Low-voltage loom, UCL Racing 2026
Replacing two separate harnesses with one integrated loom, for a car that in the end did not make it to scrutineering.
The problem
UCL Racing’s earlier cars ran two separate harnesses, one for the chassis and one for the engine. For 2026 the team moved to a single unit handling both engine management and data acquisition, which meant the two harnesses could become one integrated vehicle loom.
I joined the electronics team in October 2025 as junior IC loom designer, working on sensor selection and the low-voltage side of that loom.
Scope
The finished design is 55 connectors and 239 terminated pins. It carries:
- Engine management: crank and cam position, coolant and oil temperature, manifold absolute pressure, wideband lambda, ignition and injection
- Drive-by-wire: electronic throttle, two accelerator pedal position sensors with non-intersecting transfer functions, two throttle position sensors
- Chassis sensing: four-corner wheel speed, front ride height, steering angle, independent front and rear brake pressure
- Cockpit: dashboard, paddle shift, start and shutdown buttons
- Safety: the shutdown circuit, brake system plausibility device, inertia switch and brake-over-travel switch
Approach
The design ran on two coupled artefacts: a RapidHarness model for the physical loom, and an Excel matrix with a custom wiring macro for the pin-level bookkeeping.
Every physical device was normalised to a discrete connector node with a defined male/female mating pair, so that each terminal, wedgelock and wire seal was accounted for exactly once. Connectors were then clustered into localised sub-harnesses to keep runs short, with the shutdown-circuit components deliberately isolated from signal lines.
What broke
The car did not run.
The chassis was not fully cleared on the manufacturing side, and the team decided not to take the car to competition. The loom design was finished; it was never built into a car that turned a wheel.
A year of design work stopping there is frustrating, but a subsystem depends on every other subsystem’s schedule, and this is a normal way for a student car to fail.
Component choices, the wire specification and the pin-level matrix belong to UCL Racing and are not published here. Those are the parts a rival team could use directly, and the design is going into a car that still has to compete.
What I took from it
Sensor selection was the part I found hardest and learned most from. Picking a sensor commits you to its connector family, its current draw, its mounting, its lead time and its procurement status, not just the numbers on the datasheet. Half the design effort was bookkeeping, and the bookkeeping is what separates a loom that can be built from a diagram that looks right.
Tools
RapidHarness · Excel wiring matrix · Sensor selection · Automotive harness design