Steer-by-Wire
Status: in development
Steering transmitted by cable rather than by column: a capstan winding a brake line, so the steering input can sit anywhere on the chassis relative to the wheels.
“By wire” here is literal. This is a mechanical cable system, not an electric one.
Why not just buy it?
Barrier 4: not sold as a product at all.
Steering is where the bicycle and car ecosystems both stop being useful.
A bicycle steers because the handlebar is the fork. There is no interface to buy, because there is no separation to make. A car has a steering column, rack, and linkage, but they are designed around a specific vehicle’s packaging and sold to manufacturers building thousands of that vehicle.
Neither gives you what an intermediate vehicle actually needs: a way to put the driver’s hands somewhere ergonomically sensible while the wheels are somewhere structurally sensible, without redesigning both every time.
Why cable, and not electric
This deserves emphasis, because “steer-by-wire” normally means electric actuation, and electric steer-by-wire would likely be a bad decision for this project.
Electric steering brings: a safety-critical control loop, redundancy requirements, a failure mode where loss of power means loss of steering, future certification burden, and a repair story that no bike shop can support. Every one of those contradicts Making It Viable.
A capstan and a brake cable bring none of them. The load path is mechanical and continuous, it is inspectable by eye, the components are bicycle commodities available everywhere, the failure modes are the well-understood failure modes of cable (stretch, fray, corrosion, and cable stretch is gradual and audible rather than sudden), and it can be adjusted at the roadside with an Allen key.
It is the lowtech answer. This is what Chapter 6 means by appropriate technology.
It also pays off elsewhere in the vehicle. Because steering and braking are entirely mechanical, the CAN bus carries no safety-critical function, and a total electrical failure leaves the rider with full control authority. Keeping the last line of control mechanical is what makes the rest of the electronics safe to experiment with.
What we built
Still working on it so not much to write here for now.
What this unlocks
Combined with the corner module, this is what makes the drive assembly chassis-agnostic. The corner unit brings its own suspension, braking, and propulsion; the capstan brings its own steering path. Between them, a chassis stops having to solve any of the hard mechanical problems and becomes purely a question of ergonomics, enclosure, and structure.
That is a direct answer to a constraint from Designing for Humans: if seating position, entry and exit, visibility, and cargo layout are the variables that determine whether someone will actually use the vehicle, then those variables need to be cheap to iterate on. Every hour spent redesigning a steering linkage is an hour not spent on the thing that decides adoption.
What failed, and what we would change
Since we did not build it yet, we dont know. But we can tell you what failed with the previous approach. It was based of pushrod mounted on an axis. this axis took the torque from a handle bar trough two poorly sized cardan with ony a little screw that started eating the metal rod and introduced way too much play for it to be usable.
Open questions
- How much steering feel is lost to cable friction, and does it matter for confidence in traffic?
- What is the safe inspection and replacement interval for the cable?
- Probably more with did not think about yet