Solar sizing runs backwards. You do not start with a panel and hope; you start with the load, in daily watt-hours, then decide how many gray days you must survive, size the battery through its depth of discharge, and size the array for your worst month of sun with an honest derating. Four short steps of arithmetic, and the order is the method.
Start at the load.
The load inventory is the whole foundation: every device, its watts, and its honest hours per day. Watts × hours = watt-hours, summed down the list. Two honesty rules keep the list from lying. Measure what you can rather than trusting nameplates: a plug-in power meter for AC equipment, the spec sheet or a bench measurement for DC gear, because nameplates state the maximum, not the habit. And respect duty cycle: a 60 watt pump that runs ninety minutes a day is a 90 Wh load, not a 1,440 Wh one, while a 5 watt inverter standby that never sleeps is a quiet 120 Wh a day, often the biggest thing on a small list.
Size the battery.
Two decisions turn daily watt-hours into a battery. Days of autonomy: how many sunless days the system must ride through on storage alone. Two to three is the usual farm answer; one is optimistic, five is an expedition. Then divide by the chemistry's sane depth of discharge, because the tank is bigger than the part of it you may use:
Battery watt-hours = daily load × days of autonomy ÷ DoD.
Chemistry and its manners live on Batteries and Storage; the short version is 0.5 for lead-acid and 0.8 for LiFePO4.
Size the array.
The array's job is to replace a day's spending during a day's sun, in the month when the sun is stingiest. Your location's sun hours (the day's light expressed as equivalent hours of full-strength sun) is the honest input, and it is a December number, not a June one: the solar curve visualizer shows the shape of that valley at your latitude. Then apply derating, the discount real life takes from the panel's nameplate: angle, heat, dust, wire loss, controller loss. Planning at 0.7 to 0.8 of nameplate is honest; planning at 1.0 is how systems die in December.
Array watts = daily load ÷ worst-month sun hours ÷ derating.
A worked example.
A stock-tank setup in Tennessee: a 60 watt DC transfer pump that runs about ninety minutes a day (90 Wh) and a tank-level monitoring node (about 10 Wh a day). Call it a clean 100 Wh daily load, all DC, no inverter.
Battery, with two days of autonomy on LiFePO4: 100 × 2 ÷ 0.8 = 250 Wh, which at 12 volts is about 21 amp-hours; the common 12 V 25 Ah pack fits with a little margin.
Array, for a Tennessee December of roughly two sun hours at 0.75 derating: 100 ÷ 2 ÷ 0.75 ≈ 67 watts, so the standard 100 watt panel with room to spare. Notice the shape of that answer: the December sun, not the pump, decided the panel. In June the same panel is loafing, which is exactly how it should be.
Controller, wire, fuse.
The remaining sizes follow mechanically. The charge controller must carry the array's current with about 25 percent margin: a 100 watt panel near 12 volts pushes roughly 8 amps, so a 15 amp controller covers it honestly (PWM or MPPT is its own question). Wire is sized to the current and the run length, and the fuse is sized to the wire: that chain of reasoning, ampacity and voltage drop and DC ratings, is the safety page's whole subject, and it is the part of sizing where mistakes burn rather than merely disappoint. For pump systems specifically, the solar pump sizing tool walks this whole chain with your numbers.
The shortest version
Count the daily watt-hours honestly, duty cycle and standby included. Battery = daily load times days of autonomy divided by DoD (0.5 lead-acid, 0.8 LiFePO4). Array = daily load divided by worst-month sun hours divided by 0.75. Then controller with 25 percent margin, wire to the run, fuse to the wire. December decides; June coasts.
Words to work from
Take these terms with you. They are the sizing conversation's whole vocabulary.
- Load inventory
- The list: every device, its watts, its honest daily hours. The foundation everything else stands on.
- Duty cycle
- The fraction of time a load actually runs. A pump is its minutes, not its nameplate.
- Watt-hour
- Watts times hours: the unit loads, batteries, and days all share.
- Days of autonomy
- How many sunless days storage alone must cover. Two to three is the usual farm answer.
- Depth of discharge (DoD)
- The usable fraction of the battery. Divide by it, or the battery is smaller than you think.
- Sun hours
- A day's light as equivalent hours of full sun. Use the worst month's number.
- Worst month
- The design month, usually December. Size for it and summer takes care of itself.
- Derating
- The real-world discount on nameplate: angle, heat, dust, losses. Plan at 0.7 to 0.8.
- Nameplate
- The rating printed on the panel, achieved in laboratory sun only.
- System voltage
- The battery voltage everything must match: 12, 24, or 48 volts. Higher voltage means thinner wire for the same power.
Frequently asked questions.
How many watts of solar panel do I need?
Divide your daily watt-hours by your worst month's sun hours, then by a derating of about 0.75. A 100 Wh daily load over two December sun hours needs about 67 watts of panel, so a 100 watt panel with margin. The load and the December sun decide the number; the panel aisle does not.
What size battery do I need for solar?
Daily watt-hours, times the sunless days you must ride through (two to three is typical), divided by the chemistry's usable depth of discharge: 0.5 for lead-acid, about 0.8 for LiFePO4. A 100 Wh daily load with two days of autonomy on lithium wants roughly 250 Wh, about a 12 volt 25 Ah pack.
Why size for winter instead of the yearly average?
Because the system fails in its worst month, not its average one. An average hides December inside June: the array that averages fine can still leave the battery drained for weeks in a row in the shortest days. Size for the worst month and the rest of the year is free surplus, which batteries and equipment both prefer.
Can I add more panels later?
Usually, within two limits. The charge controller must have current and voltage headroom for the new array, and new panels should electrically match the old ones in each string, since a series string obeys its weakest panel. Buying a controller one size up at the start is the cheap way to leave that door open.