Tool
Solar sizing for an always-on node
Estimate the solar panel and 18650 battery an always-on LoRa node needs in a UK winter, using December sunshine from the EU's PVGIS service. A planning aid, not a guarantee.
Updated 29 Oct 2026
Result
Suggested setup
3 W panel + 1× 18650 (3000 mAh nominal each)
For a 25 mA node at London, panel tilted 60 degrees, sized for an average December. The panel size is rounded up to a common rating; the cell count assumes 3000 mAh per 18650.
Daily energy
- Average current
- 25 mA
- Energy per day
- 2.22 Wh
- 25 mA × 3.7 V × 24 h. That is 600 mAh per day at 3.7 V.
Battery
- Energy to cover the dark days
- 6.7 Wh
- 2.22 Wh × 3 days.
- Battery to buy (nominal)
- 10.4 Wh
- Divided by depth of discharge 80 % and cold factor 0.8. About 2,812 mAh at 3.7 V.
- 18650 cells of 3000 mAh
- 1
- Rounded up. Cells wired in parallel; check that your board or battery holder supports that.
Solar panel
- December sunshine on the panel
- 1.69 kWh/m²/day
- PVGIS PVGIS-SARAH2, 2005 to 2020, fetched 2026-10-29. Numerically the same as 1.69 peak sun hours per day.
- Panel needed (rated watts)
- 2.2 W
- 2.22 Wh ÷ (1.69 h × 60 % efficiency).
- Nearest common rating, rounded up
- 3 W
How it works
Everything runs in your browser; the sunshine data is built into the page and nothing is fetched when you use it.
- Daily energy. Average current (mA) × battery voltage (V) × 24 h ÷ 1000 = watt-hours per day. A 25 mA node at 3.7 V uses about 2.2 Wh a day.
- Battery. Daily energy × days of autonomy, divided by the depth of discharge and by a cold-derating factor. Divided by the voltage, that gives milliamp-hours. The recommendation rounds up to whole 18650 cells of 3,000 mAh nominal each.
- Panel. Daily energy ÷ (December sunshine in peak sun hours × system efficiency) = rated panel watts. "Peak sun hours" is the day's solar energy per square metre in kWh, which is the same as hours of full 1 kW/m² sun. The result is rounded up to a common rating (1, 2, 3, 5, 6, 10, 12, 15, 20, 30 W).
Assumptions you should check
Every factor below is a planning assumption we chose, not a measurement and not a manufacturer figure. Change them in the tool to see how sensitive the answer is.
| Item | Options (default marked) | What it stands for |
|---|---|---|
| Device current | 25 mA, 80 mA, 150 mA, or your own | Unsourced planning placeholders (assumed values, not measurements) for an nRF52 node with no GPS or screen, an ESP32 node with Wi-Fi and screen off, and an ESP32 node with GPS and screen on. Our power guide gives no draw figures because we could not verify any from a source we could open, so treat these as placeholders and measure with a USB power meter. |
| Depth of discharge | 90 %, 80 %, 60 % (80 %) | How much of the cell's capacity you plan to use. Using less is gentler on the cell. |
| Cold derating | 1, 0.8, 0.6 (0.8) | A cold cell delivers less than its label capacity. These are rough planning factors, not measured curves. |
| System efficiency | 80 %, 70 %, 60 %, 50 % (60 %) | Everything between the panel's rated output and energy stored in the battery: charge controller, wiring, dirt, panel temperature and battery round trip. |
| 18650 cell | 3,000 mAh nominal | A single typical figure for the recommendation. Check the label on the cells you buy; capacities vary. |
| Voltage | 3.7 V | Nominal voltage of one lithium-ion cell. Your measured current should be at the battery voltage. |
December sunshine, ten UK places
Values are kWh per square metre per day (equal to peak sun hours) from the EU Joint Research Centre's PVGIS service. We fetched them on 2026-10-29 from the PVGIS EU JRC PVGIS API v5.2 (MRcalc) monthly-radiation endpoint, which returned the PVGIS-SARAH2 radiation database with ERA5 meteorology for 2005 to 2020 (16 Decembers) and includes the shading of the terrain horizon. It does not include shade from nearby buildings or trees. Panels face due south. "Average" is the mean of the 16 Decembers; "worst" is the lowest single December in that period.
| Place | Flat, avg | 45°, avg | 60°, avg | 60°, worst |
|---|---|---|---|---|
| Plymouth | 0.72 | 1.50 | 1.61 | 1.00 |
| London | 0.69 | 1.55 | 1.69 | 1.11 |
| Cardiff | 0.63 | 1.38 | 1.49 | 1.01 |
| Birmingham | 0.62 | 1.42 | 1.55 | 1.19 |
| Manchester | 0.52 | 1.19 | 1.30 | 0.98 |
| Belfast | 0.47 | 1.14 | 1.26 | 0.97 |
| Newcastle upon Tyne | 0.47 | 1.26 | 1.40 | 0.95 |
| Edinburgh | 0.38 | 0.99 | 1.09 | 0.68 |
| Inverness | 0.25 | 0.64 | 0.71 | 0.33 |
| Lerwick (Shetland) | 0.13 | 0.37 | 0.41 | 0.25 |
- We checked all twelve monthly means for every place and plane when pulling the data. December is the lowest month everywhere except Manchester at 45° and 60°, where January is lower; the tool tells you when that applies.
- Flat panels do far worse in December than steep ones because the winter sun is low. The tool defaults to 60°, which favours winter; a panel lying flat gets far less.
- The ten places are a spread, not a promise about your garden. Pick the nearest and expect differences with altitude, coast and local cloud. For an exact site, use the PVGIS interactive tool.
What this does not tell you
- Recovery. Sizing the panel to just cover an average December means a battery drained by a dark spell refills only on brighter days. If a node must come back quickly, size up.
- Shade and direction. Trees, buildings, a panel facing away from south, snow cover and dirt all reduce what you collect, and none is modelled.
- Cold. General lithium-ion guidance is not to charge cells below 0 °C. Check that your charge controller or board protects against it; we did not verify any particular board does.
- Safety. Lithium cells can catch fire if shorted, overcharged or connected the wrong way round. Read the hazards in the power guide before wiring cells.
- Integrated products. Meshtastic's documentation lists ready-made solar repeaters from RAK Wireless; for a one-off node a ready-made unit may be simpler than building your own.
Next: power, link budget calculator, antenna placement.