BME688 for gas sensing

Hi team,

I have a quick suggestion request. I am planning to use BME688 to sense various gases e.g., ammonia with help of a custom trained AI/ML model. Sure, it will require a bit of ML model training efforts but looks a lot better option as compared to MQ137 in terms of quality and power use. Any opinion/feedback?

1 Like

Hey Riddhi,
Welcome back!

It seems like you’ve got an awesome project in the works! I think that using the BME688 is advantageous when using it for an AI/ML use case, it will be able to detect multiple gases at the same time and hopefully you can train the model to effectively interpret that data.

I wish you the best of luck with the project and make sure you keep us up to date on how it goes!


Hi Riddhi,

I had a quick look at the presentation Adafruit gave on using the machine learning with that sensor (linked here for anyone else interested later) and it sounds like a huge step up from your basic analog sensor that has to be calibrated for a single setup.
Being able to train multiple sensors with the demo code to detect a target scent and extracting the preferred sensor settings first try rather than trial and error through a bunch of options is a win.

1 Like

Yeah. I think you can proceed with BME688. It is better than MQ sensors at least for ML based projects.

1 Like