The performance of a planetary roller screw assembly is paramount, encompassing metrics such as positioning accuracy, noise, temperature rise, vibration, and maximum operational speed. Existing test methodologies often rely on static measurement techniques, which are not only operationally complex but also susceptible to significant random errors. To bridge this technological gap, this article details the design and implementation of a dynamic, integrated test system specifically for evaluating the comprehensive performance of planetary roller screw assemblies. The system is engineered to perform real-time measurements under high-speed conditions. A central innovation is the integration of a thermal error compensation module, which significantly enhances the reliability of precision testing.
The test bed is built upon a robust hardware architecture centered on a servo-controlled actuation system. The core principle involves converting the rotary motion of a servo motor into precise linear travel of the nut. The system’s schematic illustrates the strategic placement of various sensors to capture a holistic performance profile. The hardware installation includes components such as a circular grating for angular position, infrared temperature sensors at the front/rear bearing blocks and nut, vibration and acceleration sensors, a laser interferometer for ultra-precise linear displacement, a laser displacement sensor for thermal growth, a noise sensor, and a tension/compression load cell.
The servo system is designed for high performance and reliability. It utilizes a central drive unit (CU320-2PN) communicating via PROFINET protocol with a Programmable Logic Controller (PLC). The power infrastructure includes filtering modules to ensure voltage stability and a rectifying power module that maintains the DC bus voltage, crucial for preventing faults during high-speed or high-load operation. The main servo motor is equipped with an absolute encoder (AM22DQ), while the circular grating encoder connects via an SMC-30 module. To manage heat generation during prolonged tests, a dedicated water-cooling system is integrated. A block diagram of this system effectively summarizes the signal and power flow from the AC mains, through filtration and conversion, to the motor and its feedback devices.
Thermal Error Detection and Compensation Methodology
Frictional heat generated at the interfaces between the screw, rollers, nut, and bearings induces thermal expansion in the screw shaft. This expansion directly compromises the system’s positioning accuracy, making its measurement and compensation critical for high-precision applications. Our system employs a dedicated procedure to measure the axial thermal growth of the screw.
The data acquisition for thermal error utilizes a precise, interrupt-driven method. The procedure initiates by establishing a reference “cold” position. The software commands the servo motor to position the nut at a specific axial location. At this point, an interrupt routine is triggered, collecting 50 data points from the laser displacement sensor focused on the screw end. This dataset is sorted, and the median 30 values are averaged to establish a highly reliable baseline position, \( L_{initial} \). Following this, the planetary roller screw assembly is exercised through several reciprocating cycles to generate operational heat. After a predetermined time or number of cycles, the nut is returned to the exact same angular position (verified by the circular grating reset to zero) and near-identical axial location. The interrupt routine is triggered again to measure the new position, \( L_{hot} \). The temperature at the screw’s end is simultaneously recorded via a serial communication interface with the infrared sensor. The axial elongation, \( \Delta L \), is calculated as \( \Delta L = L_{hot} – L_{initial} \).
The relationship between the measured temperature rise (\( \Delta T \)) and the axial elongation (\( \Delta L \)) is then modeled. Data points are plotted with \( \Delta T \) as the abscissa and \( \Delta L \) as the ordinate. A first-order linear model is fitted using the least squares method:
$$ \Delta L = a + b \cdot \Delta T $$
The parameters \( a \) (initial offset) and \( b \) (thermal expansion coefficient) are determined by minimizing the sum of squared errors. The formulas for calculating these parameters from \( n \) data points (\( \Delta T_i, \Delta L_i \)) are:
$$ b = \frac{n \sum{(\Delta T_i \Delta L_i)} – \sum{\Delta T_i} \sum{\Delta L_i}}{n \sum{\Delta T_i^2} – (\sum{\Delta T_i})^2} $$
$$ a = \frac{\sum{\Delta L_i} – b \sum{\Delta T_i}}{n} $$
The goodness of fit is evaluated using the coefficient of determination, \( R^2 \):
$$ R^2 = 1 – \frac{\sum{(\Delta L_i – \hat{\Delta L}_i)^2}}{\sum{(\Delta L_i – \bar{\Delta L})^2}} $$
where \( \hat{\Delta L}_i \) is the predicted elongation from the model and \( \bar{\Delta L} \) is the mean of the measured elongations. An \( R^2 \) value close to 1 indicates a strong linear relationship. For a tested planetary roller screw assembly, the fitted relationship was \( \Delta L = 0.0096 + 0.0128 \cdot \Delta T \) mm, with \( R^2 = 0.9496 \). This model is subsequently used to compensate positioning accuracy measurements in real-time, subtracting the predicted thermal elongation from the raw displacement readings.
Integrated Performance Test System and Experimental Results
The dynamic test system is controlled and monitored through a custom-developed LabVIEW-based human-machine interface (HMI). This software handles motion control command generation, synchronized data acquisition from all sensors, real-time graphical display, data logging, and post-processing analysis, including the execution of the thermal error compensation algorithm.
1. Precision Positioning Test with Thermal Compensation
This test evaluates the fundamental accuracy metrics of the planetary roller screw assembly: bidirectional positioning accuracy (\(A\)), bidirectional repeatability (\(R\)), and backlash (\(B\)). The test involves programming the nut to move to five target positions over its effective travel range. The motion profile consists of five round trips to each target. The entire test sequence is repeated five times to ensure statistical significance. The actual position of the nut is measured with a laser interferometer, the gold standard for displacement metrology.
Key parameters for a sample test are summarized in the table below:
| Target Position ID | Commanded Displacement (mm) |
|---|---|
| P1 | 20 |
| P2 | 40 |
| P3 | 60 |
| P4 | 80 |
| P5 | 100 |
The measured position-time data for the five reciprocating cycles clearly shows the motion profile and the minute deviations at each target. The analysis software calculates the accuracy metrics from this data both with and without the application of the thermal error compensation model. The results demonstrate the significant impact of compensation:
| Performance Metric | Uncompensated Value (μm) | Compensated Value (μm) | Improvement |
|---|---|---|---|
| Bidirectional Positioning Accuracy (A) | 62.07 | 37.66 | 39.3% |
| Bidirectional Repeatability (R) | 18.74 | 12.41 | 33.8% |
| Backlash (B) | 28.32 | 15.26 | 46.1% |
The improvement across all metrics exceeds 30%, validating the necessity and effectiveness of the thermal error compensation system for high-precision assessment of the planetary roller screw assembly.
2. High-Speed Dynamic Performance Tests
These tests evaluate the behavior of the planetary roller screw assembly under demanding operational conditions, measuring parameters like velocity, acceleration, temperature rise, noise, and vibration.
Velocity and Acceleration: With the servo system commanded to achieve a linear nut speed of 380 mm/s and an acceleration of 2g, the system accurately tracks the profile. The measured velocity and acceleration waveforms confirm the system’s ability to follow the commanded dynamic trajectory, which is crucial for applications involving rapid reciprocation or high throughput.
Temperature and Noise Monitoring: During prolonged high-speed operation, the temperature at key points (rear bearing, nut, front bearing) is recorded. A typical profile shows temperatures rising from an ambient baseline to a steady-state operating temperature. The noise emission is measured using a calibrated sound level meter. The test protocol involves first measuring the background noise level (\(dB_1\)) using a filtered median of 50 samples. During planetary roller screw assembly operation, the measured noise (\(dB_2\)) is compared to \(dB_1\). The final reported noise level is adjusted based on the difference: if \(dB_2 – dB_1 > 10\) dB(A), \(dB_2\) is used; if the difference is between 6-8 dB(A), 1 dB is subtracted; if between 8-10 dB(A), 0.5 dB is subtracted, ensuring an accurate representation of the assembly’s noise contribution.
Vibration Analysis: Vibration is measured in three axes: axial, radial horizontal, and radial vertical. Tests are conducted at incrementally increasing nut speeds: 150, 200, 250, 300, and 350 mm/s. The results, typically plotted as vibration amplitude versus speed, show a general trend of increasing vibration with speed. However, the absolute increase is often relatively small (e.g., a few mm/s² over a 200 mm/s speed increase). This suggests that a well-manufactured planetary roller screw assembly maintains stable dynamic characteristics across its speed range. Potential causes for increased vibration include progressive wear of contacting surfaces or the development of minor clearances within the roller raceway interfaces over time.
Conclusion
The designed and implemented dynamic test system provides a comprehensive, reliable, and efficient platform for evaluating the critical performance parameters of planetary roller screw assemblies. It successfully performs synchronized, real-time measurements of positioning accuracy, noise, temperature rise, vibration, maximum speed, and acceleration under realistic high-speed operating conditions. The integration of a model-based thermal error compensation system, founded on precise measurement and least-squares modeling of the screw’s thermal expansion, proves to be a pivotal advancement. Experimental results demonstrate that this compensation leads to an improvement of over 30% in key positioning accuracy metrics, significantly enhancing the fidelity of the test outcomes. This system offers a valuable tool for quality assurance during manufacturing, performance validation for application engineering, and research into the dynamic behavior of planetary roller screw assemblies.