GETMILK

Aura — Open-Source Bionic Hand

A printable, surface-EMG controlled prosthetic hand designed to give amputees a dignified daily-driver for a fraction of the price of a clinical device. Open hardware, open firmware, and a builder community of teen makers.

complexity: hardtolerance: precision
Aura — Open-Source Bionic Hand

Snapshot

Feasibility at a glance

PT localization

8/10

High

The whole build can be sourced in Portugal.

Per unit

€180–€280 per hand (BOM)

at 50-unit volume

Starter batch

50units

minimum viable run

To first batch

20weeks

6 phases, design to ship

Budget

€9K–€20K

all-in estimate

Bottom line

FDM PETG for strength, SLA resin for the knuckles where geometry is tight.
Read full insight

Picked based on the localization score (8/10) plus the dominant process (fdm printing). For an early build at this batch size, this gives you the best balance of cost, risk, and speed of iteration.

What it does

4 capabilities

Reads tiny electrical signals on the surface of the wearer's forearm.

Translates those signals into grip patterns — pinch, key-grip, wave, fist.

Drives five independent finger motors so the user can hold, point, or rest.

Logs grip-quality data so the user and their physiotherapist can tune the device together.

How it's made

5 stations · build route

  1. 1

    Print the shells

    FDM PETG for strength, SLA resin for the knuckles where geometry is tight.

  2. 2

    Assemble the drivetrain

    Thread tendon cables through the knuckles, terminate on the servo horns.

  3. 3

    Solder the EMG boards

    SMT reflow for the analog front-end; the rest is hand-pluggable.

  4. 4

    Calibrate to the wearer

    30-minute session: the firmware learns the user's specific muscle pattern.

  5. 5

    Hand-tune the grips

    Adjust spring tension on each finger so a glass and an egg are both safe.

Risks to watch

3 identified · 1 blocking

1 high
2 medium

High

Single-source ICs become unobtainable

Modern designs almost always include one IC with ≥30 week lead times when demand spikes.

Mitigation — Pin the BOM to two qualified part numbers per critical line and place rolling forecast orders.

Medium

Firmware ships before it's stable

Hardware looks done, software is still landing fixes — early customers feel it.

Mitigation — Hold the launch ship-date on a single firmware-stability KPI, not on hardware readiness alone.

Medium

CE / FCC test fail on first attempt

Most first-time radios fail EMC; budget a re-spin.

Mitigation — Pre-scan in a partner anechoic chamber before the official lab visit.

Production timeline

20 weeks to first batch

wk 1wk 7

FDM 3D printing

wk 1–2

SLA 3D printing

wk 3

PCB fabrication

wk 4

SMT pick-and-place

wk 5

Wire-harness build

wk 6

Functional QA & batch sign-off

wk 7

wk 1–2 is the longest stretch — FDM 3D printing takes 2 weeks of the 7 weeks on this build.

What goes into it

6 materials · 7 processes

Materials

PETG (shell)TPU 95A (joint pads)Stainless steel tendon cableFR-4 PCBLithium-polymer cellSilicone fingertip pads

Processes

fdm_printingsla_printingpcb_fabricationsmt_assemblywire_harnessfinal_assemblytesting_inspection

Design philosophy and best practices

5 principles

A myoelectric hand has to be light, quiet and reliable while turning noisy muscle signals into grips the wearer trusts. Weight at the end of the arm, signal quality and finger durability decide whether people keep wearing it.

1

Keep mass close to the wrist

Every gram at the fingertips feels heavier on a residual limb. Put motors, battery and board near the socket and route tendons out to the fingers.

2

Treat EMG quality as a mechanical problem

Electrode contact pressure, sweat and socket fit cause most signal loss. Design the electrode mount and shielding with the socket, not after it.

3

Make fingers fail safe and back-drivable

A finger should open under overload instead of breaking or crushing. Compliant joints and torque limits protect both the user and the gear train.

4

Plan for medical device rules early

A prosthesis sold in the EU falls under the Medical Device Regulation (EU 2017/745). Risk management to ISO 14971 and electrical safety to IEC 60601-1 shape design choices.

5

Design tendons for easy replacement

Cables stretch and fray. Accessible anchors and standard crimps let a technician swap one in minutes.

Common pitfalls

Avoid

Testing grips only on the bench and not on a real socket with sweat and movement.

Avoid

Choosing printed parts without checking fatigue and creep under constant tendon load.

Avoid

Leaving battery placement and heat to the end, then finding the socket too heavy or warm.

Avoid

Assuming an open-source design avoids medical device obligations when it is sold or fitted.

10-year evolution forecast (2026-2036)

4 phases

Affordable bionic hands are moving from hobby projects towards fitted clinical devices. The points below are trends, not forecasts.

2026–2028

Better signal processing

Trend: pattern recognition on small microcontrollers replaces simple two-electrode open and close control.

2028–2031

Sensory feedback

Trend: vibration or pressure feedback tells the wearer how hard they are gripping.

2031–2034

Clinic-led customisation

Trend: 3D scanning and printing of sockets and shells becomes routine in fitting clinics.

2034–2036

Longer life and repair networks

Trend: modular parts and local repair reduce downtime and total cost of ownership.

Outlook

Expect lower-cost hands to compete on fit, reliability and clinical support more than on finger count.

Materials intelligence

8 materials

PETG

Low supply risk

Printed shell and brackets

Sourced from: Made from PET resins produced in the EU, US and Asia; filament widely available.

TPU 95A

Low supply risk

Flexible joint pads

Sourced from: Thermoplastic polyurethane is produced in Europe, the US and China.

Stainless steel

Low supply risk

Tendon cables and crimps

Sourced from: Produced widely in the EU and Asia; China is the largest producer.

FR-4 PCB laminate

Medium supply risk

Circuit boards

Sourced from: Most laminate and board production is in China and Taiwan; smaller European board makers exist.

Lithium-ion / lithium-polymer cells

High supply risk

Battery pack

Sourced from: Cell production is dominated by China, South Korea and Japan; EU cell capacity is growing but limited.

Shipping follows UN 38.3; EU battery rules add labelling and recycling duties.

Silicone elastomer

Low supply risk

Fingertip pads

Sourced from: Made from silicon metal and methyl chloride; large producers in China, the US and the EU.

Copper

Medium supply risk

Wiring, motor windings and traces

Sourced from: Mined mainly in Chile, Peru and DR Congo; Portugal mines copper at Neves-Corvo.

NdFeB magnets

High supply risk

Inside the small finger gear motors

Sourced from: China dominates rare earth processing and magnet output.

Usually bought inside complete gear motors.

Prototyping and rapid manufacturing in Portugal

5 companies from the atlas

Exatronic

Aveiro

PCB design, fabrication and SMT assembly

Mainboard and EMG sensor boards.

Proto-Electronics

Braga

Prototype PCB fabrication and assembly

Quick-turn EMG sensor board revisions.

3DWays

Oeiras

SLA and FDM printing

Fine SLA knuckle brackets and FDM PETG shell parts.

INEGI

Porto

FDM printing and engineering R&D

Shell and bracket iteration and fatigue checks.

NOVACABLE

Vila Nova de Paiva

Wire harness production

Internal harness from battery to motors and sensors.

482 Portuguese manufacturers matched

None cover the whole build — it splits across steps.