Status: in development

Firmware for cheap field-oriented control boards, tuned for e-bike hub motors, so that a kW-class drive can be commanded and instrumented instead of merely switched on.

Of everything in this project, this is the contribution we expect to outlast the rest.

Why not just buy it?

Barriers 2 and 3: sold but either not usable as delivered, too expensive or otherwise sold only as a closed system.

The commercial path is a bundle: motor, battery, display, and throttle, matched by the vendor and speaking a closed protocol. Inside the bundle everything works. Outside it, communication between components is undocumented, there is no supported way to add a function the vendor did not anticipate, and integrating anything custom means an OEM agreement and volume commitments. For a project whose method is change one variable and measure what happens thats a dealbreaker.

The cheap FOC boards, meanwhile, are sold into adjacent markets with thin documentation and firmware that does not anticipate an e-bike hub motor: high pole count, high inductance, low Kv, hall-sensored, and operated at low electrical frequency under high load. Buying one gets you a component, not a working drive. Getting from the first to the second is weeks of work but we are willing to do it once so that everybody can benefit from it.

Why this is the highest-leverage gap

Firmware has a property no mechanical part shares: replication costs nothing. One team writes it once, and every team afterwards gets it for the price of a download.

It also attacks the obsolescence mechanism described in Chapter 7: a vehicle whose drive electronics are open cannot be killed by a vendor discontinuing a display.

Relationship to VESC

VESC already did this for electric skateboards and light EVs, and remains the reference for what open motor control should look like. We are not trying to replace it.

The interesting claim is that the cheapest available hardware is sufficient for this application, and that only the software was ever missing.

What we built

still in progress, check the progress on our github.

Open questions

• Do the cheap boards survive sustained kW-class load thermally, or only in bursts?
• How far can hall-sensored low-speed control be pushed before sensorless is required?

do you like motor control ? Get in touch.