An undersized solar and battery setup is one of the most common and most avoidable reasons a field sensor node fails. It usually doesn't fail on a sunny day. It fails after two or three days of heavy cloud cover, which is exactly when continuous monitoring (of dissolved oxygen ahead of a storm, for instance) matters most. Sizing this properly at design time is far cheaper than an emergency field visit later.
Start with actual power consumption, not the datasheet's best case
The first step is measuring, not estimating, what your node actually draws in each of its operating states: deep sleep between readings, active sensing, and (the big one) radio transmission, which draws far more current than most people expect for the brief period it's active. Datasheet current figures are useful starting points but often reflect ideal conditions; measure your actual assembled node with a proper current meter across a full duty cycle if at all possible.
Once you know average current draw, calculate daily energy consumption in watt-hours: average current in amps, multiplied by system voltage, multiplied by 24 hours. This single number is the foundation everything else is sized against.
Sizing the battery: days of autonomy, not just daily need
"Autonomy" is the number of consecutive days a system needs to run on battery alone with zero solar input, and it's the single most important design decision in the whole system. Two or three days of autonomy might be fine for a low-stakes application in a sunny climate. For a monitoring node whose whole purpose is catching an event that often coincides with bad weather (an upwelling event is frequently associated with storms and heavy cloud cover), we recommend sizing for at least five to seven days of autonomy as a starting point, adjusted based on how critical continuous uptime is to the application.
Battery capacity in amp-hours is then: daily energy consumption in watt-hours, divided by system voltage, multiplied by your target autonomy days, then increased to account for the fact that most battery chemistries shouldn't be regularly discharged to 100% depth of discharge. For lead-acid, keeping typical discharge to around 50% of rated capacity significantly extends usable lifespan; for lithium chemistries, deeper discharge is generally tolerated but still benefits from some margin.
Sizing the solar panel: for the worst realistic month, not the best
The panel needs to fully recharge the battery's daily consumption within the available sunlight of your worst realistic month, not an average or best-case month. Using local solar irradiance data (average peak sun hours for the region, which for much of Kenya's lake basin area is a reasonable few hours per day even accounting for cloud cover, but varies meaningfully by season) divide daily energy need by peak sun hours to get a baseline panel wattage, then apply a real-world derating factor.
That derating matters more than people expect: panel soiling from dust (relevant near the lake, especially during dry, windy periods), less-than-ideal mounting angle or orientation, temperature effects on panel efficiency, and charge controller and wiring losses together commonly account for a combined 20 to 30 percent reduction from a panel's rated output in real field conditions. Size the panel against derated output, not the number printed on it.
A worked example, in outline
For a node consuming an average of 50mA at 3.7V continuously (a modest but realistic figure for a sensor node with periodic radio transmission): daily consumption is roughly 4.4 watt-hours. For seven days of autonomy at 3.7V with a 50% usable depth of discharge for a lead-acid-equivalent margin, that points toward a battery in the range of 15 to 17 amp-hours. For a location averaging around 4.5 peak sun hours in its cloudier season, and applying a 25% derating factor, the panel needs to be sized meaningfully above the naive daily-need-divided-by-sun-hours number, comfortably into the small panel range rather than the smallest one available. The exact numbers shift with your specific node's consumption and site conditions, which is why measuring your actual hardware matters more than following someone else's number directly.
Building in margin, deliberately
Every input in this calculation (consumption, irradiance, derating) carries some uncertainty, and they compound in the direction of underperformance more often than overperformance in the field. We size with deliberate margin on both the battery and panel rather than the calculated minimum, because the cost difference between a slightly larger panel and battery is small compared to the cost of a field visit to fix a node that goes dark during exactly the weather event it was meant to catch.
If you're specifying power for a new field node and want help sizing it against your actual site conditions and consumption profile, get in touch. You can also see how our water-quality monitoring works.
