Fitness Band / Activity Tracker (OEM)

Heart Rate Accuracy vs. Chest Strap Gold Standard

Optical wrist PPG is the standard technology in fitness bands. Accuracy is acceptable for lifestyle tracking but falls short of medical-grade cardiac monitoring. When evaluating factory samples:

Resting HR accuracy. At complete rest (5-minute seated test), a well-tuned optical PPG algorithm achieves ±2–3 BPM vs. a Polar H10 chest strap reference. Poorly tuned firmware may show ±10–15 BPM error. Test at least 10 subjects to account for skin tone variation — PPG algorithms that are calibrated only on light skin tones show systematic positive bias on darker skin tones (a known limitation that Apple, Fitbit, and Samsung have all publicly acknowledged).

Active HR accuracy. During walking and moderate-intensity running, wrist PPG error increases due to motion artifact. Expect ±5–10 BPM at 8 km/h for a well-tuned product. Algorithms using accelerometer data to suppress motion artifact (typical in Bosch BMI160 or InvenSense ICM-40605 combined HR+accel implementations) perform better than PPG-only solutions.

The factory’s stated “HR accuracy ±3 BPM” almost always refers to resting measurement only. Request the active HR test protocol and results.

Step Counter Algorithm Tuning

Step counting algorithms convert accelerometer data into step counts. Factory default calibration is often optimized for one body size and gait pattern:

Over-counting. Wrist movements that are not steps (typing, stirring, driving) trigger false step counts in poorly calibrated algorithms. For a 10,000-step-per-day target, a 20% over-count adds 2,000 phantom steps — statistically significant for health monitoring applications.

Under-counting. Slow walking and shuffling gaits (common in elderly users) are missed by threshold-based algorithms that require a minimum acceleration magnitude.

Test procedure. Walk a known distance (1 km on a track) at normal pace, fast pace, and slow pace. Count actual steps by manual counting a 100-step sample. Acceptable accuracy: ±5% at normal pace, ±10% at slow pace.

Tuning the step counter algorithm requires firmware-level changes. Confirm the factory will re-tune the algorithm for your target demographic if initial sample testing shows systematic errors.

OEM App Integration: HealthKit, Google Fit, and SDK Handoff

Apple HealthKit. Writing workout and health data to Apple HealthKit requires NSHealthUpdateUsageDescription permission in the iOS app Info.plist and uses HealthKit APIs (HKWorkoutBuilder, HKQuantitySample). Not all OEM app SDKs include HealthKit integration — if your target market is iOS users who use Apple Health, this must be specified and verified in the factory’s app SDK before signing.

Google Fit (now Health Connect). Google’s Health Connect API (replacing the older Google Fit REST API) requires specific data type declarations in the AndroidManifest. Confirm the factory’s Android SDK uses Health Connect rather than the deprecated Fit API for new projects.

White-label app ownership. The white-label companion app provided by the factory typically uses a shared backend (the factory’s cloud server) for data sync. Confirm:

• Where user health data is stored (China server vs. EU/US server) — critical for GDPR compliance
• Whether you receive a private backend deployment or a shared multi-tenant backend
• What happens to user data if you switch factories or discontinue the app

For GDPR-regulated EU markets, data residency and processor agreements are mandatory for health data.

Silicone Strap Skin Allergy Compliance: REACH SVHC

Silicone straps in contact with skin must comply with REACH SVHC (Substances of Very High Concern) regulations in EU markets. The relevant tests:

Nickel release test (EN 1811). Metal components (buckles, watch lugs) — not the silicone strap itself — are tested for nickel release. Rate <0.2 µg/cm²/week required for items in prolonged skin contact.

Polycyclic aromatic hydrocarbons (PAH, EN 17122). Relevant for rubber compounds — some TPU strap formulations use rubber processing aids that contain PAH. Limit: <0.5 mg/kg for 16 PAH sum.

4-(dimethylamino)benzaldehyde (DMAB) test. Specific to TPU materials — tests for aromatic amine release from azo dyes. Azo dye restriction applies to colored straps.

Request REACH SVHC screening test reports for the strap material (silicone or TPU) from the factory. These should be issued by a third-party lab (SGS, Bureau Veritas, TÜV). A factory-issued self-declaration is not sufficient for EU market compliance.