The main reducer plays a role in reducing speed, increasing torque, and changing the direction of torque rotation in the transmission system of a car. It is composed of one or several pairs of reduction gear pairs, which rely on the gear with fewer teeth to drive the bevel gear with more teeth to achieve deceleration. The bevel gear transmission is used to change the direction of torque rotation. The active bevel gear is an important transmission component in this part of the system, which inputs power and transmits it to the driven gear, Realize the control of the car steering system, adjust the linear speed of the inner and outer wheels, and thus achieve smooth turning of the car wheels.
The machining process of bevel gears is as follows: cutting → hot forging → normalizing (pre heat treatment) → rough machining → precision machining → carburizing and quenching → tempering → precision grinding and forming. Pre heat treatment can eliminate the uneven distribution of austenite structure, eliminate the stress generated during the forging process, reduce the chance of crack occurrence, improve cutting performance, and prepare for the final quenching and tempering heat treatment.
The main reducer gear studied by ZHY Gear experienced abnormal fracture before reaching its normal service life. The design standard for this part requires an effective surface hardening layer depth of 0.8-1.1mm, a surface hardness of 58-64HRC, and a matrix hardness of 32-48HRC. A comprehensive analysis and testing were conducted to provide theoretical guidance for later production.
1. Analysis and testing
1.1 Macro observation
Most of the teeth of the driving bevel gear are broken, and macroscopic cracks have also appeared at the root of the unbroken teeth. Observing the fracture surface, obvious striated crack propagation traces can be found on two of the broken teeth, indicating low cycle fatigue fracture. Figure 1 is a physical image of the driving bevel gear.

1.2 Chemical composition
The chemical composition of the surface layer of the gear teeth was determined using a direct reading spectrometer. The material of the bevel gear is 20CrMnTiH steel, and the chemical element analysis results are shown in Table 1, which meets the requirements of the GB/T3077-2015 Alloy Structural Steel Standard for each element content.
C | Si | Mn | P | S | Cr | Ti | Cu | Ni |
0. 22 | 0. 33 | 0. 65 | 0. 015 | 0. 016 | 1. 06 | 0. 056 | 0. 058 | 0. 029 |
1.3 Hardness test
Samples were taken from the residual tooth surface and the interior of the matrix, and the Rockwell hardness of each part was measured separately. The test results are shown in Table 2, and the matrix detection value is close to the lower limit of the design requirements. The average hardness value on the surface of bevel gears is lower than the design requirements, and the hardness is uneven. The gradient of surface hardness difference values is large, indicating the existence of abnormal structure on the surface. The unevenness of the structure causes the difference in hardness.
Project | 1 | 2 | 3 | 4 | Mean value |
Surface | 53.3 | 58.9 | 59.2 | 54.1 | 56.4 |
Matrix | 32.7 | 32.3 | 33.2 | 32.1 | 32.6 |
1.4 Determination of effective depth of carburized hardening layer on the surface
Take samples at 3 points on the surface and measure the average effective hardening depth of the bevel gear surface using a micro Vickers hardness tester, which is 0.86mm (required to be 0.8-1.1mm), with a hardness limit value of 550HV1. According to the standard (GB/T9450-2005 “Determination and Verification of the Depth of Carburization and Quenching Hardening Layer in Steel Parts”), the measurement and assessment of the depth of carburization and quenching hardening layer in steel parts have exceeded the minimum required value and meet the design requirements.
1.5 Metallographic analysis
A metallographic sample was taken at the fracture site, and after rough grinding, fine grinding, and polishing, non-metallic inclusions (which can damage the continuity of the matrix structure, change the distribution of material force, and generate stress concentration here) were observed under a metallographic microscope. No obvious non-metallic inclusions were found near the fracture and in the matrix, only a small amount of spherical oxides were present, all below level 1, as shown in Figure 2. Therefore, it is not the main cause of bevel gear fracture.

The polished sample was etched with 4% nitric acid alcohol, washed, blown dry, and observed under a microscope. The microstructure of the carburized hardened layer on the surface of the unbroken bevel gear is fine needle martensite+a small amount of residual austenite, as shown in Figure 3. Its microstructure belongs to the product of normal quenching; The microstructure of the center is lath bainite+sorbite (Fig. 4), which will reduce the bending strength of bevel gears. However, considering that it is located in the central position and not the main force bearing part, it has little impact on the fracture of the bevel gear.



There is a mesh like non martensitic structure on the surface of the root of the fractured bevel gear, and its depth was tested to be approximately 0.04mm (Figure 5). The ideal microstructure for carburizing (carbon nitrogen co carburizing) on the surface of quenched parts should be fine needle shaped high carbon martensite. However, due to many uncontrollable factors such as heat treatment and processing technology, a mixed structure of non martensite such as bainite and martensite (pearlite type) is formed on the surface of bevel gears, resulting in serious quality defects. If the depth of non martensitic structure exceeds the standard severely, there will be a phenomenon of low surface hardness on the mechanical properties of the parts, affecting the hardness gradient and resulting in uneven testing hardness.
There is a non martensitic structure at the root of the bevel gear teeth. According to the national automotive industry standard QC/T262-1999 “Metallographic Inspection of Automotive Carburized Gears”, the maximum depth of non martensitic structure on the surface of bevel gears shall not exceed 0.02mm, and the depth of non martensitic structure on the surface of bevel gears shall be 0.04mm, with a network like infiltration along the original austenite grain boundaries. The deeper non martensitic structure severely reduces the surface hardness, wear resistance, and fatigue limit of bevel gears, and initiates fine cracks from grain boundaries or stress concentration areas of oxides, forming crack sources that cause bevel gears to fracture due to insufficient bending strength during later service meshing.
2. Discussion on test results
There is a deep non martensitic structure on the surface of bevel gears, which infiltrates along the original austenite grain boundaries in a mesh like manner. This structure seriously weakens the strength of the bevel gear surface and grain boundaries, reduces wear resistance, and reduces the fatigue life of bevel gears (under the same force, the early initiation of crack sources or the cumulative degree of crack damage will mostly reduce the fatigue coefficient of bevel gears, thereby reducing the fatigue limit life), The presence of non martensitic structure first causes uneven surface hardness of bevel gears, making them prone to stress concentration and fatigue crack sources during service. The continued propagation of multiple crack sources ultimately leads to bevel gear fracture, greatly shortening the fatigue life of bevel gears, which is the main reason for bevel gear failure and fracture.
3. Improvement measures and effects
There are two main ways to solve the source of non martensitic structure: first, when selecting materials, try to minimize the elements that will preferentially oxidize (the order in which different elements are preferentially oxidized is C>Ce>Ba>Mg>Al>Ti>Si>B>V>Nb>Mn>Cr>Cd>Fe>P>Mo>Sn>Ni>As>Cu); The second method is to reduce the oxidizing components of the carburizing atmosphere (such as lowering oxygen partial pressure, etc.) to solve the current problem of bevel gears in China. Choosing the second method is more easily accepted by manufacturers.
The specific measures are:
(1) The formation of non martensitic structure indicates that there is an oxidizing atmosphere in the heat treatment furnace. The cleanliness of the carburizing atmosphere in the furnace should be improved, and the sealing of the heat treatment furnace should be strictly controlled. The exhaust time of the heat treatment furnace can be appropriately extended to make the carburizing atmosphere in the furnace purer.
(2) The overall hardness of the surface of the failed bevel gear is relatively low. Increasing the surface hardness appropriately can improve the contact fatigue strength of the tooth surface. Ensure the cleanliness of the workpiece surface before carburization, and improve the surface hardness and uniformity.
(3) During the carburizing process, due to the high carbon potential at the root of the tooth and the slower cooling rate compared to other parts, non martensitic structures are easily generated at the root of the tooth. It is necessary to increase the quenching and cooling rate of the bevel gear appropriately to reduce or eliminate this defect.
4. Conclusion
Based on the improvement suggestions, adjustments were made to the heat treatment process (including oxidation atmosphere and quenching cooling process), and a new batch of bevel gears were sampled and observed. No obvious network like surface non martensitic structure was found. Thoroughly eliminating the high requirements for material and heat treatment process in this organization, the adjustment of process greatly improves the structural defects on the surface of bevel gears. From the final analysis results and the improved effect, it can be seen that the analysis method in this article is highly effective and can provide certain guidance for the treatment process of bevel gears.