The development and production of a differential double-gear presented our engineering team with a significant and fascinating challenge in the realm of heat treatment, specifically in managing and minimizing heat treatment defects. This component is a critical part of a new transmission system. Its unique structure, featuring a thin-walled, fine-module gear (Gear A) directly adjacent to a thick-walled, medium-module gear (Gear B), creates a severe disparity in mass and section thickness. This inherent asymmetry makes it notoriously difficult to achieve uniformity in case depth and hardness during carburizing and quenching. However, the most formidable challenge, and the primary source of potential heat treatment defects, was the control of geometric distortion, particularly in the fragile Gear A. Successfully producing several hundred of these parts in pilot batches required a fundamental re-examination of our standard practices. Our goal was to suppress distortion-related heat treatment defects to within functional limits, achieving a qualification rate that would support initial production needs.
Technical Analysis and the Root of Distortion
The gear was manufactured from 20CrMnTi steel. The technical specifications required carburizing and quenching on the tooth surfaces of both gears and the inner孔 of Gear A. The target case depth was 0.8-1.2 mm, with a surface hardness of 58-63 HRC and a core hardness of 33-45 HRC. The key dimensional parameters are summarized below:
| Gear Designation | Pitch Diameter (mm) | Module (mm) | Number of Teeth |
|---|---|---|---|
| Gear A (1st/2nd Speed) | φ69.6 | 1.75 | 38 |
| Gear B (Forward) | φ92.3 | 2.25 | 40 |
| Gear B (3rd Speed) | φ95.3 | 2.25 | 42 |
| Gear B (Reverse) | φ98.3 | 2.25 | 43 |
The structural vulnerability is clear. Gear A has a maximum outer diameter of φ69.6mm but a large inner孔 of φ47mm, resulting in a wall thickness of only approximately 11.3mm. This thin section possesses a very low stiffness. In heat treatment analysis, we often consider a distortion sensitivity factor related to geometry. For a ring, a simplified sensitivity coefficient (k) can be related to the ratio of wall thickness (t) to diameter (D).
$$ k \propto \frac{1}{(D/t)} $$
For Gear A, D/t ≈ 69.6 / 11.3 ≈ 6.16, indicating high sensitivity to thermal stress. During rapid heating or cooling, the thin wall can heat up or cool down much faster than the massive Gear B hub, creating immense thermal gradients. The drawing required tight tolerances for Gear A: a radial runout of 0.05mm and a span measurement (over pins) variation within 0.03mm. While Gear B could be finished by grinding, Gear A’s quality had to be assured directly from the quenching process. Initial trials in conventional sealed-quench and pit furnaces, using medium-temperature carburizing and quench oil, resulted in severe and unpredictable distortion, with span changes (ΔW) scattering up to 0.15mm. This confirmed that uncontrolled thermal stress was the primary driver of heat treatment defects in this component.
We concluded that while distortion from heat treatment is inevitable, its magnitude can be managed. The contributing factors are machining stress, transformational (phase change) stress, and thermal stress. Machining stresses can be relieved prior to hardening. The transformational stress from martensite formation, while present, was considered secondary for this high-hardenability steel. The dominant, uncontrolled factor was thermal stress. This arises from:
1. Non-uniform heating/cooling due to section thickness variations.
2. Temperature gradients within the furnace.
3. Non-uniform fluid flow during quenching.
4. Orientation of the part during the quench.
Therefore, the core of our process development strategy was to understand, mitigate, and control the effects of thermal stress to prevent distortion-type heat treatment defects.
Fundamental Process Countermeasures Against Heat Treatment Defects
To systematically attack the problem of heat treatment defects, we implemented a multi-stage plan focusing on stress management and process control.
1. Stress Relief Prior to Carburizing:
A stress-relief anneal (or high-temperature temper) at 650°C was introduced after rough machining. This was critical to eliminate residual stresses from turning and drilling, which could superimpose onto thermal stresses during heating and lead to chaotic distortion.
2. Preheating:
Before loading into the carburizing furnace, parts were preheated at 450°C. This served a dual purpose: it provided a second, final relief for any stresses from finish machining, and it ensured the parts entered the high-temperature furnace uniformly, reducing thermal shock and the associated gradient-driven heat treatment defects.
3. Use of a Quench Plug (Core):
This was a pivotal innovation. A cylindrical steel plug was designed to fit snugly into the large φ47mm孔 of Gear A during the quench. The principle is twofold: First, it reduces the effective surface area of the thin section exposed to the quenchant, slowing its cooling rate. Second, and more importantly, it adds thermal mass to the thin wall, helping to balance the cooling rate between Gear A and the massive Gear B. By reducing the thermal gradient (ΔT) between sections, we directly reduce the thermal stress (σ_th) which is proportional to the gradient, material properties (Young’s modulus E, coefficient of thermal expansion α), and constrained by geometry.
$$ \sigma_{th} \propto E \cdot \alpha \cdot \Delta T $$
A smaller ΔT leads to lower σ_th, directly mitigating the root cause of distortion heat treatment defects.
4. Controlled Atmosphere Furnace Process:
We utilized a multi-chamber controlled atmosphere furnace. A key step was a “pre-carburizing” stage at 850°C. This promotes the formation of fine carbide nuclei in the surface layer, which later facilitates a more diffuse carbide distribution during high-temperature carburizing. This allowed us to slightly reduce the high-temperature exposure time, indirectly limiting grain growth and thermal stress buildup. The precise carburizing-diffusion cycle was critical for repeatability.
5. Hot Oil Quench with Delayed Agitation:
Instead of conventional room-temperature oil, we employed a 120°C hot oil quench. Hot oil reduces the severity of the quench (lower Heat Transfer Coefficient, HTC), lowering the cooling rate, especially in the martensite transformation range (Ms-Mf). This reduces both thermal and transformational stresses. For fine-module gears like our Gear A, the small tooth gaps can “trap” oil, leading to non-uniform cooling. To combat this, we implemented a delayed, slow-speed agitation protocol. The agitators were activated only after a brief, still-oil period. The controlled, slow flow ensured reproducible heat extraction conditions from batch to batch, which is essential for controlling the variability of heat treatment defects. The cooling rate (dT/dt) in oil can be modeled in the Newtonian regime as:
$$ \frac{dT}{dt} = -\frac{hA}{mc} (T – T_{oil}) $$
where h is the HTC, A is surface area, m is mass, c is specific heat, and T_oil is oil temperature. By using hot oil and controlled agitation, we effectively managed the ‘h’ and ‘(T – T_oil)’ terms to achieve a more forgiving quench.
6. Fixturing and Orientation:
The parts were loaded vertically on a tree-type fixture, with the Gear A (with plug installed) facing upward. This orientation ensured symmetrical fluid flow around the critical Gear B and allowed for consistent entry into the oil. Consistent fixturing is often an overlooked but vital factor in minimizing random heat treatment defects.
Pilot Production Execution and Results Analysis
The pilot production was conducted over five furnace charges, totaling over a hundred parts. Each batch was an opportunity to learn and refine.
First Batch (Exploratory):
A small batch of parts was run to validate initial gear cutting data and the baseline process. Parts were hung vertically. Post-treatment metallography, case depth, and hardness met specifications, as shown in Table 2. However, measurement of Gear A revealed significant distortion: span measurements (over pins) had grown beyond the upper specification limit, and radial runout was excessive. The average span growth (ΔW) was +0.05mm. This confirmed the process was not yet capable of controlling heat treatment defects.
| Batch | Carbide Structure | Retained Austenite | Martensite | Core Ferrite | Case Depth (mm) | Surface Hardness (HRC) |
|---|---|---|---|---|---|---|
| 1 | Fine, Dispersed | Level 2 | Level 2 | None | 1.0-1.1 | 60-62 |
| 2 | Fine, Dispersed | Level 2 | Level 2 | None | 1.0-1.1 | 60-62 |
| 3,4,5 | Fine, Dispersed | Level 2 | Level 2 | None | 1.0-1.1 | 60-62 |
Second Batch (Implemented Countermeasures):
Based on the first run, gear cutting data was adjusted to pre-compensate for expected growth. The quench plug was used (solid version), and parts were loaded with Gear A (plugged) facing up. The process parameters were now fixed. Results showed acceptable metallurgy (Table 2, Batch 2). Distortion data (Table 3) for Gear A showed improvement, but some parts still exhibited excessive radial runout. We hypothesized that the solid plug, while adding mass, might be creating a vapor pocket or impeding oil flow within the孔, causing asymmetric cooling—a classic source of quenching heat treatment defects.
| Part ID | Radial Runout (mm) Pre-HT | Radial Runout (mm) Post-HT | Span (mm) Pre-HT | Span (mm) Post-HT | ΔSpan (mm) |
|---|---|---|---|---|---|
| 2-01 | 0.02 | 0.05 | 21.985 | 22.032 | +0.047 |
| 2-02 | 0.01 | 0.07 | 21.988 | 22.038 | +0.050 |
Third through Fifth Batches (Optimized Plug):
The key modification was drilling a φ5mm axial hole through the center of the steel quench plug. This simple change allowed quench oil to flow through the孔, ensuring more uniform heat extraction from the inner wall of Gear A and preventing vapor lock. All other parameters remained constant. The results were markedly consistent. Metallurgical quality remained excellent (Table 2, Batches 3-5). The distortion data, as sampled and shown in Table 4, demonstrated a dramatic improvement in control. The spread of span measurements (ΔW) was tightly controlled, and crucially, radial runout was consistently brought within the 0.05mm specification. The limit of variation for critical dimensions was now within the acceptable range for assembly. This consistency proved the process was stable and capable of mass-producing these parts with minimal, predictable heat treatment defects.
| Batch | Part ID | Radial Runout (mm) | Span Measurement (mm) | Limit Variation* |
|---|---|---|---|
| 3 | 01 | 0.04 | 22.028 | Within Spec |
| 3 | 02 | 0.03 | 22.025 | Within Spec |
| 4 | 01 | 0.04 | 22.027 | Within Spec |
| 4 | 02 | 0.02 | 22.024 | Within Spec |
| 5 | 01 | 0.03 | 22.026 | Within Spec |
| 5 | 02 | 0.04 | 22.029 | Within Spec |
*A composite assessment of size, roundness, and taper against drawing limits.
Conclusion and Reflections on Managing Heat Treatment Defects
This project underscored that complex, asymmetrical components do not necessarily mandate complex, capital-intensive solutions like press quenching. By deeply analyzing the primary driver of heat treatment defects—in this case, thermal stress—and implementing a series of targeted, low-cost countermeasures, we achieved a technically and economically viable process. The synergistic combination of stress-relief cycles, thermal mass balancing via a vented quench plug, a controlled hot-oil quench regimen, and consistent fixturing proved effective. The process demonstrated high repeatability, with合格 rates supporting pilot production needs.
Looking forward, two key learnings emerged for preventing such heat treatment defects in future projects. First, material consistency is paramount. For thin-walled, distortion-sensitive parts, the use of guaranteed-hardenability (H-band) steel is highly recommended. Variations in hardenability or segregations within the material can lead to unpredictable transformation stresses and distortion, undermining even the most robust thermal stress controls. Second, this case study reinforces that process innovation often lies in simple, physics-based modifications—like drilling a hole in a plug—rather than solely in new equipment. A fundamental understanding of the heat transfer and stress states during quenching is the most powerful tool for defeating distortion and other geometry-related heat treatment defects.