F.I.D.G Motion

QDD Actuator

Problem statement

Imagine having a robotics idea that could change how people walk, lift, or recover, but not being able to test it because a single joint costs more than your whole hardware budget. It’s not a niche complaint. Talking with representatives from robotics startups like Cobot and large players such as Tesla and TE Connectivity, we kept hearing the same bottleneck; labs and small teams need compact, high torque actuators to build legged robots, exoskeletons, and manipulators, but the quasi-direct drive (QDD) actuators suited to these systems run hundreds of dollars each [1], and a multi-joint platform can need upwards of 40 of them. Commercial units like CubeMars perform well, but they’re closed designs with limited documentation and long lead times, while existing open-source options cut the price but leave the team to source parts, machine components, and integrate the motor, gearbox, and driver themselves. We’re building a third option, a capable, documented actuator that hits roughly 80 percent of commercial performance, continuous torque, peak torque, torque density, backdrive torque, and weight, at a fraction of the cost, and works out of the box with standard mechanical interfaces. The goal is to put hardware that larger programs take for granted into the hands of the small labs and startups that can’t. [1] CubeMars, “AK Series Actuators,” CubeMars Store. Accessed Sep. 18, 2026. https://store.cubemars.com/

Team members

Miller Dayton – facilitator
Spencer Funk – communicator
Sigurd Isaacson – accountant
William Greene – admin

Client

Miller Dayton
UW – Mechanical Engineering