Open Hardware · Browser Flashing · WiFi Mesh · Per-Tree AI
Build an open-source soil monitoring node from off-the-shelf parts, bring it online from your browser, and let the dashboard learn how each individual tree responds. Watch the full walkthrough, then follow the steps below.
The Walkthrough
Hardware, firmware, sensor addressing, field deployment, connectivity, and the intelligence layer — end to end.
Step by Step
Built around the right to repair: every part is documented, replaceable, and inspectable. You — not a manufacturer — remain in control of your hardware.
The complete board is open source — schematics, PCB layout, Gerber files, bill of materials, assembly docs, and firmware. Manufacture it yourself or have it fabricated by any PCB house. We ordered ours from JLCPCB and publish the exact files and ordering information. If one manufacturer disappears tomorrow, the same Gerbers go straight to another. Start on the Open Hardware page for the v4 Gerbers and BOM, then order from JLCPCB.
Start with a sturdy enclosure suited to the environment — it doesn't need to be an industrial tank, but it should keep out rain, dust, and insects. You'll need five cable entries: the WiFi antenna, 12V input from the solar charger, 12V output to the sensors, and the two RS-485 lines (A and B). Power the board from a single 12V battery with a proper BMS, split across both input connectors — this distributes current across two pins rather than overloading one. The external antenna gives the node real range in the field.
In your Sankhya Farms account, enter the WiFi credentials the node will use to join your orchard's mesh. The browser talks to the node directly over USB-C using the Web Serial API — there is no separate flashing software to install and no command-line tools. Everything happens in the browser. Once configured, the node connects automatically and begins uploading sensor data.
Connect each RS-485 sensor individually. As it's detected, the node assigns a unique Modbus slave ID so every device has its own address on the bus and conflicts are impossible. Repeat until all sensors are addressed. From then on, the node knows exactly which sensor is connected, how to read it, and where to upload. Browse the compatible RS-485 sensor library for moisture, EC, pH, temperature, salinity, and N-P-K probes.
Install the enclosure and insert the probes into the soil. The only requirement is that the node stays within coverage of your orchard's WiFi mesh. Once powered, it reconnects automatically and begins reporting. The dashboard shows battery state-of-charge so you can monitor available power remotely — connect a solar panel sized for your climate to keep the battery topped up indefinitely.
Outside WiFi range, plug a standard 4G/LTE USB modem into the dedicated USB-A port: insert a SIM, configure the connection, and the node uploads over cellular — typically around 1 MB per day, with the modem powered only while the node is awake. Or extend coverage with inexpensive WiFi repeaters (roughly 100–150 m each) powered from the same 12V sensor rail, so they wake and sleep with the node. Either way, all nodes share one synchronized schedule — about 50 minutes asleep, about 10 minutes awake — so the mesh is live only while data is moving, then powers back down.
Firmware updates happen over the air. Because the platform is built around standard RS-485 and open firmware, support for new sensors keeps expanding — and the community can contribute too. If a sensor speaks RS-485 and its protocol is documented, there's a strong chance it can be integrated. Your investment is in the platform itself, not any single product.
After irrigation, root-zone moisture rises sharply, then declines as the tree draws water through the day. The AI analyses this daily draw-down alongside recent irrigation history to judge whether what the tree used is being adequately replenished — and issues an irrigation recommendation with a confidence score. As you move nodes around the orchard, the dashboard builds a history for every tree and learns how each one responds, producing the uptake index. Read more on the moisture-curve logic. Measure, act, observe, learn — the system gets more useful every season.
Go Deeper
Questions