The technology
A short, honest look at how Diego works: the actuation, the control, the materials, and the design choices behind every part.
Status: pre-clinical. Engineering prototypes in active development.
Actuation
Motor-and-linkage drive. Closer to how a hand moves.
A human finger does not move like a gear. It moves through a coupled curl, with multiple joints flexing together because tendons pull along the length of the finger. We chose a motor-and-linkage drive because it gets closer to that than any gear-driven design we tested.
Compact micro-motors spool tendons that pass through linkage joints in each finger. The linkages constrain motion to natural arcs and spread load across joints instead of one output stage. The result is movement that feels right: quiet, smooth, and proportionate to the muscle signal driving it.
Engineering detail
Each finger module is driven by an encoder micro-motor with closed-loop position sensing, pulling a tendon through our relay-rocker bell-crank linkage (patent pending). The linkage shapes the curl path so motion looks and feels like a hand. Compared with gear-train designs, it reduces fingertip inertia, lowers acoustic noise, and lets us place motors where weight balance is best rather than where the joint forces them.
Control
Designed around EMG, not adapted to it.
When you try to move a hand that is no longer there, the muscles in your residual limb still produce small electrical signals (electromyographic, or EMG, signals). Diego reads those signals at the socket, processes them, and turns them into movement.
Many prosthetic hands treat EMG as an add-on, bolted onto a control loop designed for something else. We built the control architecture, firmware, and decoding model around EMG from the start. The payoff is lower latency, cleaner intent detection, and a calibration process measured in minutes rather than hours.
The signal chain, end to end
Acquisition
Skin-contact EMG sensors inside the socket capture muscle activity from selected sites on the residual limb.
Conditioning
Filtering and shielded routing reject mains hum, motor noise, and motion artefacts before the signal is digitised.
Processing
Onboard firmware extracts envelope, contraction onset, and intent features in real time, tuned for low-latency control.
Control
A closed-loop controller maps intent into grip patterns and proportional finger speed.
Calibration
A short clinician-led session personalises thresholds, gain curves, and grip mappings to your muscles. Re-calibration will always be available.
Research in progress
We are developing an in-house EMG decoder aimed at faster, more data-efficient calibration, so a new user spends less time training the system before it responds usefully. This work is in development. We are re-running our evaluation and will publish results once the work is complete and peer-reviewed.
Materials and manufacturing
3D printing is not the headline. It is the reason this works at all.
Customisation, speed, and cost are not separate problems. Additive manufacturing addresses all three at once.
Custom from the scan
Every shell is printed to fit one residual limb. No off-the-shelf socket pretending to be custom.
Fast iteration
A new shell can be printed and ready overnight. Comfort issues are solved in days, not months.
Cost control
Additive manufacturing avoids the tooling and inventory costs that drive traditional prosthetic prices.
Strong, light, skin-safe
SLA-printed photopolymer for structural parts. Biocompatible resin for any part that touches skin. Materials chosen for resolution, repeatability, and skin-contact safety.
Quality control
We are building the QC programme production will run on: dimensional inspection, layer-integrity inspection, mechanical testing of structural parts under cyclic load and temperature, and lot-level material traceability. The goal is that every component on every Diego traces back to the batch it came from.
Modularity
Designed in pieces. On purpose.
Most prosthetic hands are built as one sealed unit. When something breaks, the whole hand goes back to the manufacturer. We think a prosthetic hand should be more like a good bicycle than a sealed appliance: the parts that wear should be the parts you can replace.
Diego is designed in modules from the inside out. Tendons, finger linkages, motor units, batteries, and shells are separate, replaceable assemblies joined by standardised, documented interfaces. Most maintenance happens at home or in clinic. We only enter the picture when something local repair cannot reach.
You are not without your hand for weeks while parts ship.
Most service work fits inside a regular appointment.
Less device replacement means less e-waste and less disposable plastic.
Component upgrades reach existing users, not just new ones.
Research and development
What we are working on now.
Some projects run alongside clinical partners. Some are in-house benchwork. All of them feed the next Diego.
Tendon durability testing
Cyclic load and fatigue analysis to predict lifecycle, refine materials, and tighten safety margins for high-use patients.
Compliant joint design
Biomimetic joint geometries that improve comfort, expand range of motion, and reduce socket pressure points.
Adaptive grip algorithms
Machine-learning approaches that recognise grasp intent and adapt grip patterns to context. We are developing an in-house EMG decoder for faster, more data-efficient calibration, with results to follow once the work is complete and peer-reviewed.
Human–machine interaction
User studies on EMG site selection, feedback modalities, and intuitive control across daily tasks.
Further out, we are exploring sensory feedback and lower-limb extensions of the same platform.
See our research partnersTechnical resources
Documentation, papers, and the work behind the work.
We are building a small, well-organised library of technical documentation. Some will be public, some request-only. Below is what we are working on.
White papers
Drive-system design, EMG signal chain architecture, modular component validation.
First publications planned for 2026.
Conference talks
Presentations and posters from biomechanics, robotics, and rehabilitation conferences.
First publications planned for 2026.
Technical data sheets
Component specifications, interface drawings, and service manuals for clinicians and partners.
First publications planned for 2026.
Need detailed technical documentation?
Many resources will be available on request to verified clinicians, researchers, and institutional partners.