Prototyping a Custom Smart Ring: The Challenges of Flexible PCBs and Tiny Batteries

Smart rings promise continuous health tracking in a package smaller than a coin. Closing the gap between commercial products and a working DIY version is another story. Even four days of runtime inside a 2.5 mm band requires aggressive compromises on both battery capacity and board layout.

Battery Capacity vs. Ring Geometry

Commercial rings set a high bar. The Oura Ring Gen3 relies on curved LiPo cells that range from 15 mAh in the smallest sizes up to 22 mAh in the largest. The Ultrahuman Ring Air uses a single 24 mAh cell across its lineup. Both reach four-to-seven-day runtimes by combining those small packs with heavy power management on the radio and sensors.

Hobbyists run into the same physical limits. A 20 mm ring diameter leaves almost no room for anything thicker than about 4 mm. Off-the-shelf round cells that fit include the LiPol LPM0840 (8 mm diameter, 18 mAh) and LPM1040 (10 mm, 35 mAh). These parts run between 2.75 V and 4.2 V and are rated for 500+ cycles when charged at 1 C or less.

Comparison of micro LiPo cells suitable for rings

Sourcing and Integration Realities

Parts from LiPol, Benzo Energy, or PowerStream usually ship fast and need only a basic protection circuit. The real difficulty shows up once you try to install them. Most cells are rigid, so fitting one into a curved titanium or resin shell without stressing the tabs often requires custom fixturing or a flexible-tab version. Wireless charging coils also need to sit within 2 mm of the outer surface, which further squeezes the available space.

Flexible PCB Stackup Constraints

Four-layer flex looks appealing for routing sensors, the MCU, and charging circuitry, but ring geometry quickly runs into IPC-2223 bend-radius limits. A 0.15 mm stack—common in wearables—calls for a minimum static bend radius of 1.5–2 mm on single-layer boards and 3–5 mm on four-layer boards when using rolled-annealed copper.

JLCPCB and PCBWay both accept 4-layer polyimide orders. Prototypes in quantities of five to ten typically land between $50 and $200 after DFM review, with lead times of 7–12 days once the files clear. Thinner 12.5 µm polyimide with 9 µm copper improves flexibility but makes vias and traces more fragile.

Common Failure Modes Observed in Ring Designs

Medical wearable projects have surfaced the same problems again and again:

• Copper trace cracking after 200–500 flex cycles when traces run perpendicular to the bend axis
• Via barrel fractures at rigid-flex transitions from CTE mismatch
• Delamination near coverlay openings after reflow or moisture exposure
• Pad lift when components sit directly in the curved zone

Designers reduce these risks by staggering traces, using cross-hatched grounds, moving vias onto rigid stiffener areas, and keeping bend radii at 10–15× the stack thickness wherever possible.

Practical Prototyping Path

Begin with a two-layer 0.1 mm flex board to prove out the layout and firmware. Once power numbers and sensor placement look solid, move to four layers only if the shell design can accommodate a larger bend radius. Plan on $150–300 for the first 5–10 flex boards and another $30–60 per micro cell order. Run bend endurance tests on a 3D-printed mandrel before ordering final enclosures.

Key specifications to lock early

• Target runtime and sensor duty cycle
• Maximum allowable thickness (2.5–3 mm)
• Minimum bend radius the shell can support
• Acceptable charge interval (daily vs. weekly)

flowchart TD
    A[Define specs: runtime, thickness, bend radius] --> B[Build two-layer 0.1 mm flex prototype]
    B --> C[Test power draw and sensor performance]
    C --> D{Results acceptable?}
    D -->|Yes| E[Move to four-layer flex if bend radius allows]
    D -->|No| B
    E --> F[Order 5–10 boards + micro cells]
    F --> G[Bend endurance tests on 3D-printed mandrel]
    G --> H[Finalize enclosure design]

Conclusion

Tiny batteries and tight bend radii remain the two hardest constraints for custom smart rings. Commercial products succeed because they optimize every micron of the stack and accept non-replaceable cells. Hobbyists who follow IPC bend guidelines, start with two-layer flex, and source verified micro LiPo cells can reach working prototypes in a few iterations—as long as mechanical reliability gets the same attention as the electrical design.