nRF52840-Based Wearable Stress Monitoring Cuff: Design & Development

We developed our sensor-enabled stress management wrist cuff based on the nRF52840 Bluetooth® Low Energy platform from Nordic Semiconductor as part of this case study to show the integration of biosensors and low power wireless communications with embedded firmware resulting in compact, reliable wearable devices that accurately monitor physiological signals and assess stress levels in real time.

This allows users to take proactive steps towards a healthier lifestyle.

Client: Confidential
Role: Lead Embedded Systems Engineer
Timeline: 10 weeks (from initial concept to completed working prototype)
Deliverables: 3 functional prototypes, firmware source code, BOM, schematic diagrams, data logging interface

The customer needed a prototype product to check if a haptic method has an automatic way of delivering information about whether it is working well, is normal and is safe. The basic requirements for the cuff are that it is a bracelet cuff that can measure a person's stress levels, through basic physiology, and it is capable of automatically creating haptic cues for breathing and other techniques without AI, using only rules-based logic.

Critical requirements:

• PPG sensor for HR/HRV monitoring
• IMU for motion artifact rejection
• DRV2605-driven LRA motor for precise haptic patterns
• Bluetooth Low Energy (BLE) for data logging
• Rechargeable battery power
• Rule-based detection: HR↑ + HRV↓ + low motion for 60-120 seconds → trigger
• Haptic pattern: alert pulse + 4s inhale ramp + 0.5s pause + 4s exhale pattern

Main Microcontroller Selection

After evaluating available platforms, I selected the Seeed XIAO nRF52840 Sense as the core MCU.

Requirement	XIAO nRF52840 Advantage
Form factor	21mm × 17.5mm — fits inside wrist cuff
Processor	nRF52840 Cortex-M4F at 64MHz
BLE	Native Bluetooth 5.0 support
IMU	Built-in LSM6DS3TR-C (3D accelerometer + gyroscope)
Microphone	Built-in PDM mic for future expansion
Battery charger	Built-in BQ25101 charger IC
GPIO	14 pins for PPG and haptic driver

Through one board, the design accomplished two goals: removing separate circuits for the charger and IMU, freeing up valuable physical and development space.

PPG Sensor Selection

In selecting a sensor solution to allow both HR and HRV readings from wrist wear — a location prone to motion artifacts — the MAX30102 was the selected option due to the following:

• Provides integrated photodetector, red LED, and IR LED into a package that measures only 5.6mm × 3.3mm
• Various programmability options for sample rate and LED output current; and
• Provides the ability to reject ambient light on the chip; and
• Has an available open-source method to calculate HRV.

Haptic Driver and Actuator

The client specified an LRA (Linear Resonant Actuator) with DRV2605 driver for precise waveform control.

Selected: DRV2605 + LRA motor combo

Why DRV2605:

• Dedicated haptic driver with I²C interface
• Built-in ROM library of haptic effects
• Automatic LRA resonance tracking and calibration
• Programmable waveform sequencing for 4-second inhale/4-second exhale patterns

The DRV2605 requires proper calibration with the specific LRA motor. Critical parameters include:

• DRV2605_RATED_VOLTAGE: The motor's rated operating voltage
• DRV2605_OD_CLAMP: Overdrive voltage clamp limit
• DRV2605_DRIVE_TIME: LRA drive time based on motor resonance frequency

Selecting a Battery

Estimating Required Battery Capacity

• MAX30102 (Active Sensing) - 15mA
• nRF52840 (BLE and Processing) - 10mA
• DRV2605 (Active Haptic Feedback) - 100mA
• Average Usage: 30mA

Battery Selection: 601225 Lithium Polymer 150mAh, 3.7V

Why:

• Fits Dimensions (6.0±0.2 × 12±0.5 × 25±2.5 mm) to fit into Cuff Design
• 0.5C-1C Discharge for Haptic Power Max
• Compatible with Built-in Protection Circuit/Module
• Provides 300 to 500 Cycles in Lifetime.

Runtime calculation: 150mAh / 30mA = 5 hours continuous. With intermittent haptic activation, full-day wear is achievable.

Block Diagram

(Block diagram placeholder)

Pin Mapping

XIAO nRF52840 Pin	Connected To	Function
P0.26 (SDA)	MAX30102 SDA, DRV2605 SDA	I²C Data
P0.27 (SCL)	MAX30102 SCL, DRV2605 SCL	I²C Clock
P0.13	DRV2605 EN (Enable)	Haptic driver enable
P1.10	DRV2605 TRIG (Trigger)	Waveform trigger
VBAT	Battery (through charger)	Power input
3.3V	MAX30102 VCC	Sensor power
GND	Common ground	Ground reference

Firmware is developed as well as implemented through state machine detection logic wherein four states are used for stress detection via a rule based approach. The first five minutes of wear establish the baseline heart rate (HR) and heart rate variability (HRV) through a rolling average with outlier rejection.

Breathing guidance prototyping and pattern design requires precision in timing (using the following duration):

• 500ms alert
• 4 seconds ramp
• 0.5 seconds halt
• 4 seconds decremented
• Repeat for one to five minutes

The DRV2605 waveform sequencer has custom PWM profiles programmed to run through the inhale/pause/exhale sequences for repeat cycles that last between one to five minutes according to specified timing.

BLE Data Logging Interface

A custom BLE GATT service was implemented for real-time data streaming:

• Characteristic: Stress Metrics (UUID: 12345678-1234-5678-1234-56789abcdef0)
• Properties: Notify
• Data format (8 bytes):
• Bytes 0-1: Heart rate (BPM, uint16)
• Bytes 2-3: HRV (SDNN, milliseconds × 10, uint16)
• Byte 4: Motion magnitude (uint8, 0-255 scaled)
• Byte 5: Trigger flag (0/1)
• Bytes 6-7: Timestamp (ms window, uint16)

A simple React Native app paired with the device to:

• Display real-time HR, HRV, and motion metrics
• Log trigger events with pre/post HRV values
• Allow user feedback (helpful/not helpful/neutral)
• Export session data as CSV

DA103-5 (Wrist Cuff) Components

The client has provided a specification for a "wrist cuff" (not just a watch-style product) to be wider (30-40 mm) so it fits next to the wrist area.

The component placement strategy is as follows:

1. Center the XIAO nRF52840 module on the underside of the cuff (to have direct contact with skin for charging; IMU will work through fabric).
2. Place the 601225 battery adjacent to the MCU in parallel.
3. Place the MAX30102 on the inside of the cuff (will be pressed against the wrist to provide good optical contact; required for good quality PPG signal).
4. Mount the DRV2605 and LRA on the top of the cuff (vibration will transmit through fabric).
5. Encapsulate all components in either a silicone mold or sew into a fabric pocket.

The cuff will use stretch fabric and Velcro for adjustment.

Functional Test Results (Week 8-10)

Test	Target	Result
HR accuracy (resting)	±5 BPM (vs chest strap)	±3 BPM at rest, ±8 BPM during motion
HRV detection	SDNN calculation stable	Stable at 30-second rolling window
Motion artifact rejection	False trigger rate <10%	12% initial, tuned to 8% with 90-second confirmation window
Haptic pattern perception	4s inhale/4s exhale distinguishable	Users correctly identified pattern 92% of time
Battery life (continuous sensing)	4 hours minimum	5.2 hours (150mAh, 28mA average)
BLE range	10 meters	15 meters line-of-sight

User-Focused Testing of the User Experience (5 participants)

Protocol:

• Two hours of working (email and meetings)
• One hour of performing a cognitive stress test (mental math with time pressure)
• Collection of feedback about the experience after testing

Results:

• Four of the five respondents reported the haptic intervention was either "calming" or "noticeable but mild".
• Three of the five participants could accurately recreate the breathing instructional pattern as provided.
• All five agreed the haptic trigger is sufficiently reactive (defining at most a 90-120 second threshold) after a stress reaction occurs.

Deliverable	Format	Status
3 working prototypes	Assembled, tested wrist cuffs	Delivered
Firmware source code	C++ (Arduino framework), annotated	Delivered
Full BOM with supplier part numbers	Excel, 32 line items	Delivered
Wiring/schematic documentation	PDF, Fritzing diagrams	Delivered
Basic data logging (BLE + companion app)	React Native source + APK	Delivered
Haptic patterns adjustable in firmware	#define parameters in config header	Delivered

Challenge	Solution
PPG signal dropout during wrist movement	Integrated IMU gating — discard HR readings during high motion
LRA resonance frequency mismatch	Used DRV2605 auto-calibration routine before each session
Battery drain from continuous BLE advertising	Implemented connection-only advertising (advertise 30s, disconnect after 5 min idle)
False triggers from typing (wrist motion while stressed)	Increased motion confirmation window to 30 seconds consecutive low motion