Design Overview of Universal Wire Testing Machine

The design objectives for the wire testing machine are defined as follows:

• Eliminate initial wire stretch
• Measure tension loss for various combinations of wires and pulleys of different diameters
• Measure wire tension when the total wire length dynamically changes

• Multi-Pulley Pre-Stretching System
• Passive Pulley Transmission Efficiency Measurement System
• Variable-Length Wire Dynamics Measurement System

Multi-Pulley Pre-Stretching System

The Multi-Pulley Pre-Stretching System applies tension to a new wire to eliminate initial wire stretch. This system comprises three components: a wire anchoring point, a multi-pulley for redirecting the wire, and a lever that applies tension. By placing a weight on the lever, a constant tension is maintained on the wire, thereby removing initial wire stretch.

Passive Pulley Transmission Efficiency Measurement System

The Passive Pulley Transmission Efficiency Measurement System measures tension transmission efficiency for combinations of wires with diameters from 1 mm to 3 mm and pulleys of eight diameters between 12 mm and 60 mm. This system comprises four components: a wire winding module, an input wire tension sensor, a passive pulley unit, and an output wire tension sensor.

Variable-Length Wire Dynamics Measurement System

The Variable-Length Wire Dynamics Measurement System is an apparatus that measures wire tension and strain when a load is moved up and down by a wire. This system consists of four components: a wire winding module, an input tension sensor, a passive pulley unit, and a linear loading unit. Using this apparatus, it is possible to measure the dynamics under conditions in which the total wire length changes, as occurs in actual wire-driven robots.

Experiments

Measurement of Tension Transmission Efficiency in Wire Pulleys

Using the Passive Pulley Transmission Efficiency Measurement System, we measured the tension transmission efficiency for combinations of eight pulley diameters and four wire types. For all wire types and applied tensions, it was confirmed that the tension transmission efficiency increases monotonically with pulley diameter. Moreover, for a given pulley diameter, the tension transmission efficiency decreases as the wire diameter increases. These experiments demonstrate that by increasing the passive pulley diameter and reducing the wire diameter, a low-friction transmission mechanism can be realized in wire-driven robots.

Measurement of Dynamic Characteristics of Wires

Using the Variable-Length Wire Dynamics Measurement System, we investigated the frequency response of wire tension when the wire length varied and when it remained constant. In this experiment, a sinusoidal wave with constant amplitude and time-varying frequency was input as the tension command to the wire testing machine, and the actual tension was measured. The dynamic change in wire length reduces tension tracking performance by approximately 5 dB at low frequencies, while halving the resonance amplification in the mid-to-high frequency range and mitigating the progression of phase lag. This experiment revealed a significant difference in tension characteristics between the variable-length wire and the constant-length wire, demonstrating the importance of the Variable-Length Wire Dynamics Measurement System.

Wire-Driven Manipulator Control Considering Tension Transmission Efficiency in Wire Pulleys

Using the tension transmission efficiency of the passive pulley measured in previous experiments, we applied it to the force control of an actual wire-driven robot. In this experiment, a target end-effector force was commanded to the coupled wire-driven mechanism, and the error between the commanded end-effector force and the actual force was measured by a six-axis force sensor at the end effector. We then compared the end-effector force error with and without compensation for tension loss in the passive pulley.

The end-effector force error during the experiment is shown below. With tension loss compensation control, the root mean square error of the vertical force component was 9.5 N, whereas it was 10.9 N without compensation. This experiment confirmed that using data obtained from the wire testing machine can improve the force control accuracy of an actual wire-driven mechanism.

Bibtex

@inproceedings{t-suzuki2025wiretester,
  title={{A Universal Wire Testing Machine for Enhancing the Performance of Wire-Driven Robots (in press)}},
  author={Temma Suzuki and Kento Kawaharazuka and Kei Okada},
  booktitle={2025 IEEE-RAS 24th International Conference on Humanoid Robots (Humanoids)},
  year={2025},
}