BME688 | GUILCOR SENSORS

BME688
Temperature sensors

Advanced environmental sensor with gas detection and AI capabilities for air quality and smart sensing applications.

Interface

I²C

Measurement range
-40 / 125°C

Humidity range
0% → 100%

Pressure gauge (hPa)
300 → 1100

Temperature accuracy

+/-0,5K

Humidity accuracy
+/- 3 %

Pressure precision
+/- 1 hPa

Internal resolution
16 to 20 bits

power supply voltage
1,71 → 3,6 V

Sampling rate
1 to 100 Hz

Typical current
3,6µA(active)

Price
Medium

What is a BME688 sensor ?

The BME688 is Bosch's most advanced environmental sensor.

It combines:

the three classic sensors of the BME280 (temperature, humidity, pressure),
a gas sensor (MOX) capable of identifying volatile organic compounds (VOCs),
and embedded intelligence via the BSEC2 library (Bosch Software Environmental Cluster).

The great interest of the BME688 lies in its ability to be AI-trained to recognize specific odor profiles (smoke, fermentation, stale air, burnt plastic, etc.).

Operating principle

The BME688 operates in four steps:

MEMS sensors measure temperature, humidity, and pressure.
A MOX (Metal Oxide) gas sensor changes resistance according to ambient gases.
A programmable internal heater adjusts the temperature of the gas sensor (200 to 400 °C).
The BSEC2 algorithm transforms raw data into:

- IAQ (Indoor Air Quality)
- CO2 equivalent
- VOC equivalent
- Odor classification (AI mode)

I AQ = f(T, RH, P, R_gaz)

💡 The heating profiles and odor recognition are configurable via the Bosch AI Studio software.

Example of data output

Parameter	Unit	Typical example
Temperature	°C	23,6
Humidity	% RH	45,8
Pression	hPa	1012,7
Gas resistance	kΩ	68,3
IAQ (Indoor Air Quality)	0–500	54 (healthy air)
CO₂ equivalent	ppm	472
VOC equivalent	ppm	0,22

Principle diagram (I²C connection)

+3.3V
 │
[BME688]
 │ SDA ─────────┐
 │ SCL ─────────┤── [I²C Microcontroller]
 │ GND          │
 │ CSB / SDO ───┘ (config SPI / address)

💡 I²C Address: 0x76 (SDO=GND) or 0x77 (SDO=VCC).

Application areas

🌬️ Indoor Air Quality (IAQ) Monitoring

🧠 Odor Analysis / Specific Gas Detection (AI)

🏠 Home Automation, HVAC, Air Purification

🚗 Automotive Environmental Sensors

📱 Connected Devices, Smartwatches, Environmental IoT

Should I choose a BME688 sensor ?

Strengths points

🧠 Smart environmental sensor (integrated AI)
→ Capable of identifying and classifying air profiles using Bosch AI Studio.
🌍 4 sensors in 1→ Temperature, humidity, pressure, and gas — one single component, zero manual calibration.
🎯 Bosch Precision and Stability
→ Each unit is factory calibrated and thermally compensated.
🔋 Adaptive power modes
→ Allows for precise management of consumption based on the measurement frequency.

Weaknesses points

🤖 Requires AI training
→ For specific detection (odors), the model needs to be trained using Bosch AI Studio.
🔥 Gas sensor heating
→ Can consume up to 12 mA depending on the heating profile.
📦 Higher cost
→ But justified by its advanced features and embedded AI.

Useful information

Here is some useful information regarding the BME688 sensors.

Main Registers

Address (hex)	Name	Description
0xE0	reset	Sensor reset
0x74	ctrl_meas	Temperature and pressure configuration
0x72	ctrl_hum	Humidity configuration
0x71	config	Duration of monitoring and filtering
0x50 → 0x59	data registers	Raw data T, RH, P, gas
0x89 → 0xA1	coefficients	Factory calibration data

I²C Reading Example (simplified pseudo-code)

i2c_start();
i2c_write(0x76 << 1);
i2c_write(0x50);
i2c_start();
i2c_write((0x76 << 1) | 1);
read 10 bytes -> raw_t[3], raw_p[3], raw_h[2], raw_g[2];
i2c_stop();

temp = compensate_T(raw_t);
press = compensate_P(raw_p);
hum = compensate_H(raw_h);
gas = compensate_G(raw_g);
IAQ = f(temp, hum, gas);

💡 The compensation and IAQ functions are provided by the BSEC2 library.

Typical integration diagram

3.3V ───────────┐ │ [BME688] │ SDA ───────┐ │ SCL ───────┤── MCU (ESP32, STM32, Raspberry Pi…) │ CSB/SDO ───┘ (I²C or SPI) │ GND

💡 Pull-up resistors of 4.7 kΩ on SDA and SCL required if not integrated.

Use of AI mode (Bosch AI Studio)

Step	Action
1️⃣	Collect air data via BME688 + BME AI-Studio App
2️⃣	Label the scent profiles (coffee, clean air, plastic…)
3️⃣	Train the AI model on PC
4️⃣	Export the .bmeconfig file and load it into the microcontroller
5️⃣	The BME688 identifies the type of air encountered in real-time ✅