In the field of precision motion control and robotics, the strain wave gear, also known as harmonic drive, has emerged as a critical component due to its unique advantages. These include high reduction ratios in a compact package, exceptional positional accuracy with near-zero backlash, high torque capacity, and the ability to transmit motion through sealed spaces. The core of this transmission system lies in the meshing between a rigid circular spline and a flexible spline that undergoes controlled elastic deformation. The tooth profile geometry of these two components directly dictates the performance metrics such as load distribution, stress concentration, contact ratio, and ultimately, the longevity and efficiency of the strain wave gear. Therefore, the pursuit of an optimal tooth profile is a central theme in advanced mechanical design.
Traditional designs often employ involute or simple circular arc profiles. However, to enhance load-bearing capacity and smoothness of motion, more complex profiles like the double-arc tooth form have been developed. This article delves into the methodology for deriving the conjugate tooth profile of the rigid spline starting from a specified cutter profile for the flexible spline. Specifically, I explore the application of the CTC (Circular-Tangent-Circular) double-arc tooth profile for strain wave gears. The primary objective is to establish a complete design chain: from the hob’s基准齿形 (benchmark tooth profile) used to manufacture the flexspline, to the theoretical conjugate profile of the circular spline, and finally to the practical pinion cutter profile needed to machine the circular spline. The mathematical backbone of this derivation is the Tooth Profile Normal Method, a direct application of the fundamental law of gearing.
The fundamental law of gearing states that for two profiles to maintain continuous contact during motion transmission, the common normal at the point of contact must always pass through the instantaneous center of rotation, or the pitch point. The Tooth Profile Normal Method leverages this principle directly. For a given point on one tooth profile (e.g., the cutter or flexspline), I calculate the condition under which its normal intersects the pitch line at the correct instant. This condition defines the necessary relative motion between the tool and the workpiece, allowing for the transformation of coordinates to obtain the conjugate profile. This method is particularly intuitive for solving planar gear generation problems, such as those encountered in the manufacture of strain wave gear components.
The overall technical design flow for the CTC double-arc strain wave gear is systematic and is outlined in the conceptual diagram above. It begins with the selection of the hob’s基准齿形, which is a CTC double-arc profile. This hob is used to generate the tooth profile on the flexspline blank via a gear hobbing process, which is kinematically equivalent to the meshing of a rack (the hob’s tooth shape) and a gear (the flexspline). Subsequently, the obtained flexspline tooth profile is used as the known entity. Considering the specific wave generator motion (typically an elliptical cam) that deforms the flexspline, the conjugate tooth profile for the rigid circular spline is solved. Since the theoretical conjugate profile may consist of complex curves, a fitting process using the least-squares method is employed to approximate it with a standard CTC double-arc profile. This step is crucial for practical manufacturing, as it allows the use of existing, standardized cutters. Finally, from this fitted circular spline profile, the required profile for the pinion cutter used to machine the internal teeth of the circular spline is determined.
The chosen hob benchmark tooth profile, the CTC double-arc shape, consists of three distinct segments: a convex circular arc at the addendum (tooth tip), a straight tangent line connecting the two arcs, and a concave circular arc at the dedendum (tooth root). This combination aims to provide favorable contact conditions and stress distribution. The mathematical representation of this profile is established in a coordinate system attached to the hob (or the equivalent rack). Let the parameterized equations define this profile, which serves as the genesis for all subsequent derivations in the strain wave gear design process.
The profile is defined in the hob coordinate system \( o_1 – x_1y_1 \), with the \( y_1 \)-axis as the symmetry axis and the origin on the pitch line. The segments are:
1. Addendum Circular Arc (Segment AB): This is a convex arc with radius \( \rho_f \). Its center is offset from the origin. The parametric equations, where \( \alpha \) is the parameter, are:
$$ x_1 = \rho_f \cos\alpha – l_f $$
$$ y_1 = -\rho_f \sin\alpha + e_f $$
Here, \( l_f \) and \( e_f \) are geometric offset parameters defining the arc’s center location relative to the hob’s datum line.
2. Middle Tangent Line (Segment BC): This straight line is tangent to both the addendum and dedendum arcs. Its equation in slope-intercept form is:
$$ y_1 = k x_1 + b $$
The slope \( k \) and intercept \( b \) are determined by the tangency conditions with the adjacent arcs, ensuring a smooth \( C^1 \) continuous transition in the strain wave gear tooth form.
3. Dedendum Circular Arc (Segment CD): This is a concave arc with radius \( \rho_a \). Its parametric equations, with parameter \( \beta \), are:
$$ x_1 = -\rho_a \cos\beta + l_a + \frac{\pi m}{2} $$
$$ y_1 = \rho_a \sin\beta + e_a $$
The parameters \( l_a \) and \( e_a \) define its center, and \( m \) is the module of the strain wave gear.
The generation of the flexspline tooth profile via hobbing is modeled as the conjugate action between a rack (the hob tooth) and a gear (the flexspline blank). Two coordinate systems are used: \( o_1 – x_1y_1 \) fixed to the rack (hob), and \( o_2 – x_2y_2 \) fixed to the flexspline, with its origin at the gear center. Initially, these systems are aligned. According to the fundamental law of gearing, for a point \( M(x_1, y_1) \) on the rack profile to be a point of contact, the normal at that point must pass through the pitch point \( P \). The angle \( \gamma \) of the tangent at \( M \) is given by:
$$ \tan\gamma = \frac{dy_1/du}{dx_1/du} $$
where \( u \) is the parameter (\( \alpha \), \( x_1 \), or \( \beta \)) for the respective segment. The distance \( l \) that the rack must translate from its initial position for contact to occur at \( M \) is:
$$ l = x_1 + y_1 \tan\gamma $$
The corresponding rotation angle \( \phi_2 \) of the flexspline is \( \phi_2 = l / \rho_2 \), where \( \rho_2 = r_2 + x m \). Here, \( r_2 \) is the pitch radius of the flexspline, \( x \) is the addendum modification coefficient (not a coordinate), and the sign of \( l \) (and thus \( \phi_2 \)) indicates the direction of motion.
The conjugate flexspline profile coordinates \( (x_2, y_2) \) are obtained by transforming the contact point coordinates from system \( o_1 \) to system \( o_2 \) after the rack has moved by \( l \) and the gear has rotated by \( \phi_2 \). The homogeneous transformation matrix is:
$$
\begin{bmatrix}
x_2 \\
y_2 \\
1
\end{bmatrix}
=
\begin{bmatrix}
\cos\phi_2 & \sin\phi_2 & \rho_2 (\sin\phi_2 – \phi_2 \cos\phi_2) \\
-\sin\phi_2 & \cos\phi_2 & \rho_2 (\cos\phi_2 + \phi_2 \sin\phi_2) \\
0 & 0 & 1
\end{bmatrix}
\begin{bmatrix}
x_1 \\
y_1 \\
1
\end{bmatrix}
$$
This yields the general conjugate profile equations:
$$ x_2 = x_1 \cos\phi_2 + y_1 \sin\phi_2 + \rho_2 (\sin\phi_2 – \phi_2 \cos\phi_2) $$
$$ y_2 = y_1 \cos\phi_2 – x_1 \sin\phi_2 + \rho_2 (\cos\phi_2 + \phi_2 \sin\phi_2) $$
These equations are applied piecewise to each segment of the hob profile to generate the complete flexspline tooth flank for the strain wave gear.
1. Conjugation of the Hob Addendum Arc (AB) to the Flexspline Dedendum:
For points on arc AB, calculation typically yields \( l < 0 \), meaning the rack moves in the negative \( x_1 \) direction. Substituting the parametric equations for \( x_1 \) and \( y_1 \) from segment AB into the general transformation gives the flexspline root profile:
$$ x_2 = (\rho_f \cos\alpha – l_f) \cos\phi_2 + (-\rho_f \sin\alpha + e_f) \sin\phi_2 + \rho_2 (\sin\phi_2 – \phi_2 \cos\phi_2) $$
$$ y_2 = (-\rho_f \sin\alpha + e_f) \cos\phi_2 – (\rho_f \cos\alpha – l_f) \sin\phi_2 + \rho_2 (\cos\phi_2 + \phi_2 \sin\phi_2) $$
where \( \phi_2 = l / \rho_2 \) and \( l = (\rho_f \cos\alpha – l_f) + (-\rho_f \sin\alpha + e_f) \tan\gamma \), with \( \tan\gamma \) derived from the arc’s derivative.
2. Conjugation of the Hob Middle Tangent (BC) to the Flexspline Mid-flank:
For the tangent line segment, the sign of \( l \) varies along its length. The lower part near the dedendum yields \( l < 0 \), and the upper part near the addendum yields \( l > 0 \). The conjugate profile is:
$$ x_2 = x_1 \cos\phi_2 + (k x_1 + b) \sin\phi_2 + \rho_2 (\sin\phi_2 – \phi_2 \cos\phi_2) $$
$$ y_2 = (k x_1 + b) \cos\phi_2 – x_1 \sin\phi_2 + \rho_2 (\cos\phi_2 + \phi_2 \sin\phi_2) $$
Here, \( x_1 \) is the variable along the line, and \( \phi_2 \) is calculated using \( l = x_1 + (k x_1 + b) \cdot k \) (since \( \tan\gamma = k \) for a straight line).
3. Conjugation of the Hob Dedendum Arc (CD) to the Flexspline Addendum:
For points on arc CD, \( l > 0 \), indicating rack movement in the positive \( x_1 \) direction. The conjugate flexspline tip profile is:
$$ x_2 = (-\rho_a \cos\beta + l_a + \frac{\pi m}{2}) \cos\phi_2 + (\rho_a \sin\beta + e_a) \sin\phi_2 + \rho_2 (\sin\phi_2 – \phi_2 \cos\phi_2) $$
$$ y_2 = (\rho_a \sin\beta + e_a) \cos\phi_2 – (-\rho_a \cos\beta + l_a + \frac{\pi m}{2}) \sin\phi_2 + \rho_2 (\cos\phi_2 + \phi_2 \sin\phi_2) $$
with \( \phi_2 \) determined from the corresponding \( l \) value.
The next critical phase in strain wave gear design is deriving the rigid circular spline’s tooth profile that conjugates with the now-known flexspline profile under the specific motion imposed by the wave generator. The wave generator, usually an elliptical cam, deforms the flexible spline so that its teeth engage with those of the circular spline at two diametrically opposite regions. The conjugation condition is again governed by the fundamental law, but now applied between two rotating bodies with a variable center distance due to the flexspline’s deformation. A common approach uses a kinematic model where the flexspline’s neutral curve is assumed to deform into a curve defined by the wave generator’s shape (e.g., an ellipse). For each angular position of the wave generator, a contact condition between a point on the flexspline tooth and the circular spline tooth is established.
Without delving into the exhaustive derivation here, the process involves defining coordinate systems attached to the flexspline \( (X_f, Y_f) \), the circular spline \( (X_c, Y_c) \), and the wave generator. The deformation function \( \delta(\theta) \) describes the radial displacement of the flexspline neutral curve. For a given point on the flexspline tooth profile, its position relative to the circular spline is calculated as a function of the wave generator angle. The condition for contact is that the relative velocity at a potential contact point is zero along the common normal direction. Solving this condition yields the locus of contact points, which defines the required circular spline tooth profile. The resulting profile is typically a set of discrete points calculated numerically.
To make the circular spline profile manufacturable with standard tools, it is highly desirable to approximate it with a standard double-arc CTC profile. This is achieved through a least-squares fitting procedure. The goal is to find the parameters of a CTC curve (radii \( \rho_f’ \) and \( \rho_a’ \), offsets \( l_f’, e_f’, l_a’, e_a’ \), and tangent line parameters) that minimize the sum of squared distances between the theoretical discrete points \( (X_{ci}, Y_{ci}) \) and the closest points on the proposed CTC profile. The objective function \( F \) to be minimized is:
$$ F(\mathbf{P}) = \sum_{i=1}^{N} d_i^2(\mathbf{P}) $$
where \( \mathbf{P} \) is the vector of CTC profile parameters, and \( d_i \) is the perpendicular distance from the \( i \)-th theoretical point to the nearest segment (arc or line) of the CTC curve. This nonlinear optimization problem can be solved using algorithms like the Levenberg-Marquardt method. The outcome is a set of optimal CTC parameters that best represent the conjugate circular spline tooth profile for the strain wave gear.
The following table summarizes typical parameter ranges and results from such a fitting process for a medium-sized strain wave gear design. These values are illustrative and depend heavily on the module, number of teeth, and wave generator configuration.
| Parameter | Symbol | Hob Profile (Input) | Fitted Circular Spline Profile (Output) | Unit |
|---|---|---|---|---|
| Module | \( m \) | 0.5 | 0.5 | mm |
| Addendum Arc Radius | \( \rho_f, \rho_f’ \) | 0.65 m | 0.68 m | mm |
| Dedendum Arc Radius | \( \rho_a, \rho_a’ \) | 0.80 m | 0.78 m | mm |
| Addendum Center Offset X | \( l_f, l_f’ \) | 0.25 m | 0.23 m | mm |
| Addendum Center Offset Y | \( e_f, e_f’ \) | 0.30 m | 0.32 m | mm |
| Dedendum Center Offset X | \( l_a, l_a’ \) | 0.35 m | 0.38 m | mm |
| Dedendum Center Offset Y | \( e_a, e_a’ \) | 0.40 m | 0.42 m | mm |
| Tangent Line Slope | \( k, k’ \) | -0.15 | -0.14 | – |
| RMS Fitting Error | – | – | < 0.002 m | mm |
The small fitting errors confirm that the theoretical conjugate profile for the circular spline in a strain wave gear is, to a very close approximation, also a CTC double-arc profile. This is a significant finding as it allows the use of the same family of standard cutters for both splines, greatly simplifying tooling inventory and manufacturing logistics for strain wave gear production.
Once the circular spline’s CTC profile parameters are determined, the final step is to deduce the required profile for the pinion cutter that will be used to machine the internal teeth of the circular spline. This is essentially the inverse of the process used for the flexspline. The pinion cutter generation can be modeled as the meshing between an external cutter (a pinion with a small number of teeth) and the internal circular spline. The known circular spline tooth profile \( (X_c, Y_c) \) and the known relative motion (the cutter rotating and feeding into the workpiece) are used. Applying the tooth profile normal method again, but this time with the circular spline as the fixed “rack” in a coordinate transformation, the conjugate profile of the pinion cutter can be solved. The governing equation for the transformation, similar to before but for an internal gear pair, is used. The resulting cutter profile ensures that during the gear shaping process, it will accurately generate the desired CTC double-arc teeth on the circular spline of the strain wave gear.
The mathematical model for the pinion cutter derivation involves setting up the meshing condition between the internal circular spline tooth profile \( \mathbf{r}_c(u) \) and the pinion cutter profile \( \mathbf{r}_p \). The condition is that the relative velocity at the contact point is orthogonal to the common normal vector \( \mathbf{n}_c \):
$$ \mathbf{n}_c \cdot (\mathbf{v}_c^{(c)} – \mathbf{v}_p^{(c)}) = 0 $$
Here, \( \mathbf{v}_c^{(c)} \) and \( \mathbf{v}_p^{(c)} \) are the velocities of the contact point on the circular spline and pinion cutter, respectively, expressed in a common coordinate frame. Solving this equation along with the coordinate transformation from the circular spline system to the pinion cutter system yields the pinion cutter profile parameters. For the CTC profile, this results in another set of double-arc parameters specific to the cutter.
The advantages of the CTC double-arc profile in strain wave gears are multifaceted. The circular arcs allow for a larger radius of curvature at contact points compared to sharp corners, reducing contact stress and Hertzian pressure. The tangent line ensures smooth transition between the arcs, avoiding stress concentrations. The double-arc design can be optimized to maximize the contact ratio, which in strain wave gearing is inherently high due to the simultaneous engagement of many teeth (often 15-30% of total teeth). A higher contact ratio distributes the load more evenly, leading to higher torque capacity and smoother operation with reduced vibration and noise—critical for precision applications like robotics and aerospace actuators where strain wave gears are prevalent.
Furthermore, the methodology presented here, based on the tooth profile normal method, provides a robust and general framework. It is not limited to the CTC profile but can be adapted to any given rack or tool profile to derive its conjugate in a strain wave gear or conventional gear system. The use of numerical methods for fitting makes the approach practical for real-world engineering design, bridging the gap between theoretical conjugate geometry and manufacturable profiles.
In conclusion, this exploration successfully details a comprehensive procedure for designing the tooth profiles of a CTC double-arc strain wave gear. Starting from a defined hob基准齿形, I applied the tooth profile normal method to derive the conjugate flexspline tooth profile. This profile was then used, considering the wave generator kinematics, to solve for the theoretical conjugate profile of the rigid circular spline. Practical manufacturing considerations led to the use of a least-squares fitting technique to approximate this profile with a standard CTC double-arc curve, demonstrating that both splines in the strain wave gear pair can indeed share the same fundamental tooth form. Finally, the profile for the necessary pinion cutter was deduced. This end-to-end methodology enhances the design toolkit for advanced strain wave gears, potentially leading to improved performance, standardized manufacturing, and wider adoption of these efficient and compact transmission systems in high-tech industries. The consistent application of the fundamental law of gearing through the tooth profile normal method proves to be a powerful and intuitive technique for tackling complex conjugate geometry problems in gear design.
The continued evolution of strain wave gear technology will likely involve further optimization of these arc parameters using finite element analysis and multi-objective genetic algorithms to balance stress, wear, and efficiency. However, the foundational mathematical approach described herein remains a critical first step in that advanced optimization process for high-performance strain wave gear systems.