High-ratio spiral bevel gears are attractive when a single stage must replace a multi-stage train: the axes intersect, the package is compact, and the curved flanks provide gradual engagement. The difficult case is a pinion with very few teeth. A 4:41 pair gives a reduction of 10.25:1, but the pinion root section is small, positive modification tends to sharpen the tooth tip, and the wheel is conventionally much stronger. The design problem is therefore not simply to cut a four-tooth pinion. It is to redistribute tooth volume and contact location so that both members reach comparable bending safety while remaining manufacturable.
Design problem: a four-tooth pinion in a 10.25:1 stage
A four-tooth pinion is far below the usual 17-tooth undercut limit for a standard 20° involute gear. In a bevel pair, the midpoint equivalent gear is the relevant reference, and its virtual tooth number remains very low for a shallow pinion pitch cone. Standard addendum distribution makes the pinion root weaker than the wheel root, so the pair is limited by the pinion rather than by the wheel. Increasing the pinion addendum by radial modification helps, but an excessive positive shift can make the tip too thin; a negative shift on the wheel can also thin the wheel tip. The modification must therefore be chosen as a constrained balance, not as an isolated correction.
Pitch-cone external meshing addresses the balance directly. In the midpoint equivalent gear, the active contact path is deliberately placed beyond the pitch point. The pinion receives a large positive radial modification and the wheel a negative one. The pinion tooth volume grows, the wheel tooth volume is reduced, and the bending capacities of the two members move toward each other. Because the active path remains on one side of the pitch point, the sliding direction relative to the flank does not reverse at the pitch point during the mesh cycle. That kinematic feature is usually favorable for excitation, vibration, and noise, although it is a design rationale rather than a substitute for dynamic testing.
Pitch-cone external meshing and the sign of the wheel addendum
Two relative positions of the pitch cone are possible. In the front configuration, the active path lies before the pitch point. The pinion face cone moves inside the pitch cone, the pinion addendum becomes negative, and the pinion shrinks while the wheel grows. This worsens the strength imbalance and is not used for the high-ratio case. In the rear configuration, the active path lies after the pitch point. The wheel pitch cone then lies outside its face cone, so the wheel addendum becomes negative. The pinion, by contrast, receives a large positive addendum. This is the configuration used for small-tooth-count, high-reduction spiral bevel gears.
The degree of pitch-cone external meshing can be expressed on the midpoint equivalent gear as
$$
\lambda=\frac{r_{va2}-r_{v2}}{m_{mt}},
$$
where \(r_{va2}\) is the wheel equivalent tip radius, \(r_{v2}\) is the wheel equivalent pitch radius, and \(m_{mt}\) is the midpoint transverse module. For the rear configuration, the same quantity is the wheel addendum coefficient \(c_{\mathrm{ham}}\), and the meshing condition becomes
$$
c_{\mathrm{ham}}=\frac{1}{2}-\frac{x_1}{k_d\cos\beta},\qquad
\lambda=c_{\mathrm{ham}}<0,\qquad
x_2=-x_1.
$$
Here \(x_1\) and \(x_2\) are the radial modification coefficients of the pinion and wheel, \(k_d\) is the tooth-height coefficient, and \(\beta\) is the midpoint spiral angle. A negative \(\lambda\) requires
$$
x_1>\frac{k_d\cos\beta}{2}.
$$
In the 4:41 example, \(\lambda=-0.02\). The wheel addendum is therefore slightly negative, while the pinion addendum is strongly positive. The pinion is enlarged and the wheel is relieved, which is the intended strength-sharing mechanism.
Geometric constraints that decide the modification
Three constraints dominate the feasible modification range. First, both members must retain sufficient tip thickness. The pinion’s positive modification and the wheel’s negative modification can both thin the tip, so the midpoint tip thickness is checked against a lower limit of \(0.25m_{mt}\). Second, the radial modification must clear the undercut limit on the midpoint virtual gear. The minimum virtual tooth number is evaluated from the standard relation \(z_{v\min}=2k/\sin^2\alpha_{vt}\) with \(k=k_d\cos\beta/2\), and the selected radial shifts must exceed the corresponding minimum values. Third, the total contact ratio must remain high enough for continuous, smooth engagement. For spiral bevel gears, the transverse and axial components combine as \(\varepsilon_o=\sqrt{\varepsilon_\alpha^2+\varepsilon_\beta^2}\), and the design target is \(\varepsilon_o\ge1.5\).
These constraints are not independent. A larger positive pinion shift improves the pinion root section but reduces pinion tip thickness. A larger negative wheel shift balances the pinion but can thin the wheel tip and reduce contact ratio. The modification coefficients are therefore chosen inside a narrow feasible domain, with the pitch-cone external meshing condition supplying the direction of the shift rather than a single fixed value.
A 4:41 design point and its solid model
The demonstration pair uses 4 and 41 teeth, a large-end module of 4.5 mm, a normal pressure angle of 20°, a shaft angle of 90°, a midpoint spiral angle of 36°, and a pitch-cone external meshing coefficient of \(-0.02\). The pinion receives positive radial and tangential modification; the wheel receives opposite radial modification and negative tangential modification. The wheel pitch angle is 84.428°, so the wheel lies in the formate range and can be cut by the forming method. The pinion is generated to obtain the conjugate flank. A constant-tooth-depth spiral bevel geometry is used, with the whole tooth depth held at 5.845 mm for both members.
| Parameter | Pinion | Wheel |
|---|---|---|
| Number of teeth | 4 | 41 |
| Large-end module / mm | 4.5 | |
| Normal pressure angle / (°) | 20 | |
| Shaft angle / (°) | 90 | |
| Midpoint spiral angle / (°) | 36 | 36 |
| Pitch cone angle / (°) | 5.572 | 84.428 |
| Pitch diameter / mm | 18.0 | 184.5 |
| Addendum height / mm | 5.300 | -0.104 |
| Whole tooth depth / mm | 5.845 | 5.845 |
| Midpoint tip thickness / mm | 1.893 | 2.966 |
| Total contact ratio | 1.696 | 1.696 |
| Pitch-cone external meshing coefficient | -0.02 | |
The wheel flank was represented as a point cloud from the cutter surface. Coordinate transformation and the meshing equation produced the conjugate pinion flank, and the two point clouds were converted into a solid model. The solid model reproduced the 5.845 mm whole depth exactly. The midpoint pinion tip thickness was 1.880 mm against a design value of 1.893 mm, an error of 0.66%; the wheel value was 2.953 mm against 2.966 mm, an error of 0.43%. Both tip thicknesses remain well above the 1.125 mm lower limit set by \(0.25m_{mt}\), and the contact ratio of 1.696 exceeds the 1.5 requirement.
Bending strength under a 50 N·m pinion torque
The tooth model was split into sweepable blocks and meshed with hexahedral elements for the finite-element check. A 50 N·m torque was applied to the pinion. The pinion support released only the tangential degree of freedom, the wheel was fixed, and the contact pair used a friction coefficient of 0.03. One pinion tooth engagement was sampled in 10° increments at 13 positions. At each position, the maximum root bending stress was recorded for both members.
The maximum pinion root bending stress was 320.35 MPa, and the maximum wheel root bending stress was 324.52 MPa. The two maxima differ by only about 1.3%. Using a 439 MPa limiting bending stress together with the life, temperature, reliability, and design-safety factors, the allowable bending stresses were 411.99 MPa for the pinion and 428.32 MPa for the wheel. The resulting safety factors are summarized below.
| Item | Pinion | Wheel |
|---|---|---|
| Allowable bending stress / MPa | 411.99 | 428.32 |
| Maximum root bending stress / MPa | 320.35 | 324.52 |
| Bending safety factor | 1.285 | 1.322 |
Both safety factors exceed 1.0, and their difference is 0.037, below the 0.1 equal-strength criterion. The wheel carries a slightly higher stress but also has a higher allowable stress, so the two members reach nearly the same bending safety. This is the intended outcome of the pitch-cone external meshing design: the four-tooth pinion is no longer the sole weak link.
Contact ellipse and cutting trial
A kinematic contact simulation of the modified pinion produced a regular elliptical instantaneous contact zone at the midpoint of the tooth. The ellipse is consistent with localized conjugate contact after modification and does not indicate edge contact. The contact check is geometric; the bending check above supplies the load-carrying evidence.
The pair was then cut on a semi-CNC GH-35 machine. The wheel was formate-cut in three depth passes followed by a finishing pass. The pinion was generated single-side, with separate correction of the convex and concave flanks. After cutting, the pair was run on a rolling tester. No tooth-profile distortion was observed. On the wheel convex flank, the contact pattern was an inner diagonal ellipse occupying 45% of the face width and 70% of the working tooth height, positioned slightly toward the small end. That pattern meets the design target for this pair.
The manufacturing route is part of the design logic. The wheel pitch angle of 84.428° places it in the formate range, while the pinion must be generated to match the wheel flank. The cutting trial therefore tests both the modification calculation and the assumed division between formate and generated members.
What the demonstration establishes
The 4:41 pair shows a closed design chain for small-tooth-count spiral bevel gears: a pitch-cone external meshing condition sets the sign and scale of the modification; tip-thickness, undercut, and contact-ratio constraints bound the feasible domain; a solid model verifies the tooth geometry; finite-element analysis shows nearly equal bending safety; and cutting plus rolling confirms that the contact pattern can be produced in practice.
The result is a design-point demonstration rather than a universal recipe. Other ratios, shaft angles, modules, and load levels require recalculation of \(\lambda\), the radial and tangential modifications, and the contact-ratio margin. The static bending balance also does not by itself establish fatigue life, dynamic load sharing, or thermal behavior. Those are separate validation topics. What the example does establish is that pitch-cone external meshing is a practical route for high-reduction, small-tooth-count spiral bevel gears when a compact single-stage angle drive is required.