The quietest solar job on a farm is keeping the Collect layer alive: a soil probe at the far fence, a water-level sensor on a stock tank, a weather node past the tree line. None of them will ever see an outlet, and none of them need one. At this scale solar stops being a construction project and becomes a component: a panel the size of a paperback, a charge board, and one lithium cell will run a sensor node indefinitely, if you do the small math and point the panel at the winter sun.
The job: milliamps, not amps.
A well-built sensor node spends almost its whole life asleep. It wakes on a wake interval (every ten minutes, say), reads its sensors, pushes the readings, and goes back to sleep, where a good design draws a trickle (sleep current) measured in millionths of an amp. What the battery feels is the average: seconds of real work per hour diluted into a draw of a few milliamps. That average, times 24 hours, is the day's budget in milliamp-hours, and it is tiny. This is why the node problem and the pump problem are different subjects wearing the same word: the sizing method is identical, but the numbers shrink a thousandfold, and oversizing becomes cheap instead of expensive.
The four parts.
The node chain is the five-part chain with the inverter deleted and everything shrunk. A small panel, two to ten watts, glass or semi-flexible. A charge and protection board, the node-scale charge controller: a small module that charges the cell correctly, refuses to over-discharge it, and hands the node clean power; at this scale the controller is a circuit board the size of a stamp, not a box on the wall. A cell: the common cylindrical 18650 lithium cell or a small LiFePO4 pack. And the node itself, an ESP32-class board running a program from the Software Library, in a vented, shaded enclosure, because electronics cooked in a sealed box in direct sun age like milk. The physical parts live in the Hardware guide.
The math, at node scale.
A worked miniature: a node that wakes every ten minutes averages about four milliamps, which over a day is roughly 100 mAh, call it half a watt-hour. One 18650 cell holds around 12 watt-hours, so the battery alone rides out three weeks of dark sky, an autonomy a pump system could never afford. A two-watt panel in two honest December sun hours at 0.7 derating brings in nearly three watt-hours, five times the daily spend. Notice what happened: at node scale the worst-month rule is easy to satisfy lavishly. Buy the five-watt panel instead of the two and the whole question disappears for the life of the node.
Placement is the engineering.
With the electrical margin that generous, node failures are almost always placement failures. Tilt the panel steeply, latitude plus about fifteen degrees, so it faces the low winter sun and sheds snow rather than wearing it. Mount it above the snow line and the splash line. Check the shade in December, not July: a fencepost shadow that misses the panel all summer can park on it all winter. And wipe it when you walk past; dust is a slow, free derating. For nodes past Wi-Fi's reach, the reading travels by LoRa or another long-range link, which is Connectivity's subject; the power story does not change.
The cold-charging catch, again.
The lithium rule follows the chemistry down to a single cell: charging below freezing damages lithium permanently. A node that ran flawlessly all summer and died in February usually spent the winter charging frozen. The honest options at node scale: a charge board with a low-temperature cutoff (a temperature sensor on the cell, charging paused below freezing, and the three-week autonomy carries the node through the pause), a LiFePO4 cell chosen for its wider manners, or the enclosure placed where winter sun warms it above freezing during charging hours. What does not work is hoping.
The shortest version
A sleeping node averages milliamps, so a paperback-sized panel and one lithium cell run it forever: half a watt-hour a day against three watt-hours of December harvest and three weeks of battery autonomy. Oversize the panel, tilt it steep for winter, mount it above snow and out of December shadows, and give the cell a charge board that refuses to charge below freezing.
Words to work from
Take these terms with you. They are the vocabulary of the product listings and the forum threads that solve this exact problem.
- Sleep current
- What the node draws while sleeping, in microamps. The number that decides battery life more than any other.
- Wake interval
- How often the node wakes to read and report. Minutes of sleep bought per second of work.
- Milliamp-hour (mAh)
- The small-battery capacity unit. A thousandth of an amp for an hour.
- 18650
- The common cylindrical lithium cell, roughly 12 watt-hours in one finger-sized can.
- Charge and protection board
- The node-scale charge controller: charges the cell correctly and refuses to over-discharge it.
- Low-temperature cutoff
- The board feature that pauses charging below freezing. The difference between a node that survives February and one that does not.
- Tilt angle
- The panel's lean from horizontal. Latitude plus about fifteen degrees favors the winter sun and sheds snow.
- Self-discharge
- What a cell loses sitting idle. Small for lithium, but part of the winter budget.
- Brownout
- The node crash from sagging supply voltage as a battery empties. A protection board prevents the deep version.
- LoRa
- The long-range, low-power radio that carries a reading miles when Wi-Fi cannot follow.
Frequently asked questions.
How big a solar panel does an ESP32 sensor node need?
Two watts covers a well-behaved sleeping node even in December, and five watts makes the question disappear. The node's average draw is a few milliamps, roughly half a watt-hour a day, while even a small panel in weak winter sun brings in several times that. Spend the spare dollars on the panel; at this scale oversizing is cheap insurance.
Do I need a charge controller for a tiny panel?
Yes, but at this scale it is a small charge and protection board, not a wall-mounted controller. The board charges the cell on the correct curve, stops at full, and disconnects the node before the cell over-discharges. Wiring a panel straight to a lithium cell is how cells swell and burn; the board costs a few dollars.
Will a solar sensor node survive winter?
Electrically, easily: the battery margin is weeks. The two real winter killers are placement and cold charging. Tilt the panel steeply so it sheds snow and faces the low sun, check for December shadows, and use a charge board with low-temperature cutoff so the cell is never charged below freezing, which is what actually kills nodes in February.
How do I weatherproof the node?
Shade and ventilation matter as much as rain. A vented enclosure under an eave or its own small shade, with cable glands pointing down and a desiccant pack inside, outlives a sealed box in direct sun, which cooks the electronics and condenses moisture at night. Keep the panel in the sun and the electronics out of it.