How a fitness tracker is made
A fitness tracker is a sealed wrist device built around a Bluetooth Low Energy system-on-chip, a motion sensor, an optical heart-rate sensor, a small lithium-polymer battery and, in most designs, a display. Making one involves fine-pitch SMT assembly, often on a flexible or rigid-flex board, precision molded or machined housings, an optical window for the heart-rate sensor, an adhesive or welded seal, and a final test sequence that checks the radio, sensors, current draw and water tightness of every unit.
The difficulty is rarely a single component. It is how the parts interact in a very small volume. The optical sensor needs a clean light path through the housing, the antenna needs clearance from the battery and any metal, the seal has to survive sweat, sunscreen and drops, and battery life depends on a sleep current measured in microamps that one wrong component or firmware build can ruin. Most schedule risk sits in DVT, when parts from production tooling are tested together for the first time.
Architecture and key components
| Component | What to decide | Sourcing considerations |
|---|---|---|
| BLE system-on-chip | Memory for firmware and over-the-air updates, power modes, whether a second processor is needed | Long-term availability, reference designs, and whether an ODM’s firmware ties you to one chip family |
| Motion sensor (accelerometer, sometimes gyroscope) | Power modes, on-chip buffering, which algorithms run on it | Step and activity algorithms are tuned to a specific part, so a substitution means revalidation |
| Optical heart-rate sensor (PPG) | LED colors, number of photodiodes, analog front end, discrete parts or a module | Algorithm licensing, performance across skin tones and motion, supplier support for tuning |
| Battery | Custom pouch cell size, capacity, swelling allowance | Cell maker qualification, IEC 62133-2 and UN 38.3 reports for the exact cell, lead time for custom sizes |
| Display | PMOLED, AMOLED, LCD or memory-in-pixel, cover lens, lamination | Panel minimum order quantities and end-of-life notices, since small panels are revised often |
| Charging | Pogo pins, magnetic contacts or wireless | Contact corrosion from sweat, and the charger or cradle as a separate product with its own tooling |
| Strap | Silicone, fluoroelastomer, TPU or fabric | A separate supplier, color matching to the housing, skin-contact material compliance |
Many trackers start from an ODM platform that already combines a chip, sensors, firmware and an app SDK. That can save months, but the heart-rate and activity accuracy you claim are still your responsibility. Ask for validation data on the platform before you build your feature list around it, and agree ownership of firmware, algorithms and app code in writing. The glossary entry on ODM covers the model in more detail.
Core manufacturing processes
PCB assembly. Trackers use fine-pitch packages such as wafer-level chip-scale parts and small BGAs, often on rigid-flex boards that fold into the housing. Underfill under critical parts improves drop reliability. Depaneling and folding put stress on solder joints, so the process window matters as much as the layout.
Housings. Plastic housings are injection molded, often with two-shot or overmolded soft-touch areas. Metal housings are CNC machined and anodized, or made by metal injection molding for stainless steel. Cosmetic finish, color matching between housing and strap, and gate and parting-line placement all need agreement before tooling.
The optical window. The heart-rate sensor looks through a window on the back of the device. Light that leaks from the LEDs to the photodiodes inside the housing, rather than through the skin, degrades the signal. Light barriers, window inks matched to the sensor’s wavelengths, and tight control of window bonding are what make one build read cleanly and the next one noisy.
Sealing. Seals are made with dispensed adhesive, die-cut pressure-sensitive adhesive, gaskets or ultrasonic welding. Pressure sensors and microphones need vent membranes that pass air and block water. Adhesive cure time and fixture pressure become production parameters that have to be controlled and recorded.
Final assembly. Much of it is manual, with fixtures for battery connection, display lamination and closing the housing. ESD control, battery handling rules and cleanliness around the optical window all affect yield.
Testing and validation: EVT, DVT and PVT for a tracker
The general sequence is described in the EVT, DVT and PVT guide. For a tracker, each stage has its own focus.
EVT proves the electronics work in the form factor. Measure antenna performance with the device on a wrist or a wrist phantom, since the body detunes small antennas. Measure sleep and active current against the battery-life target, and check the heart-rate signal quality on a range of wearers, not only the engineering team.
DVT uses parts from production tooling and tests the design against its claims:
- Water resistance to the rated level, such as an IP68 rating under IEC 60529 or a watch-style rating like 5 ATM, typically referencing ISO 22810
- Exposure to artificial sweat, sunscreen, insect repellent and cleaning agents, which attack some plastics, coatings and adhesives
- Drop, tumble, button life, strap pull and charging-contact corrosion
- Temperature cycling and humidity, followed by repeat leak tests
- Heart-rate accuracy against a reference such as a chest-strap ECG, across skin tones, wrist sizes and activities, including motion
- Pre-compliance radio and EMC scans before formal certification testing
PVT runs the production line at rate. A typical end-of-line sequence covers firmware programming and serial number or address provisioning, a radio check, a sleep current measurement, sensor self-tests including the optical sensor against a reference target, an air leak test, display, touch and button checks, charging, and a cosmetic inspection. Track first-pass yield by station so you can see whether sealing, the optical path or the radio is losing units.
Certifications by market
Requirements depend on the radios, the battery, the claims and the markets. Treat the table as a starting list to confirm with a test lab, not a final scope.
| Market | Typical requirements |
|---|---|
| United States | FCC Part 15 for the radio and digital circuitry (a pre-certified module reduces testing, but many trackers are too small for one and certify a chip-down design in full); RF exposure evaluation; California Proposition 65 review for strap and housing materials |
| European Union | CE marking under the Radio Equipment Directive, including the cybersecurity requirements for wearable radio equipment that have applied since August 2025; RoHS; REACH, including nickel release limits for metal parts in prolonged skin contact; WEEE; the EU Battery Regulation |
| United Kingdom | UKCA, or CE marking, which Great Britain continues to recognize for radio equipment; the product security regime in force since April 2024, which bans universal default passwords |
| Other markets | National radio approvals, for example in Japan, Korea and Australia, often with local test reports or a local certificate holder |
| All markets | Bluetooth SIG qualification listing to use the Bluetooth name; IEC 62368-1 safety for the device and charger; IEC 62133-2 and UN 38.3 for the battery and for shipping |
Under the EU Battery Regulation, rules on user-replaceable portable batteries apply from February 2027, with exemptions that may cover some water-resistant designs. Check early whether they affect a sealed design. Health claims are the other big variable: features that move from fitness into detecting a condition can change the regulatory route entirely. The guide to wearable health sensors covers that boundary.
What drives the cost
- Silicon and sensors: the SoC, motion sensor and optical sensor, and any licensed algorithms
- Battery: custom cells cost more and carry minimum order quantities and longer lead times
- Display and cover lens: often the single largest component cost when present
- Housings and finishes: metal and multi-shot plastic parts raise both tooling and unit cost
- Test: fixtures for leak, radio, current and optical tests, and the line time each one adds
- Yield: sealing and optical failures found at end of line are expensive because the device is fully built
- Certification, charger, strap and packaging: frequently left out of early unit-cost estimates
Compare quotes on total program cost, not unit price. The quote comparison calculator adds one-off costs such as NRE and tooling at your volume.
What to ask an OEM/ODM partner
- Which sealed wearables with optical sensors are in production with you now, and may we see the end-of-line test stations for one of them?
- How do you leak test every unit, what is the pass limit, and how was it correlated with immersion tests?
- How is sleep current measured on the line, and what limit do you use?
- Who owns the firmware, heart-rate algorithm and app SDK, and what is licensed rather than transferred?
- Who makes the battery cell, and can we see IEC 62133-2 and UN 38.3 reports for that exact cell?
- What first-pass yield do you see on comparable products, and what are the top three failure modes?
- How do you control strap material and color lots against the housing?
Common mistakes
- Leaving heart-rate accuracy to late firmware tuning, when much of it is fixed by the mechanical and optical design
- Validating water resistance only on hand-built EVT units, then discovering seal variation at volume
- Squeezing battery capacity late in design without leaving room for cell swelling
- Leaving the charger or cradle until after the tracker’s design is frozen
- Approving a component substitution without rerunning current and sensor tests
- Skipping chemical exposure tests, so the first sunscreen complaints arrive after launch
Where Zefinity helps
Fitness trackers reward an early, specific conversation with the right partner: one that has built sealed optical wearables, can show its test stations and will be clear about what its platform does and does not include. Our Connect service identifies and qualifies OEM/ODM partners against those requirements, runs the RFQ and compares quotes on a like-for-like basis. The questions we raise across the category are on the wearables page.
Frequently asked questions
Should a fitness tracker start from an ODM platform or a custom design?
An ODM platform can shorten development when its electronics, sensors and firmware already match your feature set, and you mainly change the housing, strap and app. A custom design makes sense when your product depends on a sensor, form factor or battery life the platform cannot reach. Either way, confirm who owns the firmware and algorithms, and what you can change later.
How is water resistance tested in fitness tracker production?
Usually with an air leak test rather than immersion. A fixture seals around the device, applies pressure or vacuum and measures pressure decay against a limit. That limit has to be correlated in DVT with immersion and pressure tests on units with known leaks, otherwise the line test can pass devices that fail in a pool.
Does a fitness tracker need medical device approval?
Not if it makes only general fitness and wellness claims, such as steps, activity and heart rate during exercise. Claims about detecting, diagnosing or managing a disease or condition generally move a product into medical device regulation in the US and EU. Decide the claims before design freeze and get regulatory advice on the wording.