In the realm of parallel-axis power transmission, cylindrical gears, particularly spur and helical variants, are ubiquitous due to their efficiency and reliability. However, traditional involute cylindrical gears exhibit inherent limitations, such as sensitivity to misalignment, edge contact, and sliding-induced wear, which can compromise performance and longevity. To address these challenges, we explore a novel design methodology for non-generated cylindrical gears based on pure rolling meshing principles, specifically focusing on curvilinear cylindrical gears with a circular arc tooth trace. This approach leverages active design of meshing line functions and combined tooth profiles to achieve optimized meshing characteristics, including reduced sliding, improved stress distribution, and enhanced transmission error behavior. In this article, we present a comprehensive study encompassing mathematical modeling, tooth contact analysis (TCA), stress evaluation via finite element analysis (FEA), and comparative assessments against other gear types, aiming to elucidate the advantages of this innovative cylindrical gear design.
The foundation of our design lies in the concept of pure rolling meshing, where tooth surfaces are generated to minimize relative sliding, thereby reducing friction, wear, and heat generation. For curvilinear cylindrical gears with a circular arc tooth trace, the tooth shape along the face width is controlled by predesigned contact curves derived from pure rolling meshing functions. These curves ensure that meshing occurs primarily through rolling motion, enhancing efficiency and durability. The tooth surface is constructed by sweeping a combined transverse tooth profile along these contact curves. The transverse profile comprises three segments smoothly connected at control points: a circular arc near the tooth tip, an involute curve in the active region, and a Hermite curve for the fillet transition. This combination allows for precise control over geometry, avoiding issues like undercutting and tip sharpening common in generated gears. The design freedom offered by this non-generated approach, coupled with advancements in additive manufacturing, opens new avenues for producing high-performance cylindrical gears with tailored materials and geometries.
To mathematically define the cylindrical gear, we establish coordinate systems as shown in the referenced figures. Fixed coordinate systems \(S_p\) and \(S_g\) are attached to the frame, while \(S_1\) and \(S_2\) are attached to the pinion and gear, respectively. The pure rolling meshing point motion is described in an auxiliary fixed system \(S_k\). The meshing line \(K-K\) and the contact curves \(C_1\) and \(C_2\) on the pinion and gear are derived from the motion of meshing points \(M_a\) and \(M_b\), which symmetric about the \(x_k\)-axis. For a circular arc tooth trace, the parameterized equation of the meshing line expansion is given by:
$$
\begin{align*}
x_k &= 0, \\
y_k &= 0, \\
z_k(t) &= \pm \sqrt{r_c^2 – (k_\phi t R_1 – r_c)^2},
\end{align*}
$$
where \(t\) is the motion parameter (\(0 \leq t \leq t_{\text{max}}\)), \(k_\phi\) is the motion coefficient, \(r_c\) is the radius of the circular arc obtained by unfolding the contact curve, and \(R_1\) is the pitch radius of the pinion. The relationship between gear rotation and motion parameter is:
$$
\phi_1 = k_\phi t, \quad \phi_2 = \phi_1 / i_{12},
$$
with \(i_{12}\) as the transmission ratio. The tooth surfaces of the pinion and gear are formed by sweeping transverse profiles along \(z_k(t)\). The position vector for any point on the pinion tooth surface is:
$$
\mathbf{r}_1^{(\Omega_j^q)} =
\begin{bmatrix}
x_p^{(\Sigma_j^q)} \cos \phi_1 – y_p^{(\Sigma_j^q)} \sin \phi_1 \\
x_p^{(\Sigma_j^q)} \sin \phi_1 + y_p^{(\Sigma_j^q)} \cos \phi_1 \\
z_k(t) \\
1
\end{bmatrix},
$$
where \(q\) denotes left or right flank, \(j\) represents the segment (circular arc, involute, or Hermite), and \(x_p^{(\Sigma_j^q)}, y_p^{(\Sigma_j^q)}\) are coordinates of the transverse profile in system \(S_p\). Similarly, the gear tooth surface vector is:
$$
\mathbf{r}_2^{(\Omega_J^Q)} =
\begin{bmatrix}
-x_g^{(\Sigma_J^Q)} \cos \phi_2 + y_g^{(\Sigma_J^Q)} \sin \phi_2 \\
x_g^{(\Sigma_J^Q)} \sin \phi_2 + y_g^{(\Sigma_J^Q)} \cos \phi_2 \\
z_k(t) \\
1
\end{bmatrix},
$$
with \(Q\) and \(J\) as corresponding indices for the gear. This parametric formulation enables precise generation of the cylindrical gear surfaces, ensuring pure rolling conditions are maintained throughout meshing.
The transverse tooth profile is critical for defining the cylindrical gear geometry. It is constructed using four control points: \(P_{ai}\) at the tip, \(P_i\) at the pitch point, \(P_{di}\) at the start of the fillet, and \(P_{ei}\) at the root. The profile consists of three curves: a circular arc \(\Sigma_{\text{Cir}}^l\), an involute \(\Sigma_{\text{Inv}}^l\), and a Hermite curve \(\Sigma_{\text{Her}}^l\), smoothly connected at \(P_i\) and \(P_{di}\). In a local coordinate system \(S_{pi}\) attached at \(P_i\), the circular arc segment is parameterized as:
$$
\begin{align*}
x_{pi}^{(\Sigma_{\text{Cir}}^l)} &= \rho_{pi} \sin \xi_{pi}, \\
y_{pi}^{(\Sigma_{\text{Cir}}^l)} &= \rho_{pi} \cos \xi_{pi} – \rho_{pi}, \\
z_{pi}^{(\Sigma_{\text{Cir}}^l)} &= 0,
\end{align*}
$$
for \(\xi_{pi}^{\text{min}} \leq \xi_{pi} \leq \xi_{pi}^{\text{max}}\), where \(\rho_{pi}\) is the arc radius. The involute segment in system \(S_{\text{Inv}}\) is:
$$
\begin{align*}
x_{\text{Inv}}^{(\Sigma_{\text{Inv}}^l)} &= r_{bi} \sin u_i – u_i r_{bi} \cos u_i, \\
y_{\text{Inv}}^{(\Sigma_{\text{Inv}}^l)} &= r_{bi} \cos u_i + u_i r_{bi} \sin u_i, \\
z_{\text{Inv}}^{(\Sigma_{\text{Inv}}^l)} &= 0,
\end{align*}
$$
for \(u_{di} \leq u_i \leq u_{pi}\), with \(r_{bi}\) as base radius and \(u_i\) as the involute parameter. The Hermite curve for the fillet is defined by control points \(P_{di}\) and \(P_{ei}\), with parameter weight \(T_H\) influencing curvature. This combined profile ensures continuous stress distribution and avoids geometric discontinuities, which is essential for high-load applications in cylindrical gear drives.
To validate the design, we establish basic parameters for four gear sets: Case 1 (pure rolling cylindrical gears with circular arc tooth trace), Case 2 (pure rolling cylindrical gears with parabolic tooth trace), and Cases 3-4 (modified traditional helical gears). The parameters are summarized in the table below, highlighting key dimensions common to all cylindrical gear sets.
| Parameter | Symbol | Value for All Cases |
|---|---|---|
| Number of pinion teeth | \(Z_1\) | 30 |
| Transmission ratio | \(i_{12}\) | 2.0 |
| Normal module | \(m_n\) | 2.0 mm |
| Normal pressure angle | \(\alpha_n\) | 20° |
| Helix angle | \(\beta\) | 22.1474° |
| Addendum coefficient | \(h_{an}^*\) | 1.0 |
| Dedendum coefficient | \(c_n^*\) | 0.25 |
| Face width | \(b\) | 50 mm |
For Cases 1 and 2, additional transverse profile parameters are defined to tailor the cylindrical gear geometry. These include the motion coefficient \(k_\phi = \pi\), maximum parameter \(t_{\text{max}} = 0.1\), control point position coefficient \(k_d = 0.75\), rotation coefficients \(k_{\chi_{a1}} = 0.11\) and \(k_{\chi_{a2}} = 0.04\), Hermite weights \(T_{Hp} = 0.5\) and \(T_{Hg} = 0.7\), and coefficients for angles \(k_\eta = 0.02\) and \(k_\lambda = 0.02\). The geometric dimensions are derived using standard cylindrical gear relations. The transverse module and pressure angle are:
$$
m_t = \frac{m_n}{\cos \beta}, \quad \alpha_t = \arctan\left(\frac{\tan \alpha_n}{\cos \beta}\right).
$$
Pitch radii are \(R_i = Z_i m_t\), with center distance \(a = R_1 + R_2\). Addendum and dedendum heights are \(h_{ai} = h_{an}^* m_n\) and \(h_{fi} = (h_{an}^* + c_n^*) m_n\), respectively. Key radii for the cylindrical gear include addendum radius \(R_{ai} = R_i + h_{ai}\), control point radius \(R_{di} = R_i – h_d\) with \(h_d = k_d m_n\), and root radius \(R_{fi} = R_i – h_{fi}\). The transverse tooth thickness is controlled by angles \(\lambda_1 = (1 + k_\lambda)\pi/Z_1\) and \(\lambda_2 = (1 – k_\lambda)\pi/Z_2\), while root angles are \(\eta_1 = 2\pi k_\eta / Z_1\) and \(\eta_2 = 2\pi k_\eta / Z_2\). Positions of Hermite curve points are given by angles \(\kappa_{l1}, \kappa_{l2}, \kappa_{r1}, \kappa_{r2}\), calculated as:
$$
\begin{align*}
\kappa_{l1} &= \left(0.75 – 0.5k_\eta\right)\frac{2\pi}{Z_1} + 0.5k_\lambda \frac{\pi}{Z_1}, \\
\kappa_{l2} &= \left(0.75 – 0.5k_\eta\right)\frac{2\pi}{Z_2} – 0.5k_\lambda \frac{\pi}{Z_2}, \\
\kappa_{r1} &= \left(0.25 – 0.5k_\eta\right)\frac{2\pi}{Z_1} – 0.5k_\lambda \frac{\pi}{Z_1}, \\
\kappa_{r2} &= \left(0.25 – 0.5k_\eta\right)\frac{2\pi}{Z_2} + 0.5k_\lambda \frac{\pi}{Z_2}.
\end{align*}
$$
These parameters ensure the cylindrical gear teeth are optimally shaped for pure rolling meshing. Micro-geometry modifications are applied for comparison: Cases 1 and 2 have parabolic lead crowning of 2 µm, while Cases 3 and 4 have profile and lead modifications of 10 µm and 60 µm, respectively, to match contact pattern heights.
To analyze meshing performance, we conduct tooth contact analysis (TCA) for all cylindrical gear sets under perfect alignment conditions. The TCA algorithm solves for contact points and transmission errors by simulating meshing over two pitch angles, discretized into 21 positions. The contact patterns on pinion and gear surfaces reveal distinct characteristics. For Cases 1 and 2, the patterns are symmetric about the mid-plane, with instantaneous contact ellipses centered on the pitch radius. The ellipse major axis length increases from the ends toward the center of the face width, and its orientation relative to the tooth trace changes gradually. This symmetry aids in eliminating axial forces, a key advantage of curvilinear cylindrical gears. In contrast, Cases 3 and 4 show contact ellipses with consistent orientation and size along the path of contact, due to the applied modifications. Despite different modification types, their patterns are similar, covering comparable profile heights. The unloaded transmission error curves for all cylindrical gear sets exhibit parabolic shapes, attributable to the lead crowning in Cases 1-2 and combined modifications in Cases 3-4. This parabolic error function helps mitigate noise and vibration by avoiding discontinuous linear errors under misalignment.
We further evaluate mechanical behavior through stress analysis using finite element methods. For cylindrical gears in Cases 1 and 2, a model with five tooth pairs is built to account for load sharing; for Cases 3 and 4, seven pairs are used due to higher contact ratio. The pinion is subjected to a torque of 300 N·m, and materials are steel with elastic modulus 210 GPa and Poisson’s ratio 0.3. Mesh convergence is achieved with 65 elements along the face width and 35 along the profile. Constraints are applied to simulate rigid hub connections. The maximum von Mises stress on the pinion surface and maximum principal stress at the fillet are monitored over 21 contact positions. The results indicate that Cases 1 and 2 have similar von Mises stress fluctuations and peaks, slightly higher than those in Cases 3 and 4. Specifically, the maximum von Mises stress in Case 1 is about 3-5% higher than in Cases 3-4. The stress distributions align with TCA contact patterns, showing multiple contact ellipses for Cases 1-2 and uniform ellipses for Cases 3-4. For bending stresses, Cases 1 and 2 demonstrate lower maximum fillet stresses compared to Cases 3 and 4, with Case 2 having the lowest. Notably, in these cylindrical gears, the pinion experiences lower bending stress than the gear, which is beneficial for longevity given the pinion’s higher stress cycles. In contrast, traditional helical gears show higher pinion bending stresses. The table below summarizes key stress outcomes for the cylindrical gear sets.
| Case | Max Pinion von Mises Stress (MPa) | Max Pinion Bending Stress (MPa) | Max Gear Bending Stress (MPa) |
|---|---|---|---|
| Case 1 | ~1050 | ~320 | ~340 |
| Case 2 | ~1020 | ~312 | ~332 |
| Case 3 | ~995 | ~391 | ~374 |
| Case 4 | ~993 | ~390 | ~373 |
Loaded transmission error (LTE) functions are computed by incorporating contact deformations and tooth deflections from FEA results. The LTE curves for all cylindrical gear sets maintain parabolic shapes, with Cases 1 and 2 exhibiting significantly larger amplitudes (approximately 29″ and 27″, respectively) compared to Cases 3 and 4 (about 1.5″). This larger amplitude in curvilinear cylindrical gears may be attributed to the pure rolling design and lead modifications, which enhance misalignment tolerance but increase error magnitude. However, the parabolic nature ensures smooth meshing transitions, reducing dynamic loads.
In conclusion, the pure rolling curvilinear cylindrical gear with circular arc tooth trace offers distinct advantages over traditional cylindrical gears. The design methodology, based on active meshing line functions and combined tooth profiles, enables precise control over geometry and meshing behavior. Through comprehensive analysis, we find that this cylindrical gear exhibits similar contact patterns, stress levels, and transmission error characteristics to its parabolic tooth trace counterpart, while outperforming modified helical gears in terms of lower bending stresses and higher LTE amplitudes. The symmetric contact patterns eliminate axial forces, and the reduced sliding minimizes wear, contributing to enhanced durability. These findings underscore the potential of non-generated cylindrical gears for high-performance applications, especially where misalignment compensation and efficiency are critical. Future work could explore optimization of parameters for specific loads or additive manufacturing constraints, further advancing the capabilities of cylindrical gear technology.