ePortfolio

Project: Cross Country Ski Camber Device
Group: I-34
Sponsor / Client: Patrick Gallager, TrailSports Canmore

This project created a machine that tests how a cross-country ski bends and distributes weight when someone stands on it. Instead of relying on guesswork, it measures exactly how the ski behaves under load, helping ski shops match customers with the best skis for their weight and style. This benefits both recreational skiers and high-performance athletes by improving comfort, efficiency, and performance.

For engineers: The system is a load-controlled test bench that applies up to 250 lb of force via a linkage-actuated mechanism and captures spatial pressure distribution using an array of 120 strain-gauge load cells (≈17.5 mm resolution). Signals are conditioned through HX711 ADCs and processed across a distributed Arduino architecture, with synchronized acquisition and CSV export handled by a Python-based interface. Key constraints included cost (< CAD $2,500), portability, sensor resolution/accuracy (≈±0.05 g), and ensuring purely vertical loading without damaging the ski.

Please check the project page for a 3D model.

The core problem we set out to solve was the lack of a standardized, accurate way to measure how a cross-country ski behaves under load, specifically how force is distributed along its length. Existing methods rely heavily on manual inspection, inconsistent manufacturer labels, and subjective judgment, which fail to capture the true performance characteristics of a ski.

This mattered because it directly impacts how well a skier is matched to their equipment, affecting performance, efficiency, and overall experience. Inaccurate assessments can lead to poor ski selection, reducing performance for competitive athletes and limiting enjoyment for recreational users, while also making it difficult for ski shops to provide consistent, data-driven recommendations.

Our primary objectives were:

Objective	Target Specification	Rationale
Measure load distribution along ski	High spatial resolution (~17.5 mm sensor spacing, 120 sensors total)	Required to accurately capture pressure distribution and identify camber characteristics for ski selection
Apply controlled load to ski	Up to 250 lb (≈113 kg) vertical force	Represents expected skier weight range and client requirement for realistic loading conditions
System cost constraint	< CAD $2,500 total cost	Defined by combined project budget from UBC and client, ensuring feasibility for real-world deployment

These targets were set through a combination of client consultation, literature review, and benchmarking against existing solutions in the space.

We didn’t arrive at the final design in one shot — the solution went through several meaningful iterations before it worked the way we needed it to. Here's how that progression unfolded:

Iteration 1 — Initial Concept

Our first approach was to utilize repurposed load cells from commercial kitchen scales arranged in a “piano key” configuration along the beam to measure load distribution. This concept was selected due to its low cost and perceived simplicity, as the load cells were already available and required minimal initial fabrication. At this stage, the design focused on validating the feasibility of capturing distributed force measurements along the ski while maintaining a simple mechanical mounting approach using extruded plastic supports. The system was still in an early prototyping phase, with the primary goal being to confirm that measurable and repeatable pressure data could be obtained before committing to a more complex and integrated design.

Iteration 2 — First Prototype

Moving to a physical prototype exposed problems that simulation had hidden. Specifically, the initial “piano key” load cell configuration produced poor measurement resolution because multiple load cells were combined in a full Wheatstone bridge, reducing the ability to resolve localized pressure variations. In addition, the sensors exhibited drift and failed to return to their original baseline after loading, resulting in unreliable and inconsistent readings. This forced us to abandon the kitchen scale load cells entirely and transition to individual straight bar strain gauge load cells in a quarter-bridge configuration, allowing each sensor to provide independent, repeatable, and higher-resolution measurements along the ski.

Iteration 3 — Refined Design

The second prototype incorporated a complete redesign of the sensing system, replacing the initial kitchen scale load cells with 5 kg straight bar strain gauge load cells configured in a quarter Wheatstone bridge. Each sensor was paired with an HX711 amplifier and distributed across a modular architecture of 15 Arduino Nanos, allowing for independent, high-resolution measurements along the ski. In parallel, the mechanical system was refined to include a linkage-driven actuator capable of applying a controlled vertical load while minimizing horizontal force, and a standardized mounting approach was implemented to ensure consistent sensor preload.

Testing showed a significant improvement in both resolution and accuracy. The system achieved a spatial resolution of approximately 17.5 mm across 120 sensors and a measurement error within ±0.05 g after calibration, with consistent return-to-zero behavior across all sensors. At this point, the design was stable enough to move toward final integration and verification.

After verification testing against our original specifications, the final system achieved the following:

Metric	Target	Achieved	Status
Maximum Load Capacity	≥ 250 lb (113 kg)	250 lb sustained load with no structural failure	Met
Sensor Accuracy	Within ±2–3% of applied load	±0.05 g error after calibration	Met
System Portability	Transportable within shop and between locations	Functional but relatively heavy 7 ft beam system	Partial

System portability fell slightly short of target because the full-length aluminum beam and integrated components resulted in a relatively heavy and bulky structure. This was a tradeoff made to achieve the required stiffness, sensor resolution, and mechanical robustness. If we were to continue the project, we would investigate a modular beam design with detachable sections to improve transportability without sacrificing performance.

Overall, the project demonstrated that a low-cost, high-resolution ski camber testing system can be successfully developed to provide accurate, repeatable measurements of load distribution along a ski. The final system enables standardized, data-driven ski selection, improving both performance outcomes for athletes and efficiency for ski shop operations.

On a team of 6, my primary ownership areas were:

Area	What I Did
Electrical System Design	Selected and integrated system hardware including load cells, HX711 ADCs, microcontrollers, and communication architecture. Iterated the electrical design to balance modularity, reliability, and cost while supporting 120 sensors.
Firmware & Software Development	Developed Arduino firmware for distributed data acquisition and actuator control. Built the PC-side Python application to handle parallelized data collection, calibration routines, real-time load control, and CSV data export.
Testing & System Verification	Designed and executed tests to validate actuator performance, sensor accuracy, and full system functionality. Verified communication between all subsystems and ensured stable operation under full load conditions.

Beyond the technical work, I contributed to system integration, workflow optimization for assembly, and multiple sections of the final report. I also participated in client discussions and design iteration decisions, ensuring the final system met both technical and practical requirements for deployment in a retail environment.