Why contact behaviour sets the design limit
Spiral bevel gears transmit motion between intersecting shafts through teeth whose traces are curved rather than straight. The curved trace lengthens the axial contact ratio and enlarges the simultaneous contact area compared with straight and skew bevel gears, so torque is handed over more gradually and with less excitation. This is the property that places spiral bevel gears in the input stages of high-speed, heavy-duty reducers, where the pinion runs at motor speed and carries the full torque of the train.
Teeth in such service fail through the contact interface rather than through the shaft: temperature gradients across the flank, alternating mesh stress, scuffing, efficiency loss and progressive wear all originate in the contact patch. Classical tooth contact analysis, as formulated by Gleason, evaluates the position and shape of the contact pattern on a light-load model in which load-induced deflection is neglected. That is a geometric answer, not a stress answer. Loaded tooth contact formulations have since added elastic deformation, vibration and impact, and the alternation of single- and double-tooth meshing, and they have been used to compare bending stress under differing duty, to quantify the influence of friction on static contact, and to relate flank geometry to the spatial meshing path. What remains less well characterised is how contact stress evolves tooth by tooth and instant by instant within one mesh cycle at the loads and speeds that a reducer input actually sees.
Gear set, material and finite element model
The object of study is the input pinion of a reduction gearbox, coupled directly to the motor output and therefore operating at high speed. The full tooth surface was generated from the tooth trace and profile geometry, and a solid model was built from those surfaces for contact analysis. The defining parameters are listed in Table 1; the 14-tooth pinion running against a 59-tooth wheel gives a nominal ratio of about 4.21:1, and the 35° spiral angle at mid-face width with a 50 mm face width provides the axial overlap that keeps several teeth in contact.
| Parameter | Value |
|---|---|
| Shaft angle | 90° |
| Normal pressure angle | 20° |
| Addendum coefficient | 0.85 |
| Clearance coefficient | 0.188 |
| Spiral angle at mid-face width | 35° |
| Modification | Radial + tangential |
| Tooth height type | Equal-clearance tapered |
| Number of teeth (pinion / wheel) | 14 / 59 |
| Transverse module | 5 mm |
| Pitch cone angle (pinion) | 10.7131° |
| Face width | 50 mm |
| Tangential modification (pinion) | 0.105 |
| Radial modification (pinion) | 0.3760 |
| Mounting | Overhung (pinion); straddle-mounted (wheel) |
| Accuracy grade | 7-7-6c |
| Material and heat treatment | 20Cr2Ni4, carburised and quenched |
| Flank hardness | 62–64 HRC (pinion); 57–59 HRC (wheel) |
The elastic constants follow from the case-hardened 20Cr2Ni4 steel: Young’s modulus 206.75 GPa, Poisson’s ratio 0.3 and density 7.85 × 10³ kg/m³. Element size in the contact zone was fixed at 0.5 mm after a convergence check in which the number of elements was raised in steps from 0.5 million to 3 million; with each step increasing the count by 50–100%, stress values sampled at three arbitrary flank points changed by less than 5%, so the result is effectively insensitive to further refinement beyond that point. Because the flank surfaces are doubly curved, contact was defined surface-to-surface, with the pinion flank as the contact surface and the wheel flank as the target surface, and the nonlinear contact problem was solved with the augmented Lagrangian method.
Static contact: patch shape and load sharing across teeth
To resolve the contact state of individual teeth during one meshing period, the pinion was rotated to three successive angular positions (a, b, c) and the contact condition of three neighbouring teeth (1, 2, 3) was examined at each position. The results reproduce the expected sequence of a mesh cycle. Tooth 1, entering contact, shows a contact stress that rises as it takes load; tooth 3, leaving contact, shows a stress that falls away; and the three teeth are simultaneously engaged to different degrees. The contact regions on all three flanks are narrow and roughly linear in form, running across the flank rather than concentrating at a point. That distribution is what allows the load to be shared and the transmission to remain smooth.
The static solution also exposes a less favourable detail. In the middle and later positions of the cycle, the flank of the main loaded tooth shows contact along the tip edge. Edge contact of this kind concentrates stress in a region with little material support and is the usual precursor of tip breakage and spalling under high-speed, heavy-duty operation. The practical remedy identified is to grind the tooth flanks after heat treatment, which corrects the tip region geometry and removes the localised edge bearing left by the cutting process.
Dynamic contact: the mesh cycle and load transfer
A static solution cannot represent how load is handed from tooth to tooth when speed and torque fluctuate, nor how mesh stiffness varies periodically as the number of contacting teeth alternates. A dynamic contact model was therefore built on the static one: an angular velocity boundary was applied to the pinion, a revolute joint was defined between pinion and wheel, and the connection representing the mating interface was converted from a body-to-body constraint to a body-ground revolute joint to release rotational freedom while retaining the contact definition.
The equivalent-stress fields obtained through a full cycle show the maximum flank contact stress varying in an approximately periodic manner as teeth mesh in, carry load and mesh out, with the sequence and behaviour summarised in Table 2.
| Phase | Teeth carrying load | Contact-stress behaviour |
|---|---|---|
| (a) Single-tooth mesh | 1 | Reference level for the cycle |
| (b) Double-tooth mesh-in | 2 | Transient peak above the single-tooth level, caused by abrupt load transfer |
| (c) Double-tooth mesh | 2 | About 40% below the single-tooth level as load is shared |
| (d) Double-tooth mesh | 2 | About 40% below the single-tooth level as load is shared |
| (e) Double-tooth mesh-out | 2 | Transient peak above the single-tooth level, caused by abrupt load transfer |
| (f) Single-tooth mesh | 1 | Reference level for the cycle |
Two features dominate. First, the highest stresses occur not in single-tooth meshing as such but at the transitions, where one tooth abruptly takes over the whole load. Second, once two teeth share the torque, the flank contact stress falls by roughly 40% relative to single-tooth loading. This load-sharing margin is the quantitative expression of the multi-tooth contact advantage that motivates the use of spiral bevel gears in heavily loaded reducers, and it also means that the fatigue-critical instants are the short hand-over intervals rather than the steady mesh.
Speed dependence of contact stress
Repeating the dynamic analysis at different pinion speeds, with torque held constant, shows the maximum flank contact stress rising exponentially with rotational speed. At low speed the impact component of the mesh load is small, the stress and strain variation between mesh phases stays narrow, and the transmission is smooth. As speed increases, the same nominal torque produces a markedly larger stress amplitude, and the spread of stress and strain across the successive phases of the cycle widens. The mechanism is the impact loading that accompanies each tooth entry, amplified by the shortened time available for load transfer and by the periodic variation of mesh stiffness.
The design consequence is that a gear sized on static contact stress alone will be under-designed for high-speed service. Speed enters the contact fatigue assessment as an independent variable, not merely as a multiplier on the number of cycles, and the transition intervals identified in Table 2 are the locations where contact fatigue damage is most likely to initiate. Setting the allowable contact stress with reference to the intended operating speed, and treating the mesh-in and mesh-out peaks rather than the average mesh condition as the design case, are the direct routes to a longer service life for this class of gearing.
Design implications
For a 14/59 tooth pair with a 35° spiral angle and 50 mm face width, the analysis supports three conclusions. Multi-tooth contact is present throughout the cycle, with at least three teeth engaged to different degrees, and the resulting linear contact patches distribute load as intended; the associated reduction of about 40% in flank contact stress during double-tooth mesh is a genuine design margin. Tip-edge contact on the main loaded tooth must be removed, most directly by grinding the flanks after carburising and quenching so that the tip relief and the contact pattern are corrected together. Finally, because contact stress grows exponentially with speed, the contact fatigue life of the pinion, rather than its bending strength, is the limiting quantity in high-speed input stages, and the speed-dependent stress amplitude should be carried through the whole structural optimisation rather than checked afterwards.