IESH v0.1 — Field-Testing the Classroom Laboratory
The sensor stack
DHT22 and BMP280 carry over from v0. New: a DS18B20 soil-temperature probe on 1-Wire, a capacitive soil-moisture sensor, and MQ-7 (CO) and MQ-135 gas sensors, all read through an MCP3208 8-channel ADC. The MQ sensors output 0–5 V but the ADC reference is 3.3 V, so a two-resistor divider halves the signal and the software multiplies it back. Small detail; ruins your data if you forget it.
Calibration
Soil moisture is calibrated against recorded endpoints — raw ADC 4021 in dry air, 1673 fully wet — and reported as a percentage between them. The MQ sensors get a clean-air baseline (R0) recorded per unit with a calibration script and stored with the data, because every MQ unit is its own animal.
The sky camera
A fisheye camera captures the sky every 15 minutes and runs a simple cloud-cover estimate. One camera is shared between scheduled captures and the live stream, so captures pause while someone is watching — groundwork for the all-sky meteor camera this will eventually become.
The classroom part
The station broadcasts its own WiFi hotspot. Any phone or laptop that joins gets the live dashboard and the learning portal: five modules — climate, soil, air quality, data science, environment — with 25+ activities aligned to the national curriculum, their visualisations driven by the station's own sensors. There is no internet anywhere in that loop, which is exactly the point for the schools this is built for.
Provisioning
One command flashes a complete station onto a blank SD card; first boot installs everything unattended in about ten minutes. The data store lives outside the code directory, so re-flashing software never touches collected data.
What's next
Cloud sync to the open-data API — the local logs are ready and waiting, the uploader integration is the work — and per-device credentials before any station goes to a school.
Kathmandu, Nepal · 1350m asl