In modern mechanical transmission systems, spur and pinion gears are fundamental components due to their efficiency and reliability. The involute tooth profile is widely adopted for spur and pinion gears because of its smooth transmission and low noise characteristics. However, spur gears, especially in spur and pinion gear pairs, exhibit limitations such as end-face contact, low overlap ratio, reduced load capacity, and sensitivity to manufacturing errors. Therefore, precise error control is critical for ensuring performance. Traditional contact-based measurement methods, like gear rolling tests, suffer from drawbacks including long detection times, low universality, and potential wear on both the measuring probe and gear surface. To address these issues, non-contact measurement techniques based on laser displacement sensors have emerged as a promising alternative. This paper presents a comprehensive study on a laser-based non-contact measurement method for involute spur and pinion gears, focusing on error correction to enhance accuracy. We develop an error compensation model through systematic experiments and apply it to gear profile inspection, demonstrating significant improvements over uncorrected measurements.
The core principle of our approach relies on laser triangulation, where a laser beam is projected onto the gear tooth surface, and the reflected light is captured by a CCD array to determine displacement changes. For spur and pinion gears, this method offers advantages like rapid measurement, avoidance of surface damage, and suitability for mass inspection. However, inherent errors arise due to factors such as incidence angle and measurement depth, necessitating calibration and correction. We conduct error calibration experiments by varying incidence angles and depths, then establish a compensation model using MATLAB. This model is integrated into the measurement process for spur and pinion gears, enabling accurate evaluation of key parameters like tooth profile deviation and pitch error. Our results validate the feasibility of this non-contact method for high-precision gear inspection, with applications in quality control for spur and pinion gear manufacturing.
The experimental setup involves a four-coordinate measuring center developed in collaboration with industry partners, comprising linear axes (X, Y, Z) and a rotary axis (C). A laser displacement sensor is mounted on the X-axis, while the spur or pinion gear is fixed on the C-axis for rotation. The sensor uses a red semiconductor laser (650 nm) with a spot diameter of 50 μm, a measurement range of ±10 mm, and high repeatability of 0.025 μm. Key parameters are summarized in Table 1. The system employs Renishaw gratings for feedback, achieving positioning accuracies of 1 μm for linear axes and 0.001° for the rotary axis. This configuration allows precise trajectory planning for scanning involute profiles of spur and pinion gears, mimicking theoretical engagement paths.
| Parameter | Value |
|---|---|
| Model | LK-H050 |
| Mounting Type | Diffuse Reflection |
| Reference Distance | 50 mm |
| Measurement Range | ±10 mm |
| Spot Diameter | 50 μm |
| Linearity | ±0.02% F.S. |
| Repeatability | 0.025 μm |
| Temperature Drift | 0.01% F.S./°C |
Laser triangulation operates on the principle where a laser beam strikes the surface of a spur or pinion gear tooth, and the reflected light forms an image on a CCD. The relationship between the object’s axial displacement and the image position is given by the triangulation equation. For a simplified model, consider the laser incident at an angle $$ \theta $$ relative to the surface normal, and the reflected beam captured at an angle $$ \phi $$ on the CCD. The displacement $$ d $$ can be expressed as:
$$ d = \frac{L \cdot \Delta x}{f \cdot \sin(\theta + \phi)} $$
where $$ L $$ is the baseline distance between laser and CCD, $$ \Delta x $$ is the pixel shift on the CCD, and $$ f $$ is the focal length. In practice, for spur and pinion gears, the incidence angle varies along the involute curve, leading to measurement errors that must be corrected. We denote the theoretical laser value as $$ z_t $$ and the measured value as $$ z_m $$, with the error $$ \delta = z_m – z_t $$ being a function of incidence angle $$ \alpha $$ and measurement depth $$ h $$. Thus, $$ \delta = f(\alpha, h) $$, which we characterize through calibration experiments.
Error calibration is performed using a control variable approach. A sine bar and gauge blocks are arranged on a rotary table to simulate different surface inclinations for spur and pinion gears. The laser displacement sensor is mounted on a CNC Z-axis, and a laser interferometer provides reference measurements. By varying the incidence angle from -45° to 45° (covering typical involute profiles) and measurement depths across the sensor range, we collect error data. For instance, at angles of 10°, 20°, 30°, and 40°, and depths of -8 mm, -4 mm, 0 mm, 4 mm, and 8 mm, errors are recorded as shown in Table 2. The data reveal that error magnitude increases with both angle and depth, emphasizing the need for compensation in spur and pinion gear measurements.
| Incidence Angle (°) | Measurement Depth (mm) | Laser Sensor Reading (mm) | Interferometer Reference (mm) | Error (mm) |
|---|---|---|---|---|
| 10 | 8 | 8.012 | 8.000 | 0.012 |
| 10 | 4 | 4.008 | 4.000 | 0.008 |
| 20 | 8 | 8.025 | 8.000 | 0.025 |
| 20 | 4 | 4.015 | 4.000 | 0.015 |
| 30 | 8 | 8.045 | 8.000 | 0.045 |
| 30 | 4 | 4.022 | 4.000 | 0.022 |
| 40 | 8 | 8.062 | 8.000 | 0.062 |
| 40 | 4 | 4.030 | 4.000 | 0.030 |
To model these errors, we use MATLAB interpolation algorithms. The error function $$ \delta(\alpha, h) $$ is approximated from experimental data points via piecewise polynomial interpolation. For computational efficiency, we define a lookup table for angles between 20° and 45° and depths from -10 mm to 10 mm, as these ranges encompass typical conditions for spur and pinion gear inspection. The compensated laser value $$ z_c $$ is then:
$$ z_c = z_m – \delta(\alpha, h) $$
In practice, for each measurement point on a spur or pinion gear tooth, we compute the theoretical surface normal vector from the involute equation. The involute profile of a spur gear with base radius $$ r_b $$ is parameterized by the roll angle $$ \varphi $$:
$$ x = r_b (\cos(\varphi) + \varphi \sin(\varphi)) $$
$$ y = r_b (\sin(\varphi) – \varphi \cos(\varphi)) $$
The unit normal vector $$ \mathbf{n} $$ at a point is derived from the derivative:
$$ \mathbf{n} = \left( -\sin(\varphi), \cos(\varphi) \right) $$
The incidence angle $$ \alpha $$ between the laser beam (assumed vertical along the Z-axis) and the normal is calculated using the dot product. If the laser direction is $$ \mathbf{l} = (0, 0, -1) $$, then:
$$ \alpha = \cos^{-1} \left( \frac{|\mathbf{n} \cdot \mathbf{l}|}{\|\mathbf{n}\| \|\mathbf{l}\|} \right) $$
This angle, along with the measured depth $$ h $$ (from sensor reading), is input into the error model to obtain $$ \delta $$. For spur and pinion gears, this process is repeated across all sampled points on the tooth flank. We implement this in MATLAB to simulate gear rotation and compute angles dynamically, as shown in Figure 1 (simulated trajectory and normals). The compensation significantly reduces systematic errors, enhancing measurement fidelity for spur and pinion gear applications.
The tooth profile measurement experiment begins by securing a spur or pinion gear on the rotary table and aligning the laser sensor. The gear is rotated incrementally while laser readings are taken at a rate of 100 points per second (10 ms period). Approximately 95 valid data points per tooth are extracted for analysis. After error compensation, we evaluate key deviations: tooth profile form error and pitch error. For a spur gear with module 2 mm, pressure angle 20°, and 17 teeth, the theoretical involute is generated, and deviations are computed. Table 3 presents a subset of tooth profile deviations after compensation, showing values within micrometers. The total profile deviation $$ F_\alpha $$ is the sum of maximum positive and negative deviations from the theoretical curve. In our case, $$ F_\alpha = 8.9 \, \mu \text{m} $$, which according to ISO standards (e.g., GB/T 10095.1-2008) corresponds to Grade 6 accuracy for spur and pinion gears.
| Data Point Index | Tooth Profile Deviation (μm) |
|---|---|
| 1 | 1.1 |
| 2 | 4.7 |
| 3 | 4.8 |
| 4 | 6.1 |
| 5 | 7.6 |
| 6 | 7.8 |
| 7 | 8.7 |
| 8 | 6.3 |
| 9 | 6.1 |
| 10 | 5.6 |
Pitch error analysis involves computing single pitch deviation $$ f_{pt} $$ and cumulative pitch deviation $$ F_p $$. Using MATLAB, we fit cubic splines to the compensated data points for each tooth of the spur or pinion gear. The intersection of the fitted curve with the pitch circle (radius $$ r_p = m \cdot z / 2 $$, where $$ m $$ is module and $$ z $$ is tooth count) yields actual pitch points. The distance between consecutive points gives actual pitch, compared to the theoretical pitch $$ p_t = \pi m $$. Single pitch deviation is the algebraic difference, and cumulative deviation is the sum over all teeth. Table 4 summarizes results for our spur gear sample, where $$ f_{pt,\text{max}} = 8.2 \, \mu \text{m} $$ and $$ F_p = 27.3 \, \mu \text{m} $$, both within Grade 6 limits. This demonstrates the efficacy of our error correction method for spur and pinion gear inspection.
| Tooth Number | Single Pitch Deviation (μm) | Cumulative Pitch Deviation (μm) |
|---|---|---|
| 1 | -4.3 | -4.3 |
| 2 | -4.1 | -8.4 |
| 3 | -1.9 | -10.3 |
| 4 | -2.7 | -13.0 |
| 5 | -6.7 | -19.6 |
| 6 | -3.2 | -22.8 |
| 7 | -4.5 | -27.3 |
| 8 | 3.5 | -23.8 |
| 9 | 7.7 | -16.1 |
| 10 | 8.2 | -7.9 |
| 11 | 7.5 | -0.4 |
| 12 | 5.1 | 4.7 |
| 13 | 4.5 | 9.2 |
| 14 | 3.4 | 12.6 |
| 15 | -4.9 | 7.7 |
| 16 | -3.8 | 3.9 |
| 17 | -3.6 | 0.3 |
To further validate our approach, we compare results with contact-based measurements for the same spur and pinion gear. Contact methods using coordinate measuring machines (CMM) typically yield $$ f_{pt,\text{max}} = 5.2 \, \mu \text{m} $$ and $$ F_p = 22.8 \, \mu \text{m} $$ for the right flank, indicating higher precision due to direct probing. However, our laser-based method after correction achieves $$ f_{pt,\text{max}} = 8.2 \, \mu \text{m} $$ and $$ F_p = 27.3 \, \mu \text{m} $$, which is sufficiently accurate for Grade 6 spur and pinion gears. The minor discrepancy stems from residual errors in laser triangulation, such as surface reflectivity variations or environmental factors. Nevertheless, the non-contact advantage—avoiding wear and enabling fast inspection—makes it viable for industrial applications involving spur and pinion gears.
The error correction model can be extended to other gear types, such as helical or bevel gears, by adjusting the incidence angle calculations. For spur and pinion gears, the involute geometry simplifies the model, but similar principles apply. We also investigate the impact of sensor alignment errors. Misalignment between the laser beam and gear axis introduces additional errors, which can be modeled as a transformation matrix. If the sensor is tilted by small angles $$ \Delta \theta_x $$ and $$ \Delta \theta_y $$, the effective incidence angle changes. Compensating for this requires initial calibration using master gears, a common practice for spur and pinion gear metrology.
In conclusion, we have developed a robust laser-based non-contact measurement system for involute spur and pinion gears, incorporating an error correction model that accounts for incidence angle and depth effects. Through calibration experiments and MATLAB-based modeling, we significantly improve measurement accuracy, achieving Grade 6 precision for tooth profile and pitch deviations. This method addresses limitations of contact techniques, offering rapid, wear-free inspection suitable for mass production of spur and pinion gears. Future work will focus on real-time compensation algorithms and integration with CAD/CAM systems for closed-loop manufacturing of spur and pinion gears. The principles outlined here underscore the potential of optical metrology in advancing gear inspection technology for spur and pinion gear applications across industries.