ST STEVAL-BCN002V1B User manual

Introduction
The STEVAL-BCN002V1B Bluetooth LE enabled sensor node development kit features the STEVAL-BCN002V1 multi-sensor
board based on BlueNRG-2 SoC Bluetooth Low Energy application processor. This sensor board has accelerometer,
gyroscope, magnetometer, pressure, humidity, Time-of-Flight and microphone sensors, and is powered by a common CR2032
coin battery.
The sensor board communicates with a Bluetooth LE enabled smartphone running the ST BLE Sensor app, available on Google
Play and iTunes stores.
The STEVAL-BCN002V1D adapter board is used to program and debug the sensor board. The adapter board is powered via
USB.
Figure 1. STEVAL-BCN002V1B BlueTile kit
1. STEVAL-BCN002V1 “BlueTile” sensor node with inertial and environmental digital MEMS sensors, a digital MEMS microphone, a time-of-
flight proximity sensor and a Bluetooth 5.0 wireless system-on-chip with a Cortex-M0 core
2. STEVAL-BCN002V1D host board to Flash and debug the sensor node
2
1
How to use the BlueNRG-Tile Bluetooth LE enabled sensor node development kit
UM2501
User manual
UM2501 - Rev 3 - June 2020
For further information contact your local STMicroelectronics sales office.
www.st.com

1Safety Information
Any type of usage not specified by the manufacturer may compromise the protection mechanisms in the device.
1.1 Class 1 laser product
VL53L1X contains a laser emitter; the device is designed to limit the laser output within Class 1 laser safety limits
under all conditions including single faults, in compliance with IEC 60825-1:2014 (third edition).
Recommended settings and operating conditions specified in VL53L1X datasheet and user manual MUST be
respected.
The laser output MUST NOT be increased by any means and no optics should be to used focus the laser beam.
Use of controls or adjustments or performance of procedures other than those specified in VL53L1X datasheet
and user manual may result in hazardous radiation exposure.
UM2501
Safety Information
UM2501 - Rev 3 page 2/44

2STEVAL-BCN002V1 BlueTile sensor node
BlueTile is a small form-factor reference design ready for you to extend and customize, as well as a development
platform for BLE-enabled applications using the BlueNRG-2 SDK. It is also an evaluation tool to help you assess
and evaluate the performance of the sensors and the capabilities of the BLE.
Thanks to its small and thin design, BlueTile can be used in the field for data collection activities to test and
develop new algorithms.
Figure 2. STEVAL-BCN002V1 BlueTile block diagram
BlueNRG-2: Bluetooth 5.0 network and application processor
LSM6DSO: accelerometer and gyroscope
LIS2MDL: magnetometer
LPS22HH: barometer
HTS221: relative humidity and temperature sensor
VL53L1X: time-of-flight proximity sensor
MP34DT05-A: microphone
BALF-NRG-02D3: integrated balun
I2C
LPS22HH
RGB LED
user button
UART
MP34DT05-A
LIS2MDL
32 MHz
I2C
I2C
VL53L1
400
kHz
1.6
I2C
Test points
TEST
VDDGND
MHz
RFTEST
PDM
inductor
10 pin
connector
32 KHz
Integrated
antenna
LSM6DSO
UART_RX line
I2C
BALF-NRG-02D3
HTS221
I2C
TEST ADC
ADC
RFTEST
SWD
SDASCL
IRQ
IRQ
I2C
3 VOLTS
+
BlueNRG-2
d
Pi-network
• The integrated SMD antenna needs clearance area and passives for proper tuning (FT1, FT2 and MT).
•The Pi-network is only intended for flexibility and for testing. It is not populated as the integrated balun
provides the necessary matching.
• The 32kHz crystal allows lower power consumption BLE Sleep Mode.
• The inductor allows lower power consumption in BLE Active Mode.
• The STEVAL-BCN002V1 BlueTile sensor node is supplied with the default firmware (BLE_SensorDemo,
available in the SDK) already loaded. The firmware enables the streaming of sensor data to the reference
smartphone app (ST BlueMS, available on Android™ and iOS™ stores).
UM2501
STEVAL-BCN002V1 BlueTile sensor node
UM2501 - Rev 3 page 3/44

Figure 3. STEVAL-BCN002V1 sensor node front and rear components
1a. BlueNRG-2 Bluetooth 5.0 network and application processor 1b. BALF-NRG-02D3 integrated balun and matching network
2a. LPS22HH ambient pressure sensor 2b. LSM6DSO smart accelerometer and gyroscope 2c. LIS2MDL magnetometer 2d.
VL53L1X proximity by time-of-flight 2e. HTS221 relative humidity and temperature 2f. MP34DT05-A top-port digital microphone
3a. User button 3b. RGB LED
4a. I2C SCL 4b. I2C SDA 4c. GND 4d. ADC 4e. TEST 4f. VDD 4g. GND 4h. RFTEST
5a. II-network (not mounted) 5b. SMD antenna
6a. Inductor to enable lowest power BLE active mode 6b. 32kHz crystal to enable lowest power BLE sleep mode
7a. Battery holder 7b. CR2032 battery (not included)
8. 10-pin connector
FCC ID: S9NSTEBCN2V1 IC ID: 8976C-STEBCN2V1
1a
2a
4a
5a
6a
2b
2c
2d
2e
2f
3a
3b
4b4c 4d
4e
4f
4g
4h
1b
5b
6b
7a
7b
8
RESETN
UART RX
UART TX
SWD CLK
SWD IO
IO IO IO IO
GNDBOOT GND GND VDD
1
2
10
The BlueTile sensor node is supplied with the firmware already loaded. The default firmware includes two
libraries:
1. MotionFX fusion library for orientation estimation (visualized with a rotating cube).
2. BlueVoice library to stream the voice captured by the MEMS microphone over the BLE connection
(reproduced by the smartphone speaker, if not in vibration mode).
The STBLESensor app can be used to plot and log sensor data in real time on your smartphone.
RELATED LINKS
STBLE sensor documentation
2.1 System architecture
The BlueTile is powered by a coin-cell battery (CR2032). The voltage is not regulated because the sensors and
the BlueNRG-2 can all operate at the voltage range of the battery. In any case, the BlueNRG-2 has its own
embedded linear power regulator and switching mode power converter (DC-DC).
2.1.1 Power section for BlueNRG-2
The BlueNRG-2 in active mode can use the embedded linear voltage regulator (LDO) or the embedded DC-DC
converter (DC-DC).
•LDO regulator [Active mode, CPU Flash and RAM on]:
– RX 14.5 mA at 3 V, TX 17.2 mA when output power is +2 dBm
– range is 12-28.8 mA when output is -14 to +8 dBm
• DC-DC converter [Active mode, CPU Flash and RAM on]:
– RX 7.7 mA at 3 V, TX 9 mA at 3 V when output power is +2 dBm
– range is 6.6-15.1 mA when output is -14 to +8 dBm
UM2501
System architecture
UM2501 - Rev 3 page 4/44

Note: The BlueTile sensor node includes the inductor needed by the DC-DC converter to allow lower power
consumption in active mode.
In Sleep Mode, the BlueNRG-2 device can use its internal 32 kHz ring oscillator (RO) or the external 32 kHz
crystal oscillator (XO) available on the BlueTile sensor node, which offers lower power consumption in sleep
mode:
• Sleep Mode [32 kHz ring oscillator, 24 kB RAM retention]: 2.1 µA at 3 V
• Sleep Mode [32 kHz crystal oscillator, 24 kB RAM retention]: 0.9 µA at 3 V
2.1.2 Radio frequency section
The radio frequency section of the BlueTile includes the following elements:
1. A BALF-NRG-02D3 ultra-miniature balun which integrates matching network and harmonics filter.
2. A Pi-network which allows additional filtering and provides access points for testing.
Note: This network is not populated, as the integrated balun provides the necessary matching.
3. An SMD 2.4 GHz antenna, which requires a certain clearance area on the PCB and specific passives for
precise tuning (FT1, FT2 and MT).
2.1.3 MEMS sensor section
The MEMS sensor section of the BlueTile sensor node includes inertial and environmental MEMS sensors
connected with the BlueNRG-2 via an I2C bus operating at 400 kHz.
The MEMS microphone is connected with BlueNRG-2 through a dedicated line to transfer the PDM (Pulse
Density Modulated) stream at 1.6 MHz, which is converted to PCM (Pulse Coded Modulation) by the integrated
digital filter in the ADC block on the BlueNRG-2.
All sensors can generate interrupts, but only the interrupts from the LIS2MDL magnetometer and the LSM6DSO
accelerometer and gyroscope are connected with the BlueNRG-2 through dedicated and independent lines.
•The LIS2MDL magnetometer interrupt line is push-pull only and must have its own independent line. The
interrupt pin of this sensor is connected with BlueNRG-2 to enable applications such as “reed-switch”, which
is activated by an external magnet.
• The LSM6DSO smart accelerometer and gyroscope interrupt pin is an input at boot. It has an internal pull-
down. If the pin is low at boot, the I2C interface is selected, otherwise the I3C interface is selected. The pin
can be reconfigured as open-drain, but the external pull-up would cause additional power consumption. The
interrupt pin of this sensor is connected to BlueNRG-2 to exploit the smart embedded processing
functionality (single/double tap recognition, free-fall, wakeup, portrait/landscape, 6D/4D orientation detection;
activity/inactivity, stationary/motion detection; pedometer, step detector and step counter; up to 16 finite state
machines to process).
• The LPS22HH barometer interrupt pin exhibits the same behavior as the LSM6DSO interrupt pin: it is an
input pin at boot and must be low to activate the I2C interface. It can be configured as push-pull or open-
drain after boot. The interrupt pin of this sensor is not connected to BlueNRG-2.
• The HTS221 relative humidity and temperature interrupt pin is push-pull at boot and can be reconfigured as
open-drain after boot. The interrupt pin of this sensor is not connected to BlueNRG-2.
• The VL53L1X time-of-flight proximity interrupt line is open-drain only and would require an external pull-up.
The interrupt pin of this sensor is not connected to BlueNRG-2.
The dynamics of the environmental parameters are relatively slow, so interrupts are not necessary. The
application only needs to wake up every few seconds and perform one-shot measurements with the LPS22HH
barometer or HTS221 relative humidity and temperature sensor, and trigger an action if specific conditions are
met.
It is also not convenient to keep the VL53L1X proximity time-of-flight sensor active, the power consumption or the
latency would be too high. The power consumption of VL53L1X ranges from 0.5 mA for 1 Hz measurements, up
to 7 mA for 10 Hz measurements.
2.1.4 Pinout descriptions
In the following figure, you can see 6 pads along the edge of the board:
•RFTEST (4h) and GND (4g) connected to the RF section for test purposes.
• TEST (4e) connected to the analog TEST output of BlueNRG-2 for test purposes.
• VDD (4f) and GND (4c), for test purposes or to power additional external components.
• ADC (4d) to accept analog input signals from additional external components.
UM2501
System architecture
UM2501 - Rev 3 page 5/44

• I2C SCL (4a) and I2C SDA (4b) to connect additional external components via I2C.
Figure 4. STEVAL-BCN002V1 sensor node front and rear components
1a. BlueNRG-2 Bluetooth 5.0 network and application processor 1b. BALF-NRG-02D3 integrated balun and matching network
2a. LPS22HH ambient pressure sensor 2b. LSM6DSO smart accelerometer and gyroscope 2c. LIS2MDL magnetometer 2d.
VL53L1X proximity by time-of-flight 2e. HTS221 relative humidity and temperature 2f. MP34DT05-A top-port digital microphone
3a. User button 3b. RGB LED
4a. I2C SCL 4b. I2C SDA 4c. GND 4d. ADC 4e. TEST 4f. VDD 4g. GND 4h. RFTEST
5a. II-network (not mounted) 5b. SMD antenna
6a. Inductor to enable lowest power BLE active mode 6b. 32kHz crystal to enable lowest power BLE sleep mode
7a. Battery holder 7b. CR2032 battery (not included)
8. 10-pin connector
FCC ID: S9NSTEBCN2V1 IC ID: 8976C-STEBCN2V1
1a
2a
4a
5a
6a
2b
2c
2d
2e
2f
3a
3b
4b4c 4d
4e
4f
4g
4h
1b
5b
6b
7a
7b
8
RESETN
UART RX
UART TX
SWD CLK
SWD IO
IO IO IO IO
GNDBOOT GND GND VDD
1
2
10
On the debug connector on the rear of the board, there are 10 pins:
•VDD (pin 2) and GND (pin 4, 6, 10) to accept power from the host motherboard.
• SWD IO (pin 1) and SWD CLK (pin 3) to Flash and debug through the host motherboard.
– SWD pins can be used as digital GPIOs when the BlueTile is not connected to its host board.
• UART TX (pin 5) and UART RX (pin 7).
– UART pins can be used as digital GPIOs when the BlueTile is not connected to its host board.
• BOOT (pin 8) and RESETN (pin 9) to control the boot and keep BlueNRG-2 under reset.
– The boot pin can also be used as a digital GPIO when the BlueTile is not connected to its host board,
but it needs to be pull-down at start-up.
2.2 Features of the BlueNRG-2 device
BlueNRG-2 integrates a Bluetooth Low Energy radio (BLE), an ARM® Cortex®-M0 core, 12+12 kB of RAM,
256 kB of Flash memory, SPI (max 1 MHz in slave mode, 8 MHz in master mode), two I2Cs (standard 100 kHz or
fast 400 kHz), UART interfaces; two multi-function timers (MFT), a DMA controller, RTC and watchdog, and an
ADC with PDM stream processor.
The public key cryptography (PKA) and random number generator (RNG) are reserved for the BLE protocol stack,
but the user application can also read the RNG.
The ADC is 10-bit and features single or continuous acquisition at max. 1 MHz sampling frequency, two single
ended or one differential signal (ADC1 and ADC2 pins), embedded channels for temperature and battery voltage
sensing, embedded digital filter with down sampling. The embedded digital filter can be used to process the PDM
stream from a digital MEMS microphone (1.6 MHz or 0.8 MHz) and convert it to audio PCM (8 kHz to 50 kHz
when 1.6 MHz clock is used for the microphone).
BlueNRG-2 uses the embedded digital filter of the ADC peripheral for PDM to PCM conversion when reading the
PDM output of the MEMS microphone MP34DT05-A.
UM2501
Features of the BlueNRG-2 device
UM2501 - Rev 3 page 6/44

The low-speed clock is used in low-power mode and can be supplied by the internal RC oscillator or by an
external crystal (32 kHz ±50 ppm). The high-speed clock is supplied by a fast-starting internal RC oscillator
(16 MHz) while the external crystal is starting up. The high-speed external crystal (16 or 32 MHz) is required to
enable the BLE radio.
Only the 32 MHz XO can support the highest computational loads, allowing the Cortex-M0, the DMA and the APB
tree to run at 32 MHz, while the rest of the clock tree runs at 16 MHz. When a 16 MHz XO is used, the Cortex-M0
and all the clock tree runs at 16 MHz.
2.2.1 BlueNRG-2 states
• Preactive:
–After power-on-reset, all digital power supplies are stable.
– Internal 16 MHz and 32 kHz RC oscillators are used.
– Reachable from Reset, Standby and Sleep states.
– Goes to Active state.
• Active:
– The external high-frequency crystal is used (16 MHz ±50ppm or 32 MHz) to enable BLE
communication.
– Internal 16MHz RC oscillator is switched off
–BlueNRG-2 uses a 32 MHz ±10ppm crystal (higher crystal accuracy = lower power consumption).
– The radio can be activated for transmission (TX) or reception (RX). This is the state used by
BlueNRG-2 when the user application is running or there is a BLE event to serve.
– The programmed GPIO configuration is restored.
– reachable only from Preactive only.
• Standby:
– RAM retention is used (12 or 24 kB).
– 5 different GPIOs can be used to wake-up: IO9, IO10, IO11 with an internal pull-up; IO12, IO13 which
require an external drive).
–BlueNRG-2 IO11 is connected to the user button, IO12 connected to the interrupt pin of LIS2MDL
magnetometer, and IO13 connected to the interrupt pin of the LSM6DSO accelerometer and
gyroscope.
– Wake-up time is typically 200µs.
– reachable from Active.
– can go to Preactive.
• Sleep:
– RAM retention is used (12 or 24 kB).
– The low-frequency oscillator is switched on to serve periodic BLE connection events.
– 5 different GPIOs can also be used to wake-up, as in Standby state.
– This is the state used by the BlueNRG-2 tile to save power when the user application is not running
and the device is waiting for a BLE event to serve.
– The programmed GPIO configuration is not maintained (this may affect connected sensors; for
example, see pin IO7 of BlueNRG-2 connected to XSHUT pin of VL53L1X).
– Reachable from Active.
– Can go to Preactive.
2.2.2 BLE firmware stack and memory map
Program memory, data memory, registers and I/O ports are organized in a linear 32-bit address space.
•The SRAM is divided in two banks:
– 12 KB from 0x2000_0000 to 0x2000_2FFF (retention always-on).
– 12 KB from 0x2000_3000 to 0x2000_5FFF (retention optional).
– The FULL BLE stack needs 9.6 KB of SRAM.
– The BASIC BLE stack needs 8.0 KB of SRAM.
UM2501
Features of the BlueNRG-2 device
UM2501 - Rev 3 page 7/44

• The Flash is 256 KB from 0x1004_0000 to 0x1007_FFFF
–The FULL BLE stack needs 77 KB, leaving 179 KB for the user application. The FULL BLE stack
supports concurrent peripheral and central roles (N=0,1,2 connections to other centrals and 8-N
connections to other peripherals), LE secure connections, controller privacy, and extended data length.
– The BASIC BLE stack needs 58 KB, leaving 198 KB for the user application. The BASIC BLE supports
the peripheral role only (1 connection to a central), legacy security only, no controller privacy and no
extended data length.
– The OTA manager for over-the-air firmware updates adds 10 KB to the size of the BLE stack.
The OTA manager may be independent of the user application, or embedded in the user application:
• When the OTA manager is independent of the user application, the Flash is divided in two sections:
1. A relatively small section for the OTA application; the size is the sum of the BLE stack size and the
OTA manager size.
2. A larger section for the user application.
• When the OTA manager is embedded in the user application, the Flash is divided in three sections:
1. A section that holds the current application; the size is the sum of BLE stack size, OTA manager size
and user application size.
2. A section for the new updated application; the size is the same as the first section.
3. A very small section for the reset application that decides decides which application to run after boot,
depending on what is available and valid in the first two sections.
An independent OTA manager leaves more room for the user application, but there is no application to run if the
update does not finish successfully.
An embedded OTA manager can still use the previous application if the update does terminate successfully, but
the space available for the user application is less than half the size of the Flash.
Note: After a successful update, the new application can be copied over the old application, or the role of the first two
sections can be exchanged.
2.3 Inertial MEMS sensors
Each sensor is made of a Micro Electro Mechanical system (MEMS) with a sensing element and a dedicated
ASIC with analog acquisition chain, analog-to-digital converter (ADC), and digital signal processing (DSP) and
control logic.
2.3.1 LSM6DSO 3D accelerometer and 3D gyroscope
2.3.1.1 LSM6DSO interrupts
The LSM6DSO delivers best-in-class motion sensing able to support more sophisticated functionality than simply
orienting devices to portrait and landscape mode.
The following event-detection interrupts are supported:
•Free-fall event:
– detected when acceleration data from all enabled axes are below threshold settings for a certain
duration
• Wake-up event:
– detected when the filtered data from any of the enabled axes are above threshold settings for a certain
duration
– choose between high-pass filtered data, slope detection filter (difference between consecutive
samples, divided by two), low-pass filtered data plus an offset for each axis, or unfiltered data plus an
offset.
• 6D/4D orientation event:
– detected when an axis (positive or negative) is above the threshold setting, while the other two are
below the threshold for a certain duration
– 4D is a subset of 6D; the Z axis is not used.
– choose between unfiltered data or low-pass filtered data
UM2501
Inertial MEMS sensors
UM2501 - Rev 3 page 8/44

• Single/double tap event:
–detected when the output of the slope detection filter exceeds the threshold setting and then returns
below the same setting within a “shock” time interval
– double tap event is detected when a first tap is detected and a second tap is detected after a “quiet”
time interval, but before a maximum “duration” interval
– for reliable detection, configure the device to use a high sampling rate, like 400 Hz.
– if two or more enabled axes are over their respective thresholds, the axis with the highest priority
setting is used
– can be used to enable user interaction in addition to the button present on the BlueTile
• Activity/inactivity event:
– uses the same data as the wake-up event
– if inactivity is detected (data below the threshold setting for a certain time interval), the device
automatically goes to Low-power Mode and reduces the accelerometer sampling rate down to 12.5 Hz
to minimize power consumption
– if a wake-up event is detected, the device automatically returns to the programmed accelerometer
operating mode and sampling rate
– can be configured to put the gyroscope in power-down or sleep mode when the accelerometer is in
low-power mode
• Stationary/motion event
• a special activity/inactivity event in which event detection is the same, but the device does not change power
mode or sampling rate
2.3.1.2 LSM6DSO advanced functionality
Certain digital processing blocks allow advanced functions and algorithms when the output data rate is 26 Hz or
higher.
• Pedometer functions: step detector and step counter.
– when a step is detected, an interrupt is generated and a corresponding counter is incremented by 1
(max value is 65535)
– can also set the interrupt to only be triggered when at least one step is detected in a configurable time
interval
– de-bounce functionality to avoid false detections, where N consecutive steps are detected before the
first interrupt is generated (default is N=10)
– two additional blocks can facilitate the identification of false-positives:
only trigger when statistical data matches the walking pattern
adapt the embedded algorithm to slow pace walking patterns
• Significant motion detection: to signal a possible change in user location
– based on the pedometer function, which is automatically enabled
– interrupt is generated when the difference between the number of steps after initialization or reset is
higher than the de-bounce threshold (default is 10).
– function is reset after each interrupt
• Relative tilt detection
– generates an interrupt when there a change in the tilt of the device exceeds 35 degrees
– reference position is set when the detection logic is enabled and reset when the interrupt is generated
– detection logic requires a 2-second settling time after it is enabled or re-initialized
2.3.1.3 LSM6DSO finite state machines
The LSM6DSO can run up to 16 simultaneous finite state machines (FSMs) that consist of a sequence of
instructions or actions (parameter setting, interrupt generation, etc.), or a RESET-NEXT condition:
• if the RESET condition is satisfied, the FSM returns to the reset point
• if the NEXT condition is satisfied, the FSM moves to the next instruction
• otherwise, the RESET-NEXT conditions are re-tested.
FSM data may be accelerometer, gyroscope, integrated gyroscope or external sensor data, and can be
decimated. A data vector consists of X, Y, Z values and a vector norm V (square root of sum of squares).
UM2501
Inertial MEMS sensors
UM2501 - Rev 3 page 9/44

Each FSM has the following features:
•3 different 8-bit masks to allow tests on positive and negative values of X, Y, Z and V.
• 3 different thresholds and 1 programmable hysteresis value that is automatically added to or subtracted from
the selected threshold based on the test condition.
• 4 different counters/timers to manage event durations.
2.3.1.4 LSM6DSO smart FIFO buffer
The LSM6DSO accelerometer has a first-in first-out (FIFO) data buffer that can store up to 3 KB of data, or 512
words of 7 bytes each (1 tag byte plus 6 data bytes). The tag drives the decoding and includes a parity bit for
validation, while the data can derive from any combination of accelerometer, gyroscope, up to 4 external sensors,
embedded temperature sensor, timestamp counter and pedometer step counter. The data rates can be set
independently for each source.
For the accelerometer, you can choose to store high-pass filtered data, slope detection filter data, low-pass
filtered data or unfiltered data plus an optional offset (independent for each axis). The low and high-pass filters
are configurable.
For the gyroscope, you can enable a first optional high-pass filter, and a second optional and programmable low-
pass filter.
To maximize the amount of data stored in the FIFO, you can enable an adaptive lossless delta pulse-coded-
modulation (DPCM) compression algorithm for the accelerometer and the gyroscope data. When the difference
between consecutive data words is small, instead of sending a new data word (16 bits/axis), only the difference
with respect to the previous is sent (8 bits = 2:1 compression or 5 bits/axis = 3:1 compression).
When DPCM compression is enabled, it is also possible to select when DPCM is reset by storing uncompressed
data in the FIFO: never, every 8, 16 or 32 words. This is useful when the FIFO is used as a circular buffer
(Continuous mode) and words may be overwritten, as the decoder can only work if previous data words can be
traced back to an uncompressed data word for reference.
DPCM compression works on 3 data words (sampled at time t, t-1 and t-2) and introduces a latency of 2 data
words. The compression factor depends on the difference between a data word and the previous data word.
• When DPCM is disabled, the compression buffer is flushed and the output is the non-compressed data word
at time t-2, then data word at time t-1 and data words at current time t thereafter (tag “NC_T_2”, “NC_T_1”,
and “NC”).
• When DPCM is enabled, the first output is the non-compressed data word at time t-2 (tag “NC_T_2”), used
as the reference to start decoding. Subsequent outputs depend on D2 (difference between data word at t-2
and previous decoded word), D1 (difference between data word at t-1 and t-2), and D0 (difference between
data word at t and t-1):
– If any difference in D2 exceeds 128 LSB on any axis:
◦ the uncompressed data word at time t-2 is written to the FIFO by storing the 16-bit signed value of
each axis (low and high bytes for X, Y and Z: XL XH YL YH ZL ZH, for a total of 6 bytes)
◦ tag NC_T_2
– If any difference in D2 and D1 exceeds 16 LSB, but is within 128 LSB for all axes:
◦ a 2:1 compression ratio is applied by storing the 8-bit signed difference to reconstruct 2 data
values for each axis at time t-1 and t-2 (D1x D1y D1z, then D2x D2y D2z, for a total of 6 bytes)
◦ tag 2xC
◦ the data at time t-2 is derived by summing D2 to the previously decoded data word
◦ the data at time t-1 is derived by summing D1 to the data at time t-2 just derived
– If all differences in D2, D1 and D0 are within 16LSB for all axis:
◦ a 3:1 compression ratio is applied by storing the 5-bit signed difference to derive 3 data values for
each axis at time t, t-1 and t-2 (D0x D0y D0z and a dummy bit to make 16 bits, then D1x D1y D1z
and a dummy bit, then D2x D2y D2z and a dummy bit, for a total of 6 bytes)
◦ tag 3xC
◦ the data at time t-2 is derived by summing D2 to the previous decoded data word
◦ the data at time t-1 is derived by summing D1 to the data at time t-2 just derived
◦ the data at time t is derived by summing D0 to the data at time t-1 just derived
Interrupts can be generated when the FIFO buffer stores a given number of samples (FIFO threshold level), or
when it is full, or when it overflows (overrun). The FIFO can work in the following modes:
UM2501
Inertial MEMS sensors
UM2501 - Rev 3 page 10/44

• Bypass Mode:
–the FIFO buffer is disabled and cleared
• FIFO Mode:
– the FIFO buffer collects data until it is full, then stops
• Continuous Mode:
– the FIFO buffer collects data continuously
– when it is full, the oldest samples are overwritten as in a circular buffer
– FIFO full and FIFO threshold level interrupts allow the host microcontroller to read the data before it is
overwritten
• Continuous-to-FIFO Mode:
– the FIFO buffer collects data continuously but switches to FIFO Mode as soon as the selected interrupt
occurs
– this mode is especially useful to capture data before and after a specific event
• Bypass-to-Continuous Mode:
– the FIFO buffer is disabled, but switches to Continuous Mode as soon as the selected interrupt occurs
– this mode is useful to capture data after an event has occurred
• Bypass-to-FIFO Mode:
– the FIFO buffer is disabled, but switches to FIFO mode as soon as the selected interrupt occurs
– this mode is useful to capture data after an event has occurred
– data collection stops once the FIFO is filled
UM2501
Inertial MEMS sensors
UM2501 - Rev 3 page 11/44

2.3.2 LIS2MDL 3-axis magnetometer
2.3.2.1 LIS2MDL dynamic range, resolution and accuracy
The LIS2MDL is a 3D digital magnetometer with a ±50 Gauss dynamic magnetic field range (reduced to ±25
Gauss if the magnetic field is not aligned with one of the axes), which is well above Earth’s magnetic field (which
is typically in the range of 0.25 to 0.65 Gauss).
The output is 16 bits (1.5 mGauss/LSB ±7%), and the RMS noise level in High-performance Mode is 3 mGauss
(2 LSB). The high resolution and accuracy allows e-compass orientation estimation and distance measurements
from a reference magnet (in its simplest form, the magnetometer can emulate a magnetically activated Reed-
switch).
The LIS2MDL magnetometer can perform a single-shot measurement and then return to Power-down Mode, or
operate in Continuous Mode with a programmable sampling rate (10, 20, 50 or 100 Hz).
Single-shot measurements can be performed at max. 100 Hz in high-performance mode, or max. 150 Hz in low-
power mode. In low-power mode, the power consumption is reduced to 25%, while in high-performance mode the
RMS noise is halved. An optional low-pass filter can be activated to further reduce RMS noise, without increasing
the power consumption, the bandwidth in this case is reduced from output data rate ODR/2 down to ODR/4.
2.3.2.2 LIS2MDL intrinsic offset automatic cancellation
The LIS2MDL is based on anisotropic magneto-resistive (AMR) technology, in which a set pulse is required to set
an initial operating condition; a reset pulse may be used to enable the compensation of the intrinsic magnetic
offset.
• If intrinsic offset compensation is disabled, there is no reset pulse. The set pulse can be generated at power-
on only (Set_FREQ=1), or every 64 measurements (Set_FREQ=0).
• If intrinsic offset compensation is enabled, set and reset pulses are generated alternately on consecutive
samples. After the set pulse, the measured output is the magnetic field H plus offset; after the reset pulse,
the measured output is H minus offset.
– In Continuous Mode, the device averages consecutive measurements to compensate for the offset, the
output will be equal to the magnetic field H.
–output =H+offset +H − offset
2=H
– In Single-shot Mode, the host microcontroller averages two consecutive measurements. The averaged
measurements must occur within a short time of each other for the compensation to be effective.
When intrinsic offset compensation is performed, the residual offset is in the range of ±60 mgauss, and the
dependency on temperature is ±0.3 mgauss/°C.
2.3.2.3 LIS2MDL extrinsic offset (hard iron) compensation
Hard-iron compensation is a constant offset added to magnetic field measurements to compensate distortion
caused by ferromagnetic materials or high currents. This offset can be computed by the host microcontroller by
analysing the output of the magnetometer.
When it is available, this offset can be programmed in specific registers (OFFSET_X_REG_H/L,
OFFSET_Y_REG_H/L, OFFSET_Z_REG_H/L) so that it is automatically subtracted by the LIS2MDL device.
RELATED LINKS
DT0059: Ellipsoid or sphere fitting for sensor calibration
MotionMC software library in the X-CUBE-MEMS1 Sensor and motion algorithm software expansion for STM32Cube
2.3.2.4 LIS2MDL interrupt generation
The LIS2MDL magnetometer sensor can be configured to generate an interrupt when output data (before or after
the subtraction of hard-iron offset) exceeds a programmed threshold (INT_THS_H/L_REG) in the positive or
negative direction.
Comparisons can be enabled independently for each axis (XIEN, YIEN, ZIEN flags):
• when the output data is above the positive threshold, the corresponding flag is set (P_TH_S_X/Y/Z)
• when the output data is below the negative threshold, the corresponding flag is set (N_TH_S_X/Y/Z).
A specific interrupt can be generated if there is a measurement range overflow at the internal ADC.
UM2501
Inertial MEMS sensors
UM2501 - Rev 3 page 12/44

2.4 Environmental MEMS sensors
2.4.1 LPS22HH barometer
2.4.1.1 LPS22HH acquisition chain
The LPS22HH pressure sensor can perform a one-shot measurement and then return to Power-down Mode, or it
can operate in Continuous Mode with a programmable sampling rate (1, 10, 25, 50, 75, 100 or 200 Hz).
Measurements can be taken in normal Low-noise Mode, or in Low-power Mode to minimize current consumption.
When continuous mode is selected, Low-noise Mode is not available at 100 and 200 Hz. In all cases, an optional
low-pass filter can be enabled with a programmable cutoff frequency to reduce the noise level (the bandwidth is
reduced from output data rate ODR/2 down to ODR/9 or ODR/20).
The low-pass filter is reset when it is enabled. After reset, a specific settling time is required before the first
accurate sample on the output is available (9 or 20 samples must be discarded, respectively).
Two programmable offsets can be subtracted from measured data:
• First, offset compensation (always): the offset measured with one-point calibration (OPC) can be stored in
specific registers (RPDS) and then subtracted from subsequent measurements (OPC-compensated data =
data – RPDS*256). The low-pass filter, if enabled, will filter OPC-compensated data. Filtered and unfiltered
OPC-compensated data is stored in the FIFO.
• Second, auto-zero mode (optional): the offset-compensated (and possibly filtered) pressure measurements
are stored in specific registers (REF_P) when auto-zero is enabled (AUTOZERO or AUTOREFP set to 1)
and then subtracted from subsequent measurements (AZ-compensated data = data – REF_P). Filtered and
unfiltered OPC-compensated data or AZ-compensated data is stored in output registers.
2.4.1.2 LPS22HH interrupt generation and FIFO buffer
The LPS22HH pressure sensor can be configured to generate interrupt events related to pressure acquisitions
and FIFO status (watermark reached, full, overrun, etc.).
The interrupt can be generated when a new pressure or temperature sample is available, or when the AZ-
compensated data exceeds a programmed threshold (THS_P) in the positive (PHE flag enables the comparison
with +THS_P) or negative (PLE flag enables the comparison with –THS_P) direction. The following configurations
are available:
1. AUTOZERO=0 and AUTOREFP=0: interrupt logic is disabled; filtered and unfiltered OPC-compensated data
goes to FIFO and output registers.
2. AUTOZERO=1: interrupt logic is enabled; REF_P is set when interrupt is enabled; AZ-compensated data
goes to the interrupt logic and to the output registers, while filtered and unfiltered OPC-compensated data
goes to FIFO.
3. AUTOREFP=1: interrupt logic is enabled; REF_P is set when interrupt is enabled; AZ-compensated data
goes to interrupt logic, while filtered and unfiltered OPC-compensated data goes to FIFO and output
registers.
The FIFO buffer can store up to 128 pressure samples (24 bits each) and 128 temperature samples (16 bits
each). FIFO depth can be limited by stopping at the programmable watermark level. The FIFO can work in the
following modes:
• Bypass Mode:
– the FIFO buffer is disabled and cleared
• FIFO Mode:
– the FIFO buffer collects data until it is full or the watermark setting is reached, then stops
• Continuous (or Dynamic Stream) Mode:
– the FIFO buffer collects data continuously
– when it is full or the watermark setting is reached, the oldest samples are overwritten as in a circular
buffer
– FIFO full and FIFO threshold level interrupts allow the host microcontroller to read the data before it is
overwritten
UM2501
Environmental MEMS sensors
UM2501 - Rev 3 page 13/44

• Continuous (Dynamic Stream)-to-FIFO Mode:
–the FIFO buffer collects data continuously but switches to FIFO Mode as soon as the selected interrupt
occurs
– this mode is especially useful to capture data before and after a specific event
• Bypass-to-Continuous (Dynamic Stream) Mode:
– the FIFO buffer is disabled, but switches to Stream Mode as soon as the selected interrupt occurs
– this mode is useful to capture data after an event has occurred
• Bypass-to-FIFO Mode:
– the FIFO buffer is disabled, but switches to FIFO mode as soon as the selected interrupt occurs
– this mode is useful to capture data after an event has occurred
UM2501
Environmental MEMS sensors
UM2501 - Rev 3 page 14/44

2.4.2 HTS221 temperature sensor
2.4.2.1 LPS22HH vs HTS221 ambient temperature measurement
The temperature sensor in the LPS22HH device is designed to compensate for temperature effects in ambient
pressure measurements, while the temperature sensor in the HTS221 device is designed and characterized for
ambient temperature measurements.
The BlueTile takes ambient temperature data from the HTS221 because it is more accurate over a larger range
than the LPS22HH. However, applications requiring a high data rate (above 12.5 Hz) can take data from the
LPS22HH temperature sensor.
Table 1. Comparison of HTS221 and LPS22HH temperature sensor characteristics
Device Temperature sensor
operating range Sensitivity Temperature absolute accuracy Data rate Response time Long term drift
HTS221 -40 to +120 °C 64 LSB/°C ±1°C (0 to 60 °C)
±0.5 °C (15 to 40 °C); 1, 7, 12.5 Hz 15 s (time to
63%).
0.05 °C/
year.
LPS22HH -40 to +85 °C 100 LSB/°C ±1.5 °C (0 to 80 °C) 1, 10, 25, 50, 75,
100, 200 Hz. - -
2.4.2.2 HTS221 acquisition chain
While all the board components have an embedded temperature sensor to monitor the temperature of the silicon,
the embedded HTS221 has the physical properties and accuracy necessary for the measurement of ambient
temperature.
The raw output of the humidity acquisition chain is stored in two 8-bit registers (HUMIDITY_OUT_H and
HUMIDITY_OUT _L), which are concatenated into a 16-bit two’s complement value (H_OUT). The raw output is
already temperature compensated and calibration coefficients to derive Relative Humidity (RH) from raw humidity
data are stored in the device. Factory calibration is performed at two different humidity levels and one
temperature.
The following sequence is used to calculate RH:
1. First true RH during calibration: H0_rH_x2setRH0=H0_rH_x2
2
2. Second true RH during calibration: H1_rH_x2setRH1=H0_rH_x2
2
3. First raw H output during calibration: H0_T0_OUT setH0=H0_T0_OUT
4. Second raw H output during calib.: H1_T0_OUT setH1=H1_T0_OUT
5. Current raw H output: H_OUT setH =H_OUT
6. Current RH% by linear interpolation: RH =RH0 + RH1−RH0*H−H0
H1−H0
The raw output of the temperature acquisition chain is stored in two 8-bit registers (TEMP_OUT_H and
TEMP_OUT _L) which are concatenated into a 16-bit two’s complement value (T_OUT). Calibration coefficients to
derive the Temperature in °C from raw temperature data are stored in the device. Factory calibration is performed
at two different temperatures.
Follow the sequence below to calculate the temperature in °C:
1. MSB of true temperatures during calibration: T1T0MSB setMSB1=T1T0MSB&0x03, MSB0 = T1T0MSB
4
2. First true T in Celsius deg. (two’s comp): T0_°C_x8setT0°C=T0_°C_x8 + MSB0*256
8
3. Second true T in Celsius deg. (two’s comp): T1_°C_x8setT1°C=T1_°C_x8 + MSB1*256
8
4. First raw T output during calibration: T0_OUT setT0=T0_OUT
5. Second raw T output during calibration: T1_OUT setT1=T1_OUT
6. Current raw T output during calibration: T_OUT setT =T_OUT
7. Current Temperature in °C: T=T0deg +T1deg − T0deg *T − T0
T1−T0
UM2501
Environmental MEMS sensors
UM2501 - Rev 3 page 15/44

2.4.2.3 HTS221 system integration
To get reliable and consistent measurements, the system design should maximize sensor exposure to the
external environment while minimizing error sources.
•Mechanical design:
– if there is one vent hole in the BlueTile housing, the hole diameter should be maximized and the dead
volume enclosed should be minimized
– two or multiple vent holes are preferable in order to create a laminar airflow and minimize the response
time
– avoid materials that absorb humidity
• Mechanical stress:
– Avoid flexing or bending the BlueTile board
• Heat:
– Heat convection or temperature gradients on the board may affect the sensor
– Metal lines and planes, such as the ground plane, should be kept far from the sensor
– Milled slits further increase decoupling
– Insulation may be required to isolate the BlueTile from convective and conducted heat
• Light exposure may induce a change in temperature and humidity.
To accelerate sensor recovery time in case of condensation, the firmware running on the host microcontroller can
be coded to switch the heating element on for a certain duration. Humidity and temperature data should not be
read during the heating cycle.
The BlueTile is designed for ultra-low power operation and a long battery life (1 month to 1 year on a typical
CR2032 coin battery, depending on the application). Power and heat generated on-board is therefore very limited.
Possible sources of conducted heat such as the BlueNRG-2 and the LED are placed as far as possible from
MEMS sensors to avoid any impact on measurements. Slits are also present to further isolate the HTS221
temperature sensor from neighboring components.
RELATED LINKS
AN4722: HTS221 digital humidity sensor: hardware guidelines for system integration
UM2501
Environmental MEMS sensors
UM2501 - Rev 3 page 16/44

2.5 MP34DT05-A digital MEMS microphone
2.5.1 Features of the MEMS microphone
The MP34DT05-A omnidirectional top-port digital microphone has the following features:
•122.5 dBSPL acoustic overload point (AOP), or 0 dBFS (100% of digital Full Scale(FS))
• -26dBFS ±3dB (5% of FS) sensitivity at 1 kHz and 94 dBSPL
• 64dB A-weighted SNR
• The equivalent input noise floor is 30 (94-64) dBSPL, or -90 (-26-64) dBFS (0.003% of FS).
• The total harmonic distortion plus noise THD+N:
– 0.2% at 94 dBSPL
– 0.7% at 110 dBSPL
– 6% at 120 dBSPL
• The frequency response is generally flat from 100 Hz to 5 kHz, up to +0.5 dB at 20 kHz.
Appropriate frequency response is important to maximize the performance of beamforming applications that can
determine the direction of incoming sounds and reject noise coming from certain directions
RELATED LINKS
AcousticBF software library in the X-CUBE-MEMSMIC1 Digital MEMS microphones acquisition and processing software
expansion for STM32Cube
2.5.2 PDM to PCM conversion
The output of the digital MEMS microphone is a bit stream (1.6 MHz on the BlueTile), where frequency of ones is
proportional to the sound pressure level: this is known as pulse-density modulation (PDM). However, audio is
usually represented by words with at least 16 bits, at a rate equal or greater than twice the max. audio frequency,
typically 16, 24 or 48 kHz: this is known as pulse-coded modulation (PCM).
The PDM-to-PCM conversion can be performed with low-pass filtering. In its simplest form, it is equivalent to
averaging, or counting the number of ones over the sampling time of the PCM word.
In the BlueTile the conversion is performed by hardware in a dedicated block in the ADC peripheral in BlueNRG-2
device, but the conversion can also be performed by software.
RELATED LINKS
AN4957: How to synchronize the DFSDMs filters and how to program the pulse skipper on the STM32F413/423 line devices
AN3998: PDM audio software decoding on STM32 microcontrollers
PDM2PCM software library in the X-CUBE-MEMSMIC1 Digital MEMS microphones acquisition and processing software
expansion for STM32Cube
UM2501
MP34DT05-A digital MEMS microphone
UM2501 - Rev 3 page 17/44

2.6 VL53L1X Time-of-Flight sensor
2.6.1 Features of the Time-of-Flight sensor
The VL53L1X is a state-of-the-art laser-ranging sensor in a miniaturized package, with a 940 nm Class 1 vertical-
cavity surface-emitting laser (VCSEL), a receiving 16x16 array of single-photon-avalanche-detectors (SPAD),
physical infrared filters and optics to achieve the best ranging performance.
The VL53L1X measures the time it takes for photons emitted by the laser to be reflected back to the sensor.
The field-of-view (FoV) is programmable from 15 to 27 degrees, which sets the region of interest (ROI) in the
SPAD array from 4x4 to 16x16 detectors, respectively. The position of the ROI in the array can also be set, in
order to orient the FoV cone in a specific direction.
The sensor returns the following data:
• ranging distance in mm
• the return signal rate and the ambient signal rate (kilo-count-per-second (kcps) per SPAD)
• range status
2.6.2 Calibration for best performance
For best performance, run the calibration functions at least once after assembly is completed. Calibration data is
stored in the host microcontroller and loaded in the VL53L1X on each startup.
If you call one or more calibration functions, you must observe the sequence below:
•RefSPAD calibration:
– Optimizes device dynamics and accuracy.
– Especially recommended if a protective glass cover is used on top of the device.
– SPAD cells are classified as non-attenuated, attenuated by 5 or attenuated by 10.
– SPAD are then selected to avoid internal signal saturation.
– This part-to-part value is computed during the final test at ST and stored in the non-volatile memory
(NVM); it is automatically loaded after boot.
• Offset calibration:
– Should always be performed after assembly.
– Compensates for part-to-part variations, reflow effects, and cover glass effects.
• Crosstalk calibration:
– Performed whenever there is a protective cover glass on top of the device, which may reflect laser light
back to the sensor.
– Crosstalk can be compensated internally by the VL53L1X .
2.6.3 Ranging Mode configuration, distance and accuracy
The sensor performs ranging continuously and autonomously with a programmable inter-measurement period. If
the interrupt pin is connected, the host receives an interrupt whenever a new measurement is available. The inter-
measurement period must be longer than the selected timing budget plus 4 ms.
The measurement timing budget can be set from 20 to 1000 ms. Increasing the timing budget increases the
maximum distance and reduces the repeatability error (standard deviation of measurements), but power
consumption also increases:
•20 ms is the minimum timing budget, can be used only in Short Distance Mode
• 33 ms is the timing budget that can be used for all distance modes
• 140 ms is the minimum timing budget which allows maximum distance of 4 m in Long Distance Mode (dark
ambient light and a 54% gray target).
The minimum accurate ranging distance is 40 mm, while the maximum ranging distance depends on several
factors such as the ambient light (the lower the better), the reflectance of the target (the higher the better), on the
timing budget (the higher the better), on the selected ROI (the larger the better) and on the selected ranging
mode:
• Short Distance Mode:
– ranging distance up to 1.3 m
– independent of ambient light and target reflectance
UM2501
VL53L1X Time-of-Flight sensor
UM2501 - Rev 3 page 18/44

• Medium Distance Mode:
–ranging distance up to 2.9 m in the dark
– limited to 0.7 m in strong ambient light (200 kcps/SPAD, direct illumination on the sensor in a sunny
day, infrared component removed by glass)
• Long Distance Mode:
• ranging distance from 1.7 to 3.6 m max. in the dark for targets with 17 to 88% reflectance
• limited to 0.7 m in strong ambient light
• longest ranging distance is 4 m when timing budget is at least 140 ms
The ranging error is the sum of the accuracy and the repeatability error, and is typically between ±20mm in the
dark, and ±25mm in strong ambient light:
The software driver provided by ST uses two parameters to qualify the ranging measurement:
• VL531L1X_CHECKENABLE_SIGMA_FINAL_RANGE:
– standard deviation in mm, default is 15mm
– decrease to reduce repeatability error, but maximum ranging distance is also reduced
• VL53L1X_CHECKENABLE_SIGNAL_RATE_FINAL_RANGE:
– rate of photons reflected by target in Mcps, default is 1 Mcps
– increase to reduce repeatability error, but maximum ranging distance is also reduced
2.6.4 Power up and boot sequence
In the BlueTile, the XSHUT pin connected to the host BlueNRG-2 (IO7 pin) is used to power off the sensor
between measurements. The maximum boot time is 1.2 ms.
Note: When the sensor is powered off, the configuration is lost. The sensor must be re-configured on each wake-up.
To minimize power consumption, the application firmware may skip the configuration step after wake-up and use
the default (long ranging mode, 30 ms timing budget). The application may also perform an early readout, not
waiting for the data ready flag, and immediately shut down the sensors. This reduces power consumption but also
limits accuracy. The minimum ranging distance in this case is around 70 mm.
UM2501
VL53L1X Time-of-Flight sensor
UM2501 - Rev 3 page 19/44

3STEVAL-BCN002V1D host board for programming and debugging
The STEVAL-BCN002V1D is used to Flash and debug the STEVAL-BCN002V1 BlueTile sensor node.
Figure 5. STEVAL-BCN002V1D top and bottom
During Flashing and debugging, remove the battery as the BlueNRG-2 device on the BlueTile is powered by the
host board, which receives its supply voltage through its USB connector or on the USB connector of the ST-LINK
V3 Stamp if you are using it.
The BOOT and RESET pin of the BlueNRG-2 on the BlueTile are controlled by the STM32L1 on the host board,
or by the ST-LINK V3 Stamp if used, or by the user (user button and connector CN5).
UM2501
STEVAL-BCN002V1D host board for programming and debugging
UM2501 - Rev 3 page 20/44