In the field of precision transmission, the performance of the rotary vector reducer is paramount. The angular transmission error (ATE) stands as a critical metric for evaluating its operational precision and quality. This error is influenced by a confluence of factors, including dimensional tolerances of components, bearing clearances, and the modification of cycloidal gear profiles. Traditional analysis methods often face challenges in accurately modeling the complex, non-linear interactions between these factors within a complete system context. This work addresses these challenges by establishing a comprehensive, parameterized virtual prototype for a rotary vector reducer. We integrate advanced multi-body dynamics simulation techniques to investigate the coupling effects of key dimensional errors on the overall transmission accuracy, providing a foundational methodology for precision design and tolerance optimization.
The core of our methodology lies in a two-stage approach: a fully parameterized Computer-Aided Design (CAD) model and a detailed multi-body dynamics simulation model. The CAD model serves as the geometric foundation, allowing for rapid modification of part dimensions. The dynamics model introduces the physical interactions, including contact forces and compliance, necessary to simulate real-world behavior under load.
Parametric CAD Modeling of Components
Accurate geometric representation is the first step. A parametric model was developed for a standard rotary vector reducer (akin to an RV80E model). The key parameters for the gear train are summarized in Table 1. This parametric approach ensures that any change in a fundamental design variable automatically propagates through the entire model’s geometry.
| Component | Parameter | Value | Component | Parameter | Value |
|---|---|---|---|---|---|
| Involute Gears | Module, \( m \) (mm) | 1.75 | Cycloid-Pin Gear | Eccentricity, \( a \) (mm) | 1.5 |
| Sun Gear Teeth, \( z_1 \) | 12 | Cycloid Gear Teeth, \( z_c \) | 39 | ||
| Planet Gear Teeth, \( z_2 \) | 36 | Pin Teeth Number, \( z_p \) | 40 | ||
| Pin Center Circle Diameter, \( d_p \) (mm) | 153 | ||||
| Pin Diameter, \( d_{rp} \) (mm) | 6 |
The modeling of the involute and cycloid tooth profiles is algorithmically driven. For the involute gear, a single tooth flank profile is generated based on the involute equation. This profile is then rotated and mirrored to create a complete tooth, which is subsequently patterned around the gear’s axis. The fundamental coordinate transformation for generating the involute profile point \( P(x, y) \) for a given roll angle \( \theta \) can be expressed as:
$$
\begin{aligned}
x &= r_b (\cos(\theta) + \theta \sin(\theta)) \\
y &= r_b (\sin(\theta) – \theta \cos(\theta))
\end{aligned}
$$
where \( r_b \) is the base circle radius. For the cycloidal gear, a full set of teeth is generated in one operation to ensure geometric continuity, which is crucial for stable contact definition in subsequent dynamics simulations. The standard cycloidal profile, modified for a roller-pin design, is defined by the following parametric equations:
$$
\begin{aligned}
x &= (r_p – r_{rp}) \sin(\phi) – a \sin(z_p \phi) – \delta_x \\
y &= (r_p – r_{rp}) \cos(\phi) – a \cos(z_p \phi) – \delta_y
\end{aligned}
$$
Here, \( r_p = d_p/2 \) is the pin center circle radius, \( r_{rp} = d_{rp}/2 \) is the pin radius, \( \phi \) is the generating angle, and \( \delta_x, \delta_y \) represent profile modification amounts for backlash and lubrication. The fully parameterized assembly model allows for instant updates to critical dimensions such as the pin center circle diameter \( d_p \) and the pin housing slot diameter \( d_{sp} \), which are the focus of the subsequent error analysis.
Multi-Body Dynamics Simulation Model
The imported CAD assembly is inert. To create a functioning dynamic model, kinematic joints, forces, and, most importantly, contact conditions must be applied. The model was configured for a flange-output style rotary vector reducer: the pin housing is fixed, the sun gear is given a rotational speed input, and a load torque is applied to the output flange.
| Component | Constraint / Load | Description |
|---|---|---|
| Pin Housing | Fixed Joint | Grounded to the global reference frame. |
| Sun Gear | Rotational Speed Input | \( \omega_{in}(t) = 10890 \cdot step(time, 0, 0, 0.2, 1) \) deg/s. |
| Output Flange | Load Torque | \( T_{load}(t) = -784000 \cdot step(time, 0.2, 0, 0.4, 1) \) N·mm. |
| Pins | Planar Joint | Constrained to the housing slot, allowing only radial movement. |
Contact Force Modeling
The non-linear dynamics of the rotary vector reducer are dominated by contact forces. Two primary contact pairs exist: the involute gear mesh (sun-planet) and the cycloid-pin mesh. A third critical contact is between the pins and their housing slots. All contacts are modeled using a penalty-based method with a continuous contact force algorithm based on a modified Hertzian theory, such as the Lankarani-Nikravesh model. The general form of the normal contact force \( f_n \) is:
$$
f_n = k \delta^{m_1} + c \dot{\delta} |\dot{\delta}|^{m_2} \delta^{m_3}
$$
where \( \delta \) is the penetration depth, \( \dot{\delta} \) is its time derivative (relative normal velocity), \( k \) is the contact stiffness coefficient, \( c \) is a damping coefficient, and \( m_1, m_2, m_3 \) are exponents determining the non-linear behavior of stiffness, damping, and indentation. The contact detection for these complex curved surfaces employs a hybrid algorithm combining bounding-box checks and precise geometrical evaluations relative to local coordinate frames. The contacting surfaces (e.g., the cycloid tooth flank and the pin cylinder) are discretized into triangular facets to accurately compute the penetration and force direction.
Modeling Bearing Clearance with Bushings
Explicitly modeling every rolling element in the numerous bearings of a rotary vector reducer is computationally prohibitive. An effective and efficient alternative is to use bushing elements. A Bushing Force model simulates the radial stiffness and damping of a bearing, including its internal clearance. The bushing connects two parts via a set of discrete, radially distributed spring-damper elements. The total restoring force is the vector sum of the forces from each element. When the relative displacement between the inner and outer race is less than the radial clearance \( c_r \), no force is generated. Outside this zone, the force follows a linear or non-linear stiffness law. The radial force component for a single element oriented at angle \( \theta_i \) can be conceptualized as:
$$
F_{r,i} = \begin{cases}
0 & \text{if } \delta_r < c_r \\
k_{eff} (\delta_r – c_r) + c_{eff} \dot{\delta}_r & \text{if } \delta_r \ge c_r
\end{cases}
$$
where \( \delta_r \) is the radial displacement of the shaft center relative to the housing bore center. This method is applied to the crankshaft support bearings, the planet carrier bearings, and the output flange support bearings, effectively introducing the critical compliance and play associated with bearing internal clearance into the dynamics of the rotary vector reducer.
Dynamics Simulation and Error Analysis
Before error analysis, the basic kinematics of the model were verified. Under no-load running conditions, the simulated angular velocities of the planet gears and the output flange closely matched the theoretical values derived from the gear ratios, confirming the correct implementation of joints and constraints. The primary focus, however, is on the transmission performance under load and with dimensional variations.
Cycloid-Pin Meshing Characteristics
Under rated load, the contact forces between the cycloid gear and the pin teeth exhibit the expected periodic behavior. Figure 9 in the source material (not reproduced here) shows the normal contact force for pins 1 through 5 over several cycloid gear cycles. The forces are not equal due to the phase differences between the two cycloid gears and load sharing characteristics. The contact force pattern validates that the cycloid-pin meshing in the virtual rotary vector reducer model is functioning as intended, with forces transitioning smoothly between pins.
Analysis of Angular Transmission Error Under Dimensional Variations
The Angular Transmission Error (ATE) is the definitive measure of precision for a rotary vector reducer. It is defined as the difference between the theoretical output position (input position divided by the ideal reduction ratio \( R \)) and the actual simulated output position:
$$
\theta_{er} = \frac{\theta_{in}}{R} – \theta_{out}
$$
To investigate the influence of manufacturing tolerances, we focused on two key dimensions within the same assembly chain: the pin center circle diameter \( d_p \) and the pin slot diameter in the housing \( d_{sp} \). Their nominal values and tolerance ranges form the basis for an error combination study. Multiple CAD models were programmatically generated by varying these two diameters within their specified tolerance bands.
| Dimension | Nominal & Tolerance (mm) | Variation Levels (mm) |
|---|---|---|
| Pin Center Circle Diameter, \( d_p \) | \( \phi153_{-0.061}^{-0.038} \) | 153.039, 153.045, 153.051, 153.056, 153.062 |
| Pin Slot Diameter, \( d_{sp} \) | \( \phi6_{+0.004}^{+0.012} \) | 6.004, 6.006, 6.008, 6.010, 6.012 |
This created a matrix of 25 distinct virtual prototypes (M_1 to M_25), each with a unique combination of \( d_p \) and \( d_{sp} \). Each model was subjected to a multi-body dynamics simulation under the rated operating conditions defined earlier. The key results extracted were the peak-to-peak ATE and the mean ATE over a stable operating period.
The simulation results reveal a clear and non-linear relationship between the dimensional errors and the performance of the rotary vector reducer. The data for peak-to-peak ATE and mean ATE are consolidated into the following analysis tables.
| \( d_p \) \ \( d_{sp} \) | 6.004 mm | 6.006 mm | 6.008 mm | 6.010 mm | 6.012 mm |
|---|---|---|---|---|---|
| 153.039 mm | 0.319 | 0.303 | 0.313 | 0.296 | 0.318 |
| 153.045 mm | 0.313 | 0.299 | 0.280 | 0.297 | 0.338 |
| 153.051 mm | 0.282 | 0.275 | 0.292 | 0.294 | 0.282 |
| 153.056 mm | 0.280 | 0.307 | 0.276 | 0.316 | 0.291 |
| 153.062 mm | 0.302 | 0.297 | 0.284 | 0.292 | 0.393 |
| \( d_p \) \ \( d_{sp} \) | 6.004 mm | 6.006 mm | 6.008 mm | 6.010 mm | 6.012 mm |
|---|---|---|---|---|---|
| 153.039 mm | 3.051 | 3.116 | 3.180 | 3.242 | 3.308 |
| 153.045 mm | 3.185 | 3.249 | 3.314 | 3.377 | 3.440 |
| 153.051 mm | 3.315 | 3.382 | 3.444 | 3.507 | 3.571 |
| 153.056 mm | 3.426 | 3.490 | 3.552 | 3.630 | 3.679 |
| 153.062 mm | 3.555 | 3.619 | 3.681 | 3.745 | 3.832 |
The results demonstrate several important trends for the rotary vector reducer:
1. Combined Effect on Error Magnitude: Both the peak-to-peak (dynamic) error and the mean (systematic offset) error are sensitive to changes in \( d_p \) and \( d_{sp} \). The mean error shows a more predictable trend, generally increasing as either dimension increases. This can be attributed to a systematic shift in the meshing action and load distribution within the cycloid stage. The relationship is approximated by a linear regression model for the mean error \( \bar{\theta}_{er} \):
$$
\bar{\theta}_{er} \approx \alpha (d_p – d_{p,base}) + \beta (d_{sp} – d_{sp,base}) + \bar{\theta}_{er, base}
$$
where \( \alpha \) and \( \beta \) are sensitivity coefficients determined from the data.
2. Non-linear Interaction and Inflection Points: The peak-to-peak error does not follow a simple monotonic trend. For a fixed \( d_p \), varying \( d_{sp} \) can lead to a minimum in the peak-to-peak error, indicating an optimal pin-slot clearance for that specific pin circle size. Similarly, for a fixed \( d_{sp} \), varying \( d_p \) shows inflection points. This non-linearity highlights the complex coupling between these two errors. An optimal combination (e.g., around M_11 or M_13 in this study) minimizes the dynamic fluctuation in the rotary vector reducer’s output.
3. Phase Shift in Error Curve: Beyond altering the amplitude of the ATE, different error combinations also induce a phase shift in the periodic error signal. This is analogous to changing the initial contact condition of the cycloid gear with the pin set. The error waveform for configurations with similar peak-to-peak values (e.g., M_4, M_10, M_23) are offset in time, implying that the “signature” of the error is altered by the dimensional setup.
Conclusion and Implications for Rotary Vector Reducer Design
This integrated study successfully establishes a high-fidelity, parameterized virtual prototyping framework for the rotary vector reducer. The combination of CAD parameterization and advanced multi-body dynamics simulation provides a powerful tool for analyzing transmission precision. The model accurately reflects the influence of key design and manufacturing parameters.
The systematic investigation into the coupled effects of the pin center circle diameter and the pin slot diameter reveals that dimensional errors do not merely additively contribute to the total angular transmission error. Instead, they interact in a non-linear manner, creating opportunities for error compensation through selective assembly or targeted tolerance allocation. The existence of inflection points in the error curves suggests that for a given nominal design, there exists an optimal combination of these within-tolerance dimensions that minimizes the dynamic transmission error of the assembled rotary vector reducer.
These findings have direct implications for the precision manufacturing of the rotary vector reducer. They move beyond viewing tolerances merely as constraints and instead treat them as potential design variables for optimizing performance. Future work will expand this analysis to include other periodic error sources such as eccentricity errors, tooth profile modifications of the cycloid gear, and the interaction with bearing clearance grades. Furthermore, a spectral analysis of the ATE output could decompose the total error into frequency components directly linked to specific geometrical periods (e.g., cycloid tooth frequency, planet carrier revolution frequency), providing even deeper insight into the error coupling mechanisms within the complex drivetrain of the rotary vector reducer.