Status: in development

Chainless pedal generation: the rider turns a generator, the generator charges the pack, the motor drives the wheel. We get full freedom to put the pedal wherever we need.

Built from a large drone BLDC motor and a gearbox, because the commercial version unobtainium.

Why not just buy it?

Barrier 1: minimum order quantity.

The purest example of it in the whole project.

Chainless pedal generation is not experimental. It is in production, and shipping in commercial cargo bikes today. It is a solved problem, and we are not claiming to have invented anything.

We found three suppliers. All three sell business-to-business only, and none of them publishes enough to evaluate the part before entering a sales process.

No torque curve, no efficiency figure, no cadence range, no mass, no mechanical interface. Every route to a specification runs through a salesperson.

Cixi is the only one we actually pursued, and the answer was unambiguous: we could not afford it a development kit, for a part whose unit price is probably closer to 1 000 CHF, with the documentation needed to integrate it available only under NDA. That is an entirely reasonable offer to a manufacturer planning a production run. It is not an offer to a team asking whether the concept is worth pursuing at all. and an NDA on the datasheet would leave us unable to publish what we learned and is imcompatible with our ideal of repairability.

We did not contact Mando or Schaeffler, and that is a decision rather than an oversight. With nothing published, opening a conversation means asking a large supplier to spend time qualifying a customer who wants one unit, cannot commit to volume, and intends to publish. We already know how that ends, and we would not be allowed to write down the result. The absence of public data is itself the finding, and it is the finding this page exists to record.

What they feel like

From our personal experience, pedal feel differs enormously between two products that do nominally the same thing.

Mando SPM, which we rode at Spezi 2026 : Pedal input and vehicle response are plainly decoupled: you pedal, and roughly two to three seconds later the vehicle begins to accelerate. The lag dominates the experience. It works, but it does not feel like riding anything.

Cixi PERS is the opposite: a very well finished pedal experience with genuine resistance feedback, close to what a mechanical drivetrain gives you.

That contrast is worth highlighting, because it partly answers one of our own open questions before we started. Pleasant pedal feel with no mechanical connection is clearly achievable. The remaining question is about user reception and which one is actually preffered.

What we built

To document later but big drone BLDC, an ODrive clone to drive it and a pedal gearbox found on aliexpress, the later need a redesign because its very noisy but generally the principle seems sound.

First, what the pedals are actually for

The efficiency of this system is worse than a chain. We know, and it does not matter, because the pedals are not there to move the vehicle.

A person sustains on the order of 75–150 W. A vehicle in this class at realistic commuting speeds needs several times that. Even with a perfect, lossless generator, rider input would cover only something like 20–30% of the energy required to move the vehicle. The remainder comes from the pack regardless of how the pedals are connected.

So optimising the pedal path for energy recovery is optimising a rounding error. A few percent of a fifth of the total is not a design driver.

The pedals are there for a public health reason.

Living Within Limits identifies sedentary mobility as one of the externalities of car dependency, alongside emissions and noise, and it is one of the few that a vehicle can address directly rather than by consuming less. A commute that is currently an hour of sitting becomes an hour of light, self-paced physical activity, without requiring the user to be fit enough to cycle it unassisted, arrive sweating, or own the confidence in traffic that Designing for Humans identifies as a real barrier.

That reframes the entire module. Pedalling is not a power source that happens to be good for you. It is a health feature that happens to return some energy.

Why chainless follows from that

Once the pedals are understood as an ergonomic and health system rather than a drivetrain, the case for decoupling them becomes much stronger.

Effort becomes a tunable parameter. This is the decisive one. If the purpose of pedalling is physical activity rather than propulsion, then the right amount of effort is a property of the rider, not of the gearing: different fitness levels, different ages, different willingness to arrive warm. With no mechanical coupling, that relationship is configuration rather than hardware. The question “how much effort should this vehicle ask of this person?” can be tested across many users and conditions without building a physical variant for each answer which is what makes the research method affordable, and is exactly the argument the impact strategy makes for software-configurable assistance.

Packaging freedom. A chain line is a hard geometric constraint running the length of the vehicle, forcing the pedals, the seat, and the driven wheel into a fixed relationship. Removing it makes seating position and pedal position independent, which is precisely the ergonomic freedom that decides whether people will use the vehicle at all.

Drivetrain simplicity. No chain, no derailleur, no chain wear, no chain line maintenance, nothing oily exposed to weather along the length of the vehicle.

The counter-argument

The efficiency objection is answered above. The repairability objection is not.

A chain is the most universally serviceable component in transport history. A generator and gearbox are not, and Making It Viable is explicit that repairability is a disqualifier rather than a trade-off. A twenty-year vehicle whose pedal system cannot be fixed by a bike shop in 2045 has a real problem, and choosing commodity parts for the generator is still an open challenge.

What failed, and what we would change

The current implementation is wayyy too noisy and heavy.

Reproduce it

We will add later the part we used for that.

Open questions

• Is there a serviceable version of this that a bike shop could repair in fifteen years? The one we most need to answer.
• If it matter, how much of Cixi’s pedal feel is hardware and how much is the control loop? We now know good feel is achievable without a chain; we do not know what it costs to reach it with commodity parts.
• What assistance curve do people actually choose when they are free to choose one, and does it change over weeks of daily use?
• Does the health benefit survive contact with reality, or do riders simply turn the effort down to zero and leave it there?

Feel free to contact us if you have some light to shed.