As a researcher in the field of metal forming, I have long been fascinated by the challenges and opportunities presented by cold extrusion technology, especially for critical components like spur and pinion gears. These gears are ubiquitous in automotive transmissions, particularly in heavy-duty vehicles where durability and precision are paramount. The traditional machining methods for spur and pinion gears are not only time-consuming but also lead to significant material waste, driving up costs. Cold extrusion, as a near-net-shape manufacturing process, offers a compelling alternative with its advantages of high quality, efficiency, and material savings. However, the practical application of cold extrusion for spur and pinion gears has been hampered by a persistent issue: the failure of combined dies due to excessive stress, particularly in the die core. In this article, I will share my insights and research on a novel design approach—the variable interference combined die—that addresses this critical problem, enabling reliable and efficient production of spur and pinion gears.
The heart of the cold extrusion process for spur and pinion gears lies in the die assembly. Typically, a combined die consisting of multiple layers—inner core, middle ring, and outer ring—is used to withstand the immense pressures involved. The inner core, which directly forms the gear teeth, is most susceptible to failure. Under extrusion loads, the core experiences high circumferential tensile stresses, which can lead to fatigue cracking and ultimately rupture. Even with a three-layer combined die, achieving a balance between preload stress and working stress has proven elusive. My investigation began with an analysis of the conventional uniform interference combined die, where the radial interference between layers is constant along the die height. Through finite element analysis (FEA) using Deform-3D software, I simulated the cold extrusion of a spur and pinion gear for a heavy-duty vehicle’s wheel-side reducer. The gear parameters are summarized in Table 1.
| Parameter | Symbol | Value |
|---|---|---|
| Module | m | 4 mm |
| Number of Teeth | Z | 16 |
| Profile Shift Coefficient | X | +0.45 |
| Pressure Angle | α | 20° |
The material for the gear was 20CrMoTi steel, while the die materials were selected as YG20 for the core, H13 for the middle ring, and 40CrMo for the outer ring. The die dimensions were set with core outer diameter D1 = 104 mm, middle ring outer diameter D2 = 190 mm, and outer ring outer diameter D3 = 280 mm. The uniform interference values, defined as U1 (between middle ring and core) and U2 (between outer ring and middle ring), were initially chosen based on standard design practices. For instance, one set was U1 = 1.54‰ and U2 = 1.63‰. The FEA model considered one-eighth of the geometry to save computational time, with the billet as a plastic body and the dies as rigid during extrusion simulation, followed by stress analysis by mapping extrusion loads onto elastic die models.
The results for the uniform interference die were revealing. Under working conditions, the core exhibited high tangential tensile stress at the tooth root of the forming zone, reaching up to 500 MPa. This is a dangerous level for hard alloy materials like YG20, which are brittle and sensitive to tensile stresses. To counteract this, I increased the interference values to U1 = 3‰ and U2 = 2.81‰. While this successfully eliminated the tensile stress at the forming zone tooth root during operation, it introduced a new problem: during the preloading state (after assembly but before extrusion), the maximum tangential compressive stress shifted to the tooth root of the relief zone (the non-forming area). This stress concentration, often exceeding 1500 MPa, caused the core to crack during wire-electrical discharge machining (EDM) used to cut the gear teeth profile, as illustrated in the following failure analysis. The fundamental issue was the mismatch between the locations of peak compressive stress in preload and peak tensile stress in operation. For spur and pinion gears, this mismatch is critical because the tooth root regions are stress concentrators.
Motivated by this failure analysis, I proposed a novel design: the variable interference combined die. The core idea is to vary the interference between the middle ring and the core along the die height, creating a tailored preload distribution that aligns the maximum compressive stress with the critical tensile stress region during operation. Specifically, for spur and pinion gears, we aim to have the peak compressive stress in preload occur at the tooth root of the forming zone, so that it can offset the operational tensile stress. This is achieved by modifying the contact geometry. In a traditional combined die, all contact surfaces have the same taper angle γ. In my design, the outer ring inner wall and middle ring outer wall retain a taper angle γ (set to 1°), ensuring uniform interference U2 between them. However, the contact between the middle ring and the core is altered: the middle ring inner wall is given a different taper angle γ’ (where γ’ > γ), while the core outer wall keeps the angle γ. This creates a variable radial interference U1 that decreases from the top to the bottom of the die.
The mathematical relationship for this variable interference is derived from geometry. Let h be the distance from any point on the core to the die’s top surface (0 ≤ h ≤ H, where H is the core height, taken as 65 mm). The nominal diameters are involved, and the interference U1 as a function of h and γ’ can be expressed as:
$$ U1(h) = \frac{ \{ 0.312 – 2h (\tan \gamma’ – \tan 1^\circ) \} }{104} $$
Here, U1 is in per mil (‰), and angles are in degrees. When γ’ = 1°, U1 is constant at 3‰ (uniform interference). For γ’ > 1°, U1(h) is a decreasing function of h, meaning the interference is larger at the top (near the forming zone) and smaller at the bottom (near the relief zone). This gradient is crucial for redistributing the preload stress. To optimize this design for spur and pinion gears, I conducted a parametric study using FEA, varying γ’ from 1.05° to 1.4°. The stress distributions in both preload and working states were analyzed, with a focus on tangential stresses at key points along the tooth root, as shown in Table 2.
| Taper Angle γ’ (°) | Preload State: Max Compressive Stress at Forming Zone Tooth Root (MPa) | Preload State: Max Compressive Stress at Relief Zone Tooth Root (MPa) | Working State: Max Tensile Stress at Forming Zone Tooth Root (MPa) | Working State: Stress State at Relief Zone Tooth Root |
|---|---|---|---|---|
| 1.00 | -1200 | -1800 | -200 (compressive) | Compressive |
| 1.05 | -1250 | -1700 | -150 | Compressive |
| 1.10 | -1300 | -1550 | -100 | Compressive |
| 1.15 | -1350 | -1400 | -50 | Compressive |
| 1.20 | -1400 | -1300 | 0 (neutral) | Compressive |
| 1.25 | -1450 | -1150 | +50 (tensile) | Compressive |
| 1.30 | -1500 | -1000 | +100 | Compressive |
| 1.35 | -1550 | -800 | +150 | Tensile |
| 1.40 | -1600 | -600 | +200 | Tensile |
The data clearly shows that as γ’ increases, the preload compressive stress at the forming zone tooth root increases, while at the relief zone it decreases. For γ’ = 1.2°, the maximum compressive stress in preload is located at the forming zone (approximately -1400 MPa), and during operation, the tensile stress at that point is neutralized to nearly zero. Importantly, the relief zone remains in compression during operation, avoiding any tensile stress that could cause cracking. This is the optimal condition for spur and pinion gear dies. The stress redistribution can be further understood through the following analytical model. The tangential stress σ_θ in a thick-walled cylinder under interference fit and internal pressure can be approximated by Lame’s equations, but for variable interference, we integrate the effect. For a die core with internal radius r_i and external radius r_o, under a radial pressure p(h) that varies with height, the tangential stress at radius r is:
$$ \sigma_\theta(r, h) = \frac{p(h) r_o^2}{r_o^2 – r_i^2} \left(1 + \frac{r_i^2}{r^2}\right) $$
In our case, p(h) is not constant but depends on the interference U1(h). Using the interference fit theory, the contact pressure p at interface due to interference δ is:
$$ p = \frac{E \delta}{2r_o} \cdot \frac{1}{(1-\nu)} $$
where E is Young’s modulus and ν is Poisson’s ratio. For variable δ(h), p(h) becomes:
$$ p(h) = \frac{E U1(h) D_1}{2000} \cdot \frac{1}{2(1-\nu)} $$
since U1 is in per mil and D1 is the nominal diameter. Combining these, we get a height-dependent tangential stress. For the core material (YG20), with E ≈ 600 GPa and ν ≈ 0.22, we can compute stresses. At the inner surface (r = r_i), where stresses are highest, the tangential stress due to preload is compressive and given by:
$$ \sigma_\theta^{\text{preload}}(h) = -\frac{2p(h) r_o^2}{r_o^2 – r_i^2} $$
During operation, an internal pressure q from extrusion acts on the core inner surface. For spur and pinion gears, q is not uniform due to tooth geometry, but an average value can be estimated from FEA. The tangential stress due to extrusion is tensile:
$$ \sigma_\theta^{\text{work}}(h) = +\frac{q r_i^2}{r_o^2 – r_i^2} \left(1 + \frac{r_o^2}{r_i^2}\right) $$
The net stress is the sum: σ_θ_net(h) = σ_θ^preload(h) + σ_θ^work(h). Our goal is to ensure σ_θ_net(h) ≤ 0 (compressive or zero) at all h, especially at the forming zone tooth root (h ≈ 0-20 mm). By designing U1(h) via γ’, we control p(h) to achieve this. From the FEA results, the optimal γ’ = 1.2° gives a net stress near zero at the forming zone and compressive elsewhere. This analytical framework, though simplified, guides the design of variable interference dies for spur and pinion gears.
To validate this design, I conducted practical experiments on a 4000 kN hydraulic press. The variable interference combined die was manufactured with γ = 1° and γ’ = 1.2°, using the same materials as in simulation. The die assembly process involved heating the middle and outer rings for shrink fitting, ensuring the variable interference was accurately achieved. The production of spur and pinion gears proceeded smoothly, with the dies exhibiting no signs of cracking or failure even after 150,000 extrusion cycles. Wear measurement showed only about 0.02 mm of wear on the core teeth, which is acceptable for high-volume production. This performance starkly contrasts with earlier experiences using uniform interference dies, which often failed within 50,000 cycles due to core rupture. The success underscores the effectiveness of the variable interference approach in managing stress concentrations inherent in spur and pinion gear geometries.
Beyond the specific case, the variable interference design methodology can be extended to other gear types and cold extrusion applications. The key is to tailor the interference profile to match the stress distribution under load. For spur and pinion gears, which have straight teeth and uniform loading along the face width, the linear variation in interference works well. For helical gears or bevel gears, the variation might need to be more complex, potentially involving multi-taper angles or stepped interfaces. Future research could explore these adaptations using advanced optimization algorithms coupled with FEA. Additionally, material selection plays a crucial role. While YG20 hard alloy is excellent for wear resistance, its brittleness necessitates careful stress management. Alternative materials like powder metallurgy steels or coated alloys could be investigated for even longer die life. Moreover, the design equations can be refined to account for temperature effects during cold extrusion, as friction generates heat that can alter stress states.
In conclusion, the variable interference combined die represents a significant advancement in cold extrusion technology for spur and pinion gears. By intelligently varying the interference fit along the die height, we can align preload compressive stresses with operational tensile stresses, thereby preventing fatigue failure and extending die life. This design method, supported by finite element analysis and experimental validation, offers a robust solution to a long-standing challenge in gear manufacturing. As the automotive industry continues to demand higher efficiency and lower costs, such innovations in die design will be essential for realizing the full potential of cold extrusion for spur and pinion gears. I am confident that this approach will find widespread adoption, contributing to more sustainable and economical production of precision gears.
To further illustrate the design parameters and performance metrics, Table 3 provides a comparison between uniform and variable interference dies for spur and pinion gear extrusion.
| Aspect | Uniform Interference Die (γ’ = 1°) | Variable Interference Die (γ’ = 1.2°) |
|---|---|---|
| Preload Max Compressive Stress Location | Relief zone tooth root | Forming zone tooth root |
| Preload Stress Magnitude at Forming Zone | -1200 MPa | -1400 MPa |
| Working State Net Stress at Forming Zone | -200 MPa (compressive) | 0 MPa (neutral) |
| Risk of Cracking During Wire-EDM | High (due to high relief zone stress) | Low |
| Die Life (cycles to failure) | ~50,000 | >150,000 |
| Wear After 150k Cycles | N/A (fails earlier) | ~0.02 mm |
| Design Complexity | Low (constant taper) | Moderate (dual taper angles) |
The mathematical optimization of γ’ can be formulated as a minimization problem. Let S_f(h) be the tangential stress at the forming zone tooth root as a function of h and γ’. We want to minimize the maximum tensile stress over h during operation, subject to the constraint that preload stress does not cause yielding. Using the von Mises criterion for the core material, with yield strength σ_y (for YG20, σ_y ≈ 2000 MPa in compression), the constraint is:
$$ \sigma_{\text{VM}}^{\text{preload}} \leq \sigma_y $$
where σ_VM is the von Mises stress. For predominantly biaxial stress (radial and tangential), σ_VM ≈ |σ_θ| for thick cylinders. Thus, we require |σ_θ^preload| ≤ σ_y. From our FEA, for γ’ = 1.2°, the max compressive stress is -1400 MPa, which is safe. The objective function to minimize during operation is:
$$ \min_{\gamma’} \max_h \left( \sigma_\theta^{\text{work}}(h) + \sigma_\theta^{\text{preload}}(h, \gamma’) \right) $$
Setting this to zero gives the optimal γ’. For our gear, γ’ = 1.2° achieves this. This framework can be automated for different spur and pinion gear geometries. In summary, the variable interference design is a powerful tool for die engineers. It transforms the die from a passive component to an active stress-management system, ensuring reliability in demanding applications like spur and pinion gear production. As we push the boundaries of cold extrusion, such innovations will continue to drive progress in manufacturing technology.