Guide · Off-grid power
The solar-powered sensor box that runs itself: size the power first.
A solar-powered sensor box in a paddock has no mains and no one to babysit it. Get the power budget wrong and it runs fine on the bench, then dies on the third overcast day while you are 40 km away. Building one that truly runs itself starts with the power system, long before you pick a single sensor.
Last reviewed: 21 July 2026 · by Rural IoT
Size the power before you buy a single sensor
Everything else is negotiable; the power system decides whether the box survives winter. Work it out in this order.
1. Add up the load
Measure or estimate the average current: the MCU and radio, plus each sensor, plus a margin. A modest node whose radio transmits periodically lands around 1.24 A average at 12 V in this example (about 15 W). Duty-cycling, sleeping between readings, is the biggest lever you have; a box that is awake 100% of the time needs a bigger everything.
2. Turn load into daily energy
1.24 A over 24 h is about 30 Ah per day, around 360 Wh. That is the hole you refill every day.
3. Size the battery for ride-through
Not just a day. A 40 Ah LiFePO4 is the floor, barely a day at this load, so add reserve for the no-sun stretches you actually get. Each extra day of autonomy is roughly another 30 Ah; two to three days of reserve is sane for a remote box.
4. Size the solar for a bad week
Not a sunny afternoon. A panel's watt rating is peak-sun; real yield is a few peak-sun hours a day, and far less under cloud. Around 150 W covers the daily energy plus recharge in good sun, so fit roughly double (about 300 W) to keep up through overcast, short winter days.
5. Use an MPPT charge controller
Matched to the panel voltage and to LiFePO4. MPPT harvests noticeably more than cheap PWM when the light is poor, which is exactly when you need it.
| Item | Figure (worked example) | Why |
|---|---|---|
| Average load | About 1.24 A at 12 V (around 15 W) | MCU, radio and sensors, duty-cycled |
| Daily energy | About 30 Ah, around 360 Wh | 1.24 A over 24 h |
| Battery | 40 Ah LiFePO4 floor, size up for 2 to 3 days reserve | Ride through cloudy stretches, never run flat |
| Solar (clear-sky need) | About 150 W | Daily energy plus daytime recharge |
| Solar (actually fit) | About 300 W, roughly double | Real yield on overcast or short days |
| Controller | MPPT, matched to panel and LiFePO4 | More harvest in poor light than PWM |
A sensor kit that earns its power
Pick sensors that survive outdoors and do not waste current:
Microwave-radar motion, RCWL-0516. Sees movement through a plastic enclosure and does not go blind in heat the way a PIR does on a hot day. Good for presence and intrusion.
Ultrasonic distance, HC-SR04. Tank level, gate position, stock trough, silo. Cheap; the speed of sound drifts with temperature, so calibrate for it and weather-seal the transducers.
Temperature probe, DS18B20. One-wire, comes in a waterproof stainless probe. Reliable, tiny current draw. If soil is your target, the same low-draw thinking runs through soil moisture monitoring without agtech prices.
Add as needed. An air-quality or dust and smoke sensor (fire and dust early warning), and a tilt or accelerometer sensor (tamper or knock-over alert).
An independent health logger. The part people skip and regret. A small, separate watchdog, its own tiny MCU or a hardware watchdog, that logs battery voltage, box temperature and uptime, and reports even when the main application has hung. Without it a silent box is ambiguous: is nothing happening, or is the box dead? The health logger answers that.
Build notes that keep it alive in a paddock
Seal it properly. Enclosure IP65 or better, cable glands on every entry, and a breather or membrane vent so condensation does not pool inside.
Aim the panel. Facing north for the southern hemisphere, tilted roughly to your latitude, sited where a tree or the tank will not shade it at 4 pm.
Mind the cold. LiFePO4 will not accept charge below about 0 degrees C, a real problem in frost country. Insulate the battery, add a low-power self-heater, or accept a winter derate and size for it.
Protect the battery from itself. A low-voltage disconnect so a bad week does not deep-discharge it to death, plus brown-out protection on the MCU.
Fuse both lines. Fuse the panel and the battery lines, close to the source.
The design rule underneath all of it is the same one behind how LoRa sensor monitoring works: put the always-on parts where power exists, and let the paddock devices be the sleepy ones.
The short version
Size power first. A load around 1.24 A becomes about 30 Ah per day, which sets a 40 Ah LiFePO4 as the floor with two to three days of reserve added, and about 150 W of solar for clear-sky need but about 300 W actually fitted for cloudy weeks, on an MPPT controller. Sensor kit: RCWL-0516 radar, HC-SR04 ultrasonic, DS18B20 temperature, plus air-quality or tilt as needed, and a separate health logger so you can tell quiet from dead. Seal it IP65, face the panel north, and mind LiFePO4's cold-charging limit.
General build guidance. Size, fuse and protect your own system for your site and load; grid-tie or mains work needs a licensed electrician.
Want a box that runs itself?
Tell us what you need to sense and where it lives. We will size the power, pick the sensors, and build a box that is still reporting after the third cloudy week, no pressure, no jargon.
Get a quote