See the website for a demo video: https://oasis-terrarium.com
This project is an enclosure for growing plants. Features include:
- high-power LED lighting
- a built-in water tank and mister
- fans for airflow
- temperature and humidity sensing
- wi-fi connectivity and a web-based interface for control and monitoring (see https://oasis-terrarium.com/docs/usage_guide/#web-interface)
The entire project is open-source: - electronics: designed in KiCad
- 3D models: designed in CadQuery
- software: written in Rust
I initially started this project during COVID and built a working prototype using a Raspberry Pi and off-the-shelf parts. The first prototype worked, but was pretty rough. Several months ago, I picked up the project again and began a complete re-design. After two rounds of circuit board design and countless iterations of the 3D-printed parts, I'm finally happy with the results. Along the way I learned a lot about electronics design and fabrication, 3D modeling, and Rust.Happy to answer any questions - thanks for checking it out!
Here are some direct links to parts of the project you may find interesting:
- demo video: https://oasis-terrarium.com
- detailed build guide with pictures: https://oasis-terrarium.com/docs/build_guide/
- interactive 3D model: https://oasis-terrarium.com/docs/3dmodel/
- detailed info on electronics design: https://oasis-terrarium.com/docs/electronics/
- 3D-printable models: https://www.printables.com/model/1315117-oasis-smart-terrarium
Thanks!
> To protect the ultrasonic mister from bad firmware frying it, you could probably replace R6 with a sufficiently sized capacitor to make an effective DC block
That is an excellent idea - I'll definitely keep that in mind for a future version.
> I've also had issues with a thermal sensor just reading board temp, best solution is to make the sensor a separate board. Short of that, one can mill slots around it and reduce copper pour to reduce thermal conductivety from the PCB.
I ended up just moving the sensor offboard and as far away from the heat source as I reasonably could. I did consider the milled slots + reduced pour option, but given the amount of heat generated by the LEDs, it seemed best to move the sensor further away.
Most of my time designing hardware is spent trying to find ways to prevent the doofus writing the firmware from releasing the magic smoke. I am that doofus