blank, we can machine the tooth slot from the heel (large end) to the toe (small end). While using a single cutter inevitably introduces tooth profile errors, these errors are symmetrically distributed and can be complementary during meshing. With appropriate modification, the final gear can meet rigorous operational requirements. This method is especially advantageous for large modulus miter gears where module sizes often exceed 20 mm, making traditional hobbling or shaping less feasible.
To illustrate the process, let us consider a specific pair of miter gears. The primary parameters are summarized in Table 1. For machining, the pinion (smaller gear) is typically more susceptible to profile errors, making it the focus of our analysis. The milling cutter is selected as an involute module cutter with a module corresponding to the midpoint of the face width.
| Parameter | Symbol | Value | Unit |
|---|---|---|---|
| Number of teeth (Gear) | $Z_1$ | 65 | – |
| Number of teeth (Pinion) | $Z_2$ | 22 | – |
| Module at Heel (Large end) | $M_E$ | 30 | mm |
| Module at Toe (Small end) | $M_I$ | 22.131 | mm |
| Face Width | $B$ | 270 | mm |
| Reference Pressure Angle | $\alpha$ | 20 | ° |
| Cutter Module (Midpoint) | $M_M$ | 26.065 | mm |
| Pitch Cone Angle (Pinion) | $\delta_2$ | Calculated | ° |
The fundamental requirement for proper meshing of miter gears is the consistency of the chordal tooth thickness or space width at the reference pitch circle across different cone distances. During machining, the cutter axis is kept perpendicular to the pitch cone element of the gear blank. The actual tooth profile on the back-cone and the slot depth are controlled solely by the cutter profile and its radial infeed (depth of cut), denoted as $K$. There is no swing of the cutter axis; only linear movement along its axis relative to the workpiece is used. To ensure constant space width at the reference pitch circle of the equivalent spur gear at any cone distance $R$, we derive the necessary cutter offset $K$.
For an involute module cutter, the tooth thickness at an arbitrary radius $r_i$ on the equivalent gear is given by:
$$ S_i = r_i \left[ \frac{S}{r} – 2(\text{inv} \alpha_i – \text{inv} \alpha) \right] $$
where:
– $r$ is the reference pitch radius of the equivalent gear at the cone distance in question.
– $S$ is the tooth thickness at the reference pitch circle, typically $S = \pi m / 2$ for standard gears.
– $\alpha_i$ is the pressure angle at radius $r_i$, calculated as $\alpha_i = \arccos(r_b / r_i)$.
– $r_b$ is the base circle radius, $r_b = r \cos \alpha$.
– $\text{inv} \alpha = \tan \alpha – \alpha$ is the involute function.
The space width at the reference pitch circle is $e = \pi m / 2$. The space width of the cutter at an arbitrary radius $r_i$ is:
$$ e_i = p_i – S_i $$
where $p_i = 2\pi r_i / Z$ is the circular pitch at radius $r_i$, and $Z$ is the number of teeth of the equivalent gear ($Z = Z_2 / \cos \delta_2$ for the pinion).
To match the reference space width, we set $e_i = e$. Substituting the expressions and solving for $r_i$ yields the radial position where the cutter profile matches the required space width. The depth of cut variation $K$ is then:
$$ K = r_i – r $$
Since the cutter profile is based on the midpoint of the face width, $K=0$ at that location. For positions towards the toe (small end), $K$ becomes negative (cutter retracts); towards the heel (large end), $K$ becomes positive (cutter advances). By calculating $r_i$ and thus $K$ at multiple points along the face width, we can generate a continuous tool path. Table 2 shows computed values for eight segments across the face width for the pinion example.
| Position from Toe (mm) | Cone Distance $R$ (mm) | Equivalent Gear Ref. Radius $r$ (mm) | Calculated $r_i$ (mm) | Depth Variation $K$ (mm) |
|---|---|---|---|---|
| 0 | $R_I$ | 243.44 | 237.12 | -6.32 |
| 33.75 | $R_I + B/8$ | 256.89 | 252.11 | -4.78 |
| 67.5 | $R_I + B/4$ | 270.34 | 267.10 | -3.24 |
| 101.25 | $R_I + 3B/8$ | 283.79 | 282.09 | -1.70 |
| 135 | $R_I + B/2$ (Midpoint) | 297.24 | 297.24 | 0.00 |
| 168.75 | $R_I + 5B/8$ | 310.69 | 312.39 | 1.70 |
| 202.5 | $R_I + 3B/4$ | 324.14 | 327.54 | 3.40 |
| 236.25 | $R_I + 7B/8$ | 337.59 | 342.69 | 5.10 |
| 270 | $R_E$ (Heel) | 351.04 | 357.84 | 6.80 |
Fitting a polynomial curve through these $(B, K)$ points provides a continuous function for the CNC program. A quadratic or cubic fit is often sufficient. The resulting tool path ensures that the space width at the reference pitch circle remains constant, satisfying the fundamental condition for correct gear mating and constant velocity ratio in miter gears.
However, using a single finger-type cutter for the entire miter gear tooth slot leads to inherent tooth profile deviations. The error distribution is symmetrical about the face width midpoint. Figure 3 (referenced conceptually from the source) illustrates that excess material (positive error) concentrates near the toe tip and heel root, while undercut (negative error) occurs at the toe root and heel tip. Fortunately, in a mating pair of miter gears, these errors can be complementary, reducing their net impact on transmission. Nevertheless, for high-performance applications, modification is necessary to optimize contact patterns and reduce stress concentration.
To quantify these errors, we perform a geometric analysis. For any machining cross-section (a specific cone distance $R$), we establish a coordinate system $O-XY$ with its origin at the center of the equivalent spur gear and the Y-axis aligned with the cutter axis. Let $(X_i, Y_i)$ be a point on the theoretical tooth profile (the planar involute on the back-cone). The equation of the theoretical profile is:
$$ \begin{cases} X = r_i \sin(T + \theta_i) \\ Y = r_i \cos(T + \theta_i) \end{cases} $$
where $T$ is the starting angle of the involute, and $\theta_i = \text{inv} \alpha_i$.
The actual cutter profile in the same coordinate system is given by:
$$ \begin{cases} X = r_o \sin(T’ + \theta_o) \\ Y = r_o \cos(T’ + \theta_o) \end{cases} $$
where $r_o = r_i + t + K$. Here, $t = e \cdot \tan \phi$ accounts for the lateral shift due to the conical shape ($\phi$ is the base cone angle, $e$ is the distance from the face width midpoint to the current section). $K$ is the depth variation from Table 2. $T’$ is the cutter profile starting angle, and $\theta_o = \text{inv} \alpha_o$ with $\cos \alpha_o = r_b’ / r_o$ ($r_b’$ is the base circle radius of the cutter).
To find the deviation in the direction of the meshing line (which is the critical error affecting transmission), we take a point $(x_0, y_0)$ on the theoretical profile and construct the line normal to the profile at that point. The normal line equation is:
$$ \begin{cases} X = t \\ Y = y_1 – \frac{y_1 – y_0}{x_0} t \end{cases} $$
where $y_1 = \frac{r_b}{\cos(T + \tan \alpha)}$ and $\cos \alpha = r_b / \sqrt{x_0^2 + y_0^2}$.
Intersecting this normal line with the cutter profile equation yields the point $(x’, y’)$ on the machined surface along the same line of action. The profile error in the meshing direction, $E$, is the distance between these points:
$$ E = \sqrt{(x’ – x_0)^2 + (y’ – y_0)^2} $$
The sign of the error (excess or deficit) is determined by comparing the positions of $(x_0, y_0)$ and $(x’, y’)$. Calculating $E$ for multiple points from tip to root across several cross-sections provides a comprehensive error map. For our example miter gear pinion, the maximum errors are found at the extremities. The error at the toe tip can be significant, often exceeding 2 mm for very large modulus gears, which necessitates modification.
The principle of miter gear modification aims to relocate the contact pattern towards the center of the tooth face, eliminating edge contact and the associated high stress concentration. Based on the error analysis, the toe region (especially the tip) has excess material that impinges on mating, while the heel tip has a deficit that may be tolerated due to complementary action. Therefore, our modification strategy focuses primarily on relieving the toe portion of the tooth, from the tip down to the pitch line, tapering off towards the face width midpoint.
For the modification cut, we employ a standard module milling cutter, but its effective profile is now the involute segment from the tip to the pitch circle of the toe cross-section’s equivalent gear. The goal is to calculate a new set of depth variations $K’$ for this finishing pass. We require that the space width at the tip circle of the modified section matches the theoretical space width at that circle for the correct toe tooth thickness. This ensures material is removed only from the tip region without affecting the correct profile near the root.
The space width at an arbitrary radius $r_i$ on the cutter (for the equivalent gear at cone distance $R$) is:
$$ e = \frac{2\pi r_i}{Z} – r_i \left[ \frac{S}{r} – 2(\text{inv} \alpha_i – \text{inv} \alpha) \right] $$
At the tip circle of the section, with radius $r_i’$, the desired space width is:
$$ e’ = \frac{2\pi r_i’}{Z} – r_i’ \left[ \frac{S’}{r’} – 2(\text{inv} \alpha_i’ – \text{inv} \alpha’) \right] $$
Here, $S’$ and $r’$ are the tooth thickness and reference radius at the tip circle’s equivalent gear. Setting $e = e’$ allows us to solve for the required $r_i$, and subsequently $K’ = r – r_i$, where $r$ is the reference radius at that section. Performing this calculation for sections from the toe to the midpoint yields the modification tool path. A sample of calculated $K’$ values is presented in Table 3.
| Position from Toe (mm) | Cone Distance $R$ (mm) | Ref. Radius $r$ (mm) | Calculated $r_i$ for Mod (mm) | Modification Depth $K’$ (mm) |
|---|---|---|---|---|
| 0 | $R_I$ | 243.44 | 243.44 | 0.00 |
| 33.75 | $R_I + B/8$ | 256.89 | 255.91 | -0.98 |
| 67.5 | $R_I + B/4$ | 270.34 | 268.38 | -1.96 |
| 101.25 | $R_I + 3B/8$ | 283.79 | 280.85 | -2.94 |
| 135 | $R_I + B/2$ | 297.24 | 293.32 | -3.92 |
This modification pass, executed after the roughing and semi-finishing passes with the main finger cutter, effectively removes the excess material at the toe tip and blends the profile towards the center. It is crucial that the tip relief is smooth and continuous to avoid introducing new stress raisers. The modification significantly improves the load distribution across the tooth face of the miter gear.
To validate the entire machining and modification process before physical trials, we employ VERICUT simulation software. A virtual model of a 3-axis vertical milling machine with a rotary table (simulating the C-axis) is constructed. The gear blank and the custom finger-type cutter are modeled. The cutter profile is defined by importing a set of $(X, Y)$ coordinates calculated from the involute equations for the midpoint module $M_M=26.065$ mm. The CNC program, containing the tool paths for both the main cutting and modification passes, is loaded into the simulation project. The simulation runs, showing material removal in real-time and allowing for collision detection and verification of tool motions.
After the virtual machining, VERICUT’s built-in comparison tool is used to measure the deviation between the simulated gear and a perfect CAD model of the theoretical miter gear. The “Auto-Diff” module generates color-coded maps showing excess and deficient material. Prior to modification, the error map clearly shows high positive residuals (red) at the toe tip and heel root, and negative residuals (blue) at the toe root and heel tip, confirming our analytical error distribution. After applying the modification pass, the error map for the toe region shows a dramatic reduction in positive residuals. The maximum residual in the modified zone is well under 0.1 mm, indicating successful relief of the toe tip. This simulation step confirms the correctness of the tool path calculations and the effectiveness of the proposed modification strategy for miter gears.
Following the successful simulation, we proceed to physical machining on a CNC vertical milling machine equipped with a fourth-axis rotary table. The workpiece material is AISI 4140 alloy steel, a common choice for high-strength gears. The same finger-type milling cutter, fabricated from solid carbide, is used. The CNC program derived from our calculated tool paths is executed. After the main machining operation, the modification pass is performed using a standard off-the-shelf module milling cutter with the appropriate profile. The entire process is completed on the same setup, ensuring accuracy and repeatability.
To quantitatively assess the tooth profile accuracy, a precision optical coordinate measuring machine (CMM) is employed. For both the unmodified and modified gears, several teeth are measured. The CMM data is aligned with the theoretical CAD model, and deviations along the profile (in the direction of the meshing line) are computed at a grid of points from tip to root and from toe to heel. The results are summarized in Table 4 and Table 5. The data unequivocally demonstrates the improvement brought by modification. The maximum error in the meshing direction before modification was 2.158 mm at the toe tip, which is unacceptable for smooth operation. After modification, the maximum error in the critical toe region is reduced to a mere 0.041 mm, which is within acceptable limits for many heavy-duty applications involving large modulus miter gears.
| Height from Tip (Fraction of Addendum) | Position Along Face Width from Toe (mm) | ||||
|---|---|---|---|---|---|
| 0 (Toe) | B/8 (33.75) | B/4 (67.5) | 3B/8 (101.25) | B/2 (135) | |
| 0 (Tip) | 2.158 | 1.588 | 1.051 | 0.523 | 0.003 |
| A/4 | 1.669 | 1.211 | 0.764 | 0.410 | 0.001 |
| A/2 | 1.117 | 0.804 | 0.520 | 0.246 | 0.004 |
| 3A/4 | 0.563 | 0.397 | 0.233 | 0.107 | 0.005 |
| A (Pitch Line) | 0.003 | 0.009 | 0.007 | 0.002 | 0.004 |
Note: A denotes the addendum height from tip to pitch line.
| Height from Tip (Fraction of Addendum) | Position Along Face Width from Toe (mm) | ||||
|---|---|---|---|---|---|
| 0 (Toe) | B/8 (33.75) | B/4 (67.5) | 3B/8 (101.25) | B/2 (135) | |
| 0 (Tip) | 0.002 | 0.003 | 0.002 | 0.004 | 0.001 |
| A/4 | 0.005 | 0.011 | 0.015 | 0.013 | 0.001 |
| A/2 | 0.001 | 0.009 | 0.018 | 0.022 | 0.003 |
| 3A/4 | 0.003 | 0.010 | 0.021 | 0.041 | 0.004 |
| A (Pitch Line) | 0.001 | 0.008 | 0.006 | 0.002 | 0.003 |
Finally, a meshing test is conducted using a gear rolling tester or a dedicated test rig. The machined pinion is mated with its gear counterpart, and a light load is applied. Contact patterns are recorded using precision marking compounds (e.g., Prussian blue). Before modification, the contact pattern is heavily biased towards the toe and heel edges, indicating severe edge contact and stress concentration. After modification, the contact pattern shifts markedly towards the center of the tooth face, becoming a well-defined oval shape located near the midpoint of the face width and slightly below the pitch line. This optimal contact pattern ensures even load distribution, minimizes bending and contact stresses, reduces the risk of pitting and tooth breakage, and contributes to quieter operation and longer service life for the miter gear pair.
In conclusion, this paper has presented a comprehensive methodology for machining and modifying large modulus miter gears using standard CNC milling equipment. The finger-type milling cutter approach, coupled with precise tool path calculation based on equivalent gear geometry, offers a viable alternative to specialized bevel gear machinery. The inherent tooth profile errors of this method are systematically analyzed using geometric models, and a targeted modification strategy using a secondary module cutter is developed to correct the most detrimental errors at the toe. VERICUT simulations provide a robust virtual validation platform, minimizing trial-and-error in physical machining. Experimental results confirm the theory: modification drastically reduces profile errors and centralizes the contact pattern. This integrated process—from calculation to simulation to physical verification—enables the production of high-quality, large modulus miter gears for demanding applications, even in facilities lacking dedicated gear manufacturing equipment. Future work could explore the optimization of cutter profiles, dynamic analysis of the modified miter gears under load, and extension of the method to spiral bevel or hypoid gears.