The pursuit of high precision and reliability in robotic joints and precision machinery has placed significant emphasis on the performance of core transmission components. Among these, the RV reducer stands out for its compact structure, high reduction ratio, and excellent torsional rigidity. The performance of an RV reducer is a complex, nonlinear, and strongly coupled system, where transmission accuracy is paramount. A critical yet often less-explored factor influencing this accuracy is the behavior of the main bearings. These bearings directly support the output mechanism, and their stiffness characteristics, which are inherently nonlinear, significantly affect the system’s dynamic response, load distribution, and ultimately, positional fidelity. This study focuses on investigating the contact stiffness of the main bearings within an RV reducer, employing a combined theoretical and numerical approach to unravel its nonlinear nature and influencing factors.
The main bearing in an RV reducer is typically a customized angular contact ball bearing. Its outer ring is mounted with a prescribed interference fit within the reducer’s housing, while the inner ring is seated on the output planet carrier. This configuration means the bearing directly withstands the complex load spectrum from the two-stage reduction process—the primary planetary stage and the secondary cycloid-pinwheel stage—and transmits the final output torque. Any elastic deformation within this bearing introduces compliance into the system, degrading positional accuracy. While extensive research exists on the meshing stiffness of the cycloid drive and the torsional stiffness of the overall RV reducer transmission, detailed studies on the nonlinear contact stiffness of the main bearing, considering actual load conditions and geometric parameters, are less common. This work addresses this gap by developing a theoretical model for the main bearing’s contact stiffness and validating it through finite element analysis.
Theoretical Modeling of Main Bearing Load and Stiffness
Static Load Analysis of the RV Reducer Main Bearing
To determine the contact forces on the main bearing balls, one must first resolve the static indeterminate force system within the RV reducer. The main bearing is subjected to radial forces ($F_{r1}$, $F_{r2}$) originating from the crankshaft, which itself is loaded by forces from the planetary gears ($F_r$) and the cycloid gears ($F_A$, $F_B$). A simplified force diagram of the crankshaft and main bearing system is considered. Based on force and moment equilibrium equations for the crankshaft, the radial forces on the bearing can be calculated.
Considering the geometry where distances A, B, C, and D represent specific spacings on the crankshaft, the equilibrium equations are:
$$F_{r1} \times (A+B+C) + F_A \times (B+C) – F_B \times C – F_r \times D = 0$$
$$F_{r2} \times (A+B+C) + F_A \times A – F_B \times (A+B) + F_r \times (A+B+C+D) = 0$$
The forces $F_A$, $F_B$, and $F_r$ are derived from the transmitted torque and the geometry of the planetary and cycloid stages. For a given input power ($P$) and speed ($n_{input}$), the output torque can be calculated from the total reduction ratio ($i$). The forces on the cycloid disc and pins can then be determined using established formulas for cycloid drives. Once $F_{r1}$ is known, the normal contact force ($Q$) on a single ball in the main bearing, assuming a quasi-static load distribution, is approximated by:
$$Q = \frac{F_{r1}}{Z \cos\alpha}$$
where $Z$ is the number of balls and $\alpha$ is the initial contact angle. Due to the kinematics of the RV reducer, where one revolution of the crankshaft corresponds to a small rotation of the cycloid disc (one tooth), the radial load $F_{r1}$ varies periodically. Consequently, the ball contact force $Q$ also exhibits a periodic variation with the crankshaft angle, superimposed on a slight ascending trend over the cycloid disc’s rotation cycle. The basic parameters used for an example E-series RV reducer and its main bearing are summarized below.
| Parameter | Symbol | Unit | Value |
|---|---|---|---|
| RV Reducer Transmission Ratio | $i$ | – | 153 |
| Input Speed | $n_{input}$ | rpm | 1050 |
| Input Power | $P$ | kW | 0.65 |
| Number of Balls (Main Bearing) | $Z$ | – | 40 |
| Ball Diameter | $D_w$ | mm | 10 |
| Pitch Circle Diameter | $d_m$ | mm | 132 |
| Initial Contact Angle | $\alpha$ | ° | 10 |
Hertz Contact Theory and Stiffness Derivation
The contact stiffness between the ball and the raceway is derived from Hertzian elastic contact theory. The theory assumes homogeneous, isotropic materials, small contact area compared to the component dimensions, smooth surfaces, and perfectly elastic deformation obeying Hooke’s law. For two elastic bodies (ball and raceway) in point contact, the geometry is defined by principal curvatures.
For the ball (Body I):
$$\rho_{1}^{I} = \rho_{1}^{II} = \frac{2}{D_w}$$
(Note: $\rho = 1/r$, where $r$ is radius)
For the inner or outer raceway (Body II, using inner race as example):
$$\rho_{2}^{I} = -\frac{2}{f D_w}, \quad \rho_{2}^{II} = \frac{2}{d_i}$$
The curvature $\rho_{2}^{I}$ is negative as the raceway groove is concave. Here, $f$ is the curvature radius coefficient (raceway groove radius / ball radius), and $d_i$ is the inner raceway diameter at the contact point. The sum and difference of the principal curvatures are critical:
$$\Sigma\rho = \rho_{1}^{I} + \rho_{1}^{II} + \rho_{2}^{I} + \rho_{2}^{II}$$
$$F(\rho) = \frac{(\rho_{1}^{II} – \rho_{1}^{I}) + (\rho_{2}^{II} – \rho_{2}^{I})}{\Sigma\rho}$$
The normal elastic approach (mutual approach of the two bodies) $\delta$ due to load $Q$ is given by a simplified formula for steel-on-steel contact:
$$\delta = \delta^* (\Sigma\rho)^{-1/3} Q^{2/3}$$
where $\delta^*$ is a dimensionless deflection coefficient dependent on the curvature difference function $F(\rho)$. This coefficient is obtained from standard Hertzian contact tables linking $F(\rho)$ to $\delta^*$. The contact stiffness $K$ for a ball-raceway contact is defined as the rate of change of load with respect to this elastic deformation. From the relationship $Q \propto \delta^{3/2}$, the instantaneous contact stiffness can be expressed as:
$$K = \frac{dQ}{d\delta} \propto Q^{1/3} (\Sigma\rho)^{1/2}$$
A more specific operational formula for the stiffness of an angular contact ball bearing under radial load considers the load-deflection constant. For a single ball contact, the stiffness can be evaluated as:
$$K_{ball} = 1.5 \times C_b \times \delta^{1/2}$$
where $C_b$ is the load-deflection constant from Hertz theory, $C_b = \left( \frac{1}{\delta^*} \right)^{3/2} \left( \frac{1}{\Sigma\rho} \right)^{1/2} \frac{1}{K_{H}}$, and $K_{H}$ incorporates material properties (Elastic modulus $E$ and Poisson’s ratio $\nu$). The total axial or radial stiffness of the bearing is the summation of the contributions from all loaded balls, which varies with the rotational angle due to the changing load distribution. Calculating $\delta$ and subsequently $K$ over one revolution of the crankshaft reveals the periodic and nonlinear nature of the main bearing contact stiffness in the RV reducer.
Influence of Curvature Radius Coefficient on Stiffness
The curvature radius coefficient $f$ is a pivotal design parameter for the main bearing in an RV reducer. It directly influences the contact ellipse dimensions, the maximum contact stress, and consequently, the contact stiffness. By varying $f$, the principal curvatures $\rho_{2}^{I}$ change, altering $\Sigma\rho$ and $F(\rho)$, which in turn affect the deflection coefficient $\delta^*$ and the final stiffness calculation.
Analysis shows that the contact stiffness of the RV reducer main bearing is not monotonic with respect to $f$. There exists an optimal value that maximizes stiffness for a given load condition. For the specific bearing geometry studied, comparing stiffness values over a crankshaft rotation cycle for different $f$ coefficients yields the following insight:
| Curvature Radius Coefficient (f) | Approx. Max Contact Stiffness (N/μm) | Trend |
|---|---|---|
| 0.525 | 3.29 × 103 | Lower than optimal |
| 0.550 | 3.31 × 103 | Maximum (Optimal) |
| 0.575 | 3.16 × 10103 | Lower than optimal |
The results indicate that an $f$ value of 0.55 provides the highest contact stiffness for this RV reducer main bearing configuration under the defined loading. This optimal value represents a balance, likely minimizing the contact ellipse size and thus the elastic deformation for a given load, leading to higher stiffness. Straying from this value in either direction reduces stiffness, which could negatively impact the dynamic performance and positioning accuracy of the RV reducer.
Finite Element Analysis for Validation
Theoretical Basis for Nonlinear FEA
Validating the theoretical stiffness model requires a numerical approach capable of handling the nonlinearities inherent in contact problems. The Finite Element Method (FEM) is employed for this purpose. The underlying theory for this structural analysis involves solving the governing equations for a deformable body: equilibrium, compatibility (strain-displacement), and constitutive (stress-strain) equations.
In a general 3D continuum, these are expressed in index notation as:
Equilibrium: $\sigma_{ij,j} + b_i = 0$
Compatibility: $\varepsilon_{ij} = \frac{1}{2}(u_{i,j} + u_{j,i})$
Constitutive (Linear Elastic): $\sigma_{ij} = D_{ijkl} \varepsilon_{kl}$
where $\sigma_{ij}$ is the stress tensor, $b_i$ is the body force, $\varepsilon_{ij}$ is the strain tensor, $u_i$ is the displacement vector, and $D_{ijkl}$ is the fourth-order elasticity tensor.
For contact problems, these equations are subject to additional inequality constraints (non-penetration) and nonlinear boundary conditions. The solution is often sought using variational principles and weighted residual methods like the Galerkin method. This involves approximating the displacement field with shape functions $N_i$, leading to a system of nonlinear algebraic equations solved incrementally and iteratively (e.g., using the Newton-Raphson method). This FEA framework allows us to model the precise geometry of the RV reducer main bearing, apply the calculated periodic loads, and extract the resulting elastic deformations and strains for comparison with theory.
FEA Model Setup and Results
A submodel of the RV reducer main bearing was created, including a segment of the inner ring, outer ring, and a single ball. This submodeling approach reduces computational cost while capturing the essential local contact behavior. The model was meshed with a high density of elements in the potential contact regions, primarily using hexahedral elements for accuracy. Contact pairs were defined between the ball and both raceways with a standard augmented Lagrange formulation. The material was defined as GCr15 bearing steel with a Young’s Modulus of 210 GPa and a Poisson’s ratio of 0.3.
The boundary conditions replicated the constraints within the RV reducer: the outer ring’s outer surface was fixed (simulating the interference fit in the housing), while the inner ring’s bore was subjected to a time-varying radial nodal force equivalent to $F_{r1}(\theta)$, where $\theta$ is the crankshaft angle. Two types of analyses were performed:
- Static Analysis: Multiple load steps representing different crankshaft angles (different $Q$ values) were solved independently.
- Transient Dynamic Analysis: A single analysis where the load $Q(t)$ was applied as a smooth periodic function over time, simulating rotation, with inertial effects considered.
The key output was the normal elastic strain (directly related to deformation $\delta$) in the ball at the contact region. For the static analysis, the computed normal strain (deformation) values across different load steps were plotted against the crankshaft angle and compared to the theoretical $\delta$ values from the Hertz model. The results showed strong agreement, with the FEA-predicted deformation curve closely following the theoretical periodic trend. Both curves exhibited the same phase and similar amplitude, with minor deviations attributable to the approximations in the analytical load distribution model and the numerical nature of FEA. The maximum deformation values were nearly identical, confirming the validity of the static Hertz-based stiffness model for the RV reducer main bearing.
$$ \text{Max Theoretical } \delta \approx 1.30 \times 10^{-5} \, m$$
$$ \text{Max FEA (Static) } \delta \approx 1.29 \times 10^{-5} \, m$$
The transient analysis provided further insight into the dynamic response. The strain history showed a stable, periodic response matching the load periodicity after an initial transient phase. This confirmed that the contact stiffness behavior is consistent under dynamic loading conditions relevant to the operational state of the RV reducer. Furthermore, a path was plotted along a diameter of the ball through the contact point. The strain distribution along this path from both static and transient analyses showed the characteristic Hertzian pattern: maximum compressive strain at the contact edges, decreasing towards the center of the contact ellipse and becoming minimal at the ball’s core. This strain field pattern serves as a direct visual validation of the Hertzian contact assumptions within the complex assembly of the RV reducer.
Conclusion
This investigation into the contact stiffness of the main bearing within an RV reducer reveals its significant and complex role in system dynamics. By integrating a static force analysis of the RV reducer with Hertzian contact theory, a nonlinear model for the main bearing’s contact stiffness was developed. The model demonstrates that the stiffness varies periodically with the crankshaft rotation, driven by the periodic load variation from the cycloid drive mechanism. This periodicity is a crucial characteristic that should be considered in high-fidelity dynamic models of RV reducers aimed at predicting vibration or positional error.
A key finding is the existence of an optimal curvature radius coefficient ($f$) for the bearing raceways, which maximizes contact stiffness. For the studied configuration, this optimum was found at $f=0.55$. This insight provides valuable guidance for the tailored design of main bearings in RV reducers, where maximizing stiffness can directly enhance positional accuracy and load capacity.
The theoretical results were rigorously validated using nonlinear finite element analysis. The FEA simulations, encompassing both static and transient regimes, confirmed the periodic variation of contact deformation and the resulting stiffness. The close agreement between the theoretical elastic deformation and the FEA-predicted strain, along with the observed Hertzian stress distribution within the ball, firmly substantiates the correctness of the proposed analytical approach. This combined analytical-numerical methodology offers a reliable tool for analyzing and optimizing the critical interface of the main bearing, contributing to the advancement of high-performance RV reducer design for precision applications.