In the realm of mechanical transmission systems, the cylindrical gear stands as a fundamental and widely utilized component. Among various gear types, the involute spur gear, characterized by its smooth transmission and low noise, has become predominant due to its reliability. However, spur gears exhibit limitations such as only end-face contact, low overlap ratio, reduced load-bearing capacity, and sensitivity to manufacturing errors. Consequently, precise error control is paramount to ensure the stability and longevity of gear-driven systems. Traditional contact-based measurement methods, while accurate, suffer from drawbacks including prolonged inspection times, low universality, and potential wear on both the measuring probe and the gear surface. To address these issues, non-contact measurement techniques, particularly those leveraging laser displacement sensors, have garnered significant attention. This article explores a non-contact measurement methodology based on laser displacement sensing for evaluating the tooth profile of involute spur gears, with a focus on error correction to enhance measurement accuracy.
The evolution of gear measurement technology has progressed from mechanical methods to integrated information systems. Modern gear measurement can be categorized into several approaches, with gear integrated error measurement technology representing a current trend. This integration aims to merge measurement data with computer-aided design and manufacturing (CAD/CAM) systems, fostering advanced closed-loop gear manufacturing processes. Instrumentation has similarly advanced, with CNC gear measuring centers emerging as sophisticated electromechanical systems capable of versatile measurements through numerical control rather than purely mechanical linkages. Despite these advancements, contact methods like gear meshing roll tests remain common but are inherently slow and prone to inducing surface damage. In contrast, laser-based non-contact methods offer rapid, wear-free inspection suitable for high-volume production environments.
Laser ranging technology, especially the laser triangulation principle, has found extensive application in modern metrology, including online surface roughness measurement and dimensional inspection. A laser displacement sensor operates by projecting a focused laser spot onto the target surface. The reflected light is captured by a linear CCD array, and the displacement of the imaged spot correlates uniquely with the depth variation of the target point along the sensor’s axis. This principle allows for high-resolution, non-contact measurement of displacement changes. The technique is material-agnostic, capable of measuring metals, non-metals, and even delicate surfaces without contamination or wear, making it ideal for precision gear inspection.
The experimental setup for this study utilized a four-coordinate measuring center developed in collaboration with a machine tool manufacturer. This apparatus comprises three linear axes (X, Y, Z) and one rotary axis (C). A laser displacement sensor is mounted at the terminus of the X-axis, while the cylindrical gear specimen is secured between centers on the C-axis, allowing full 360-degree rotation. Each axis is equipped with Renishaw linear encoders for feedback, achieving positioning accuracies of 1 μm for the linear axes and 0.001° for the rotary axis. The core measurement device is a Keyence LK-H050 laser displacement sensor, which employs a red semiconductor laser (650 nm wavelength) and a high-density CCD to achieve superior precision. Its key parameters are summarized in Table 1.
| Parameter | Value |
|---|---|
| Measurement Principle | Diffuse Reflection, Laser Triangulation |
| Reference Distance | 50 mm |
| Measurement Range | ±10 mm |
| Beam Spot Diameter | 50 μm |
| Linearity | ±0.02% of Full Scale (F.S.) |
| Repeatability | 0.025 μm |
| Temperature Coefficient | 0.01% F.S./°C |
The laser triangulation principle can be mathematically described. Let the laser beam be incident along the sensor’s optical axis. When the target surface is at the reference position, the laser spot forms an image at a nominal position on the CCD. If the surface moves by a distance $$ \Delta z $$ along the sensor axis, the image displacement on the CCD, denoted $$ \Delta x $$, is related by the triangulation geometry. For a setup with laser incidence angle $$ \theta $$ and receiving lens angle $$ \phi $$, the relationship is approximated by:
$$ \Delta z = \frac{L \cdot \Delta x}{f \cdot \sin(\theta + \phi)} $$
where $$ L $$ is the baseline distance between the laser projector and the lens, and $$ f $$ is the focal length of the receiving lens. However, in practical sensors, this relationship is calibrated and often linearized over the measurement range. A critical factor affecting accuracy is the inclination angle between the laser beam and the surface normal. As this angle deviates from zero, systematic errors emerge due to changes in the reflection path and spot deformation.
To quantify and correct these errors, a comprehensive error calibration experiment was conducted. The setup, as depicted in the methodology, involved a数控加工中心 (CNC machining center) where the laser sensor was mounted on the Z-axis. A precision sine bar placed on a rotary table (indexing head) beneath the sensor allowed for precise adjustment of surface inclination. A laser interferometer, fixed via magnetic bases on the Z-axis and worktable, served as the reference standard for displacement. The procedure began by aligning the laser beam to be vertically incident (0° inclination). Then, for a series of controlled inclination angles (e.g., 10°, 20°, 30°, 40°), the Z-axis moved the sensor through its full measurement range (-10 mm to +10 mm). At discrete depth intervals (e.g., every 1 mm), readings from both the laser sensor and the interferometer were recorded. The difference between these readings constitutes the sensor’s error at that specific depth and inclination. The control variable method ensured systematic data acquisition.
The error data revealed significant trends. For a fixed inclination, the error varied with measurement depth. More importantly, as the inclination angle increased, the magnitude of the error grew substantially. Furthermore, the influence of depth on error became more pronounced at larger angles. This is critical for cylindrical gear measurement because during a tooth profile scan, the laser incidence angle on the involute flank changes continuously. From the planned measurement path for a standard involute cylindrical gear, the theoretical incidence angles (relative to the surface normal) typically range from approximately 23° to 43°. Therefore, error compensation within this angular domain is essential.
A subset of the experimental error data is presented in Table 2, illustrating the error values at different depths for a 30° inclination.
| Nominal Depth (mm) | Laser Sensor Reading (mm) | Interferometer Reference (mm) | Error (mm) |
|---|---|---|---|
| -8.0 | -7.942 | -8.000 | +0.058 |
| -4.0 | -3.945 | -4.000 | +0.055 |
| 0.0 | 0.051 | 0.000 | +0.051 |
| +4.0 | 4.048 | 4.000 | +0.048 |
| +8.0 | 8.043 | 8.000 | +0.043 |
To generalize the correction, an error model was established using MATLAB. The experimental data points (error as a function of inclination angle $$ \alpha $$ and measured depth $$ d $$) were interpolated to create a continuous two-dimensional error surface. The model covers an inclination range from 20° to 45° and a depth range of -10 mm to +10 mm, encompassing the operational conditions for the cylindrical gear inspection. The model can be represented as a correction function:
$$ E_{corr}(\alpha, d) = f_{interp}(\alpha, d) $$
where $$ E_{corr} $$ is the error compensation value to be added to the raw laser reading. This function was implemented as a lookup table or a fitted polynomial within the data processing software.
The core of the measurement process for the cylindrical gear involves precise trajectory planning. The gear is rotated incrementally on the C-axis while the laser sensor, fixed in position, captures distance data. For each angular position of the cylindrical gear, the laser beam intersects the tooth flank at a specific point. The theoretical profile of an involute spur gear is given by the parametric equations:
$$ x = r_b (\cos(\theta) + \theta \sin(\theta)) $$
$$ y = r_b (\sin(\theta) – \theta \cos(\theta)) $$
where $$ r_b $$ is the base circle radius and $$ \theta $$ is the roll angle. The surface normal vector at any point on the involute can be derived from its geometry. During measurement, for each acquired data point, the corresponding theoretical point on the ideal cylindrical gear profile is calculated based on the C-axis rotation angle. The theoretical normal vector at that point is then computed. The angle between this normal vector and the direction of the incident laser beam (which is fixed in the machine coordinate system) is the instantaneous incidence angle $$ \alpha_i $$. This angle is crucial for error correction.
A MATLAB simulation was developed to emulate this process. The program generates the theoretical involute profile of the cylindrical gear under test. It then simulates the rotational motion, identifying the start and end points of the measurement on the tooth flank. For each simulated measurement point, it calculates the theoretical normal and the corresponding incidence angle $$ \alpha_i $$. This workflow is illustrated conceptually: the laser trajectory is a line in space, while the gear rotates, causing the intersection point to trace the involute curve. The colored lines representing surface normals show how the angle changes along the profile. Using the calculated $$ \alpha_i $$ and the raw laser distance value $$ d_{raw} $$, the error compensation value $$ E_{corr}(\alpha_i, d_{raw}) $$ is retrieved from the pre-established model. The corrected distance value is then:
$$ d_{corrected} = d_{raw} + E_{corr}(\alpha_i, d_{raw}) $$
This corrected data set forms the basis for all subsequent gear error analyses.
The actual measurement procedure for a cylindrical gear specimen commences with system initialization: powering on, homing all machine axes, and mounting the gear between centers. The laser sensor is positioned via the X, Y, and Z axes so that the tooth flank to be measured lies within the sensor’s ±10 mm range. The sensor is activated, and the C-axis rotates the gear incrementally. At each angular step, the laser displacement value is recorded. The data acquisition rate of the LK-H050 sensor is 100 Hz (10 ms period), yielding approximately 100 data points per second. For a single tooth flank scan, the effective portion containing the involute profile typically comprises around 95 data points. These raw data are processed offline using the MATLAB routine implementing the angle calculation and error compensation.
Key gear accuracy parameters, as per standards like GB/T 10095.1-2008 (similar to ISO 1328), include tooth profile deviation and pitch deviation. The tooth profile total deviation $$ F_\alpha $$ is the distance between two design profile lines that enclose the measured actual profile over the evaluation range. After compensation, the measured profile points are compared to the theoretical involute. The deviation at each point is computed, and $$ F_\alpha $$ is determined as the maximum positive deviation plus the absolute value of the maximum negative deviation. For a cylindrical gear, this reflects the accuracy of the involute form. A sample of calculated profile deviations for one tooth from our experiment is shown in Table 3.
| Data Point Index | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 |
|---|---|---|---|---|---|---|---|---|---|
| Profile Deviation | 1.1 | 4.7 | 4.8 | 6.1 | 7.6 | 7.8 | 8.7 | 6.3 | 6.1 |
| Data Point Index | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | … |
| Profile Deviation | 5.6 | 4.7 | 3.5 | -1.1 | 2.4 | 2.8 | 3.3 | 3.9 | … |
From such data, the tooth profile total deviation $$ F_\alpha $$ was found to be 8.9 µm (7.8 µm max positive + 1.1 µm abs(max negative)). According to the precision grade table for cylindrical gears, an $$ F_\alpha $$ value of 8.9 µm falls within the tolerance for Grade 6 accuracy (where the limit for a gear of this module and diameter might be, for example, 13.0 µm). This indicates that the measured cylindrical gear meets a relatively high precision grade.
Pitch deviation analysis is another critical aspect for cylindrical gears. The single pitch deviation $$ f_{pt} $$ is the algebraic difference between the actual pitch and the theoretical pitch between two adjacent teeth at the reference circle. The cumulative pitch deviation $$ F_p $$ is the maximum algebraic difference between the actual and theoretical cumulative pitches over any number of teeth. To compute these from the laser-measured data, the compensated profile points for all teeth are first fitted with cubic splines to reconstruct the complete tooth flanks. The intersection points of the reference circle (typically the pitch circle) with each fitted flank are calculated numerically. The distances between consecutive intersection points give the actual pitches. The deviations are then derived. Results for a 17-tooth cylindrical gear segment are presented in Table 4.
| Tooth Sequence Number | Single Pitch Deviation, $$ f_{pt} $$ | Cumulative Pitch Deviation, $$ F_p $$ |
|---|---|---|
| 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 |
From Table 4, key statistics can be derived: the maximum single pitch deviation $$ \max(f_{pt}) = 8.2 \, \mu m $$, the minimum single pitch deviation $$ \min(f_{pt}) = 1.9 \, \mu m $$, and the average single pitch deviation $$ \overline{f_{pt}} \approx 6.9 \, \mu m $$. The total cumulative pitch deviation $$ F_p $$ is the maximum range of the cumulative deviation column, which is $$ 27.3 \, \mu m $$ (from -27.3 µm to +0.3 µm). Comparing these values to Grade 6 tolerance limits (e.g., $$ f_{pt} $$ limit of 9.0 µm and $$ F_p $$ limit of 28.0 µm for this gear size), the cylindrical gear qualifies for Grade 6 precision in pitch parameters as well.
To validate the effectiveness of the error correction method, a comparison was made with traditional contact measurement results on the same cylindrical gear. The contact method, using a coordinate measuring machine (CMM) with a tactile probe, reported a maximum single pitch deviation of 5.2 µm and a cumulative pitch deviation of 22.8 µm for the right flank. The laser non-contact method, after error correction, yielded 8.2 µm and 27.3 µm respectively for the same flank. While the contact method shows slightly better (smaller) deviation values, likely due to its inherently high point-measurement accuracy and established calibration, the laser-based results are very close and, crucially, remain within the same precision grade (Grade 6). This demonstrates that the non-contact laser method, augmented with the inclination/depth error correction model, achieves a level of accuracy suitable for inspecting precision cylindrical gears. The minor discrepancy may be attributed to factors like surface reflectivity variations on the cylindrical gear or residual calibration uncertainties, but the overall agreement confirms the methodology’s viability.
The error correction model’s robustness can be further analyzed. The primary error source in laser triangulation on sloped surfaces is the deviation of the reflected speckle pattern from the assumed centroid position on the CCD. For a surface with inclination $$ \alpha $$, the effective displacement calculation assumes the surface normal is aligned with the laser axis. When misaligned, the error $$ \epsilon $$ can be modeled as a function of $$ \alpha $$ and the measurement depth $$ z $$. A simplified theoretical expression might be:
$$ \epsilon(\alpha, z) \approx k_1 \cdot z \cdot \sin(2\alpha) + k_2 \cdot \alpha^2 $$
where $$ k_1 $$ and $$ k_2 $$ are constants determined by sensor geometry. Our empirical model via MATLAB interpolation effectively captures this nonlinear relationship without requiring explicit analytical form, providing a practical correction lookup.
In industrial applications, the speed advantage of non-contact laser scanning for cylindrical gears is substantial. A full gear with 50 teeth can be scanned in a matter of minutes, capturing dense point clouds for comprehensive analysis of profile, pitch, runout, and even surface texture if the sensor resolution permits. This facilitates 100% inspection in production lines. The wear-free nature also allows for measuring delicate or finished cylindrical gear surfaces without risk of marring. Future work could integrate real-time error compensation directly into the machine’s CNC system, enabling on-the-fly correction during measurement. Additionally, expanding the error model to account for different material reflectivities or surface coatings would enhance the universality of the method for various cylindrical gear types.
In conclusion, this research presents a validated non-contact measurement approach for involute spur cylindrical gears using a laser displacement sensor. By systematically characterizing and correcting errors induced by surface inclination and measurement depth through a dedicated calibration experiment and MATLAB-based modeling, the measurement accuracy is significantly enhanced. The corrected results for key parameters like tooth profile total deviation and cumulative pitch deviation demonstrate that the method can reliably assess cylindrical gears to Grade 6 precision levels, comparable to traditional contact methods. The technique offers the distinct benefits of rapid, non-destructive inspection, making it a promising solution for quality control in high-volume cylindrical gear manufacturing. The integration of such laser-based systems with advanced data processing paves the way for more intelligent and efficient gear metrology, aligning with the industry’s trend towards digitized and integrated manufacturing systems.