We are a final-year Biomedical Engineering team developing a breath analysis system using the BME688 for VOC detection and potential early respiratory disease screening.

Our system is fully functional and includes a custom breath chamber, airflow tubing, an automatic purge fan, lithium battery power, and an ESP32-S3 N16R8 microcontroller connected to the BME688 development board.

We initially attempted to use BME AI-Studio, but it did not recognize our ESP32-S3 platform. Therefore, we migrated to Visual Studio Code with PlatformIO.

At first, we encountered several library conflicts, dependency issues, and version compatibility problems. After resolving them, we successfully verified via the serial monitor that all eight sensors on the development board are operating correctly.

However, we are still facing two major challenges:

We have not yet achieved the desired heating profile configuration for breath analysis. We are currently developing a custom heater profile and optimized firmware, but the sensor response is not yet as stable and informative as required.

We currently lack sufficient clinical datasets for training our machine learning model. Access to real clinical breath data has been difficult.

Our questions are:

What are the recommended best practices for designing and optimizing custom heating profiles for BME688 in breath analysis applications?

Can synthetic or simulated sensor data be useful for the initial training phase of an AI model before collecting real clinical data?

Are there any recommended workflows, examples, or resources for projects using BME688 with custom ESP32-based platforms outside BME AI-Studio?

Any technical guidance or recommendations would be greatly appreciated.

Thank you very much for your support.

Everything works fine as long as the pendulum doesn't move too fast. As it was built to move fast, the BNO085 doesn't match the use case. For fast movements, the sensor "blacks out" and doesn't deliver new values until the motion slows down.

Is there a sensor in your portfolio that does on-board sensor fusion for a gravity report, but comes with higher ranges for the accelerometer and gyroscope (ideally at low noise levels)? A magentometer is not required. Some sort of breakout board would be great as well as the availablility of a micropython driver.

Unfortunately, I have no experience in sensor electronics, so I cannot answer the question myself by looking at the datasheets...

If there is a better way to measure rotation angles of a chaos pendulum, please let us know.

Thanks, Timo

I am required to know the thermal resistance junction to ambient (RthJA) of the BMA456 in order to complete certification for our product. Can this information please be provided?

Best wishes,

SA

We are using the BME690 sensor in one of our projects and have conducted chamber temperature testing in the range of −20 °C to +60 °C.

Based on the test results, the positive temperature readings are in good agreement with the chamber set temperature. However, we are observing a deviation in the measured values at negative temperatures when compared against the chamber temperature.

The chamber test data has been attached for your reference and review. We would like to confirm whether the observed behavior at negative temperatures is expected for the BME690 ambient temperature measurement, or if any specific compensation, configuration, or calibration needs to be applied.

Kindly review the attached data and provide your confirmation or recommendations.

I want to ask you guys about this, I am trying to connect BME680 to Raspberry Pi 5, using I2C to communicate with the sensor... I need to know how to program it and show the AQI result, can you guys help me?

Thank you.

The sensor is used intermittently (typically <1 hour per session), and the system performs a fresh calibration at each power-up. We are primarily measuring relative motion/orientation changes over short durations, not long-term inertial navigation.

The IMU is mounted on a small PCB (~25 mm × 20 mm), with board thickness options of 0.8 mm or 1.6 mm. The current concept is to clamp the PCB around its edges within a plastic housing.

My questions:

1. Mounting Sensitivity:
   Given the short operational duration and recalibration at startup, how critical is strict adherence to recommended PCB mounting guidelines (e.g., stiffness, isolation from stress)?
   Specifically, is edge clamping likely to introduce meaningful drift or bias instability in this type of use case?

2. PCB Thickness / Stiffness:
   For a board of this size, is there a meaningful benefit in going from 0.8 mm to 1.6 mm thickness in terms of reducing mechanical coupling and signal error?

3. Encapsulation / Potting:
   We are considering fully encapsulating the PCB in a medical-grade silicone for environmental protection.

• Are there known concerns with IMU performance (e.g., bias shift, thermal effects, damping) when using soft silicone encapsulants?

• Any guidance on air-cure vs. heat-cure silicones for this type of application?

4. Real-World Experience:
   Are there examples of similar use cases where IMUs are mounted on small PCBs with edge constraint and/or encapsulated, and what lessons were learned?

Any insights or practical experience would be appreciated.

To ensure 150% reliability for this student project, I would appreciate your feedback on these layout points:

1. BMP388 Mechanical Stress: Since the FPC will be curved around the nose, I’ve placed a Polyimide stiffener under the BMP388. Is this needed and does prevent MEMS offset shifts caused by mechanical strain, or do you recommend a rigid FR4 stiffener for better stability?

• also do you recommend some other models for this very small kind of pcb?

2. Venting & Silicone: The BMP388 needs to sense breathing air. I have planned a 0.4mm vent hole through the silicone case. Do you have any guidelines do i need hole in the pcb also? Would a small air cavity around the sensor be recommended?

3. Thermal Isolation: The sensors are located approximately 10mm from an nRF52840 SoC. On such a thin FPC, should I take extra steps (like specific copper cutouts) to prevent thermal drift on the BMP388 pressure readings?

4. BMI270 Placement: For activity tracking on the nose, are there any specific orientation or mounting recommendations to ensure the highest accuracy for gesture/movement detection on a flexible substrate?

Development kit is expensive(shipping to Vietnam) it cost 3.000.000- 5.000.000 VNĐ (about 100-200 USD).

The bme688 boar is "heart" for my project to detect vaping smells in bathroom school

I am a high school student in Vietnam. I am building a project to detect vaping smells for science and technology competition.