How many solar panels do I need

When someone asks how many panels they need, they are almost always asking something else: what will this cost. And the panel count is, oddly enough, the last figure you calculate. First you need to know how much energy you want to back up, how much battery it takes to store it, and how much sun your roof actually receives. Panels fall out of that, by division.

This guide follows the same order as the calculator, which is the order any installer follows in a spreadsheet. The difference is that here you see where each factor comes from, so you can argue with a quote instead of accepting it.

First: energy, not power

This is the mistake that inflates the most budgets. Watts (W) measure power: how fast an appliance consumes. Watt-hours (Wh) measure energy: how much it consumed in total. Panels are sized by energy, not by power.

The classic example is the refrigerator. The label says 150 W, and many people multiply 150 W × 24 h and conclude it uses 3,600 Wh a day. It does not: the compressor does not run continuously, it cycles on and off to hold temperature. In a hot climate, a modern unit runs about 40 % of the time. Its real consumption is closer to 150 W × 9.6 h ≈ 1,440 Wh a day.

The reverse happens with the water pump. A 750 W pump looks enormous next to the fridge, but if it starts for five minutes an hour it uses about 60 Wh in that interval. It is a high-power, low-energy appliance: it matters enormously to the inverter and very little to the battery.

For outage backup there is one more nuance: you do not care about all-day consumption, only about the hours you are off the grid. If your outages last six hours, the calculation covers those six hours and only what you genuinely want to keep running in them.

Second: the battery comes before the panels

In a grid-tied system without batteries, panels are sized against consumption. In a backup system they are sized against the battery, because the panel’s job is to leave the bank full before the next outage.

Bank capacity is not equal to your consumption: you have to leave margin. No battery should be discharged to 100 % every day, not even lithium. A 20 % margin is the reasonable minimum, and it is what this calculator uses.

Watch for a detail datasheets hide: nominal capacity is not usable capacity. A lead-acid battery of 2.4 nominal kWh only gives you about 1.2 kWh without shortening its life, because you should not go past 50 % discharge. A 2.4 kWh LiFePO4 gives you close to 2.16 kWh. If you compare two quotes by nominal capacity, you are comparing different things — that thread is picked up in the guide on lithium versus lead-acid.

Third: peak sun hours, not daylight hours

Another frequent misunderstanding slips in here. The tropics get about twelve hours of daylight, but that does not mean twelve hours of generation. The useful measure is called a peak sun hour (PSH): one peak sun hour equals receiving 1,000 W per square metre for one hour, which is the condition under which a panel’s rating is measured in the factory.

A real day does not deliver 1,000 W/m² from sunrise. At 7 in the morning irradiance is a fraction of that; around noon it peaks; by 4 in the afternoon it drops again. Integrate that whole curve and a clear day in a sunny region comes to about 5 to 5.5 peak sun hours. With afternoon cloud it falls to 3.5 or 4. In a northern European winter, to 1.

• Arid and desert — Sahara, Persian Gulf, Atacama, US Southwest, Australian interior: 5.5 to 6.5 PSH.
• Dry tropical and savanna — much of sub-Saharan Africa, India, northern Australia, Pacific Latin America: 5 to 5.5 PSH.
• Mediterranean and subtropical — Spain, Italy, Greece, Türkiye, California, central Chile, South Africa: 4.5 to 5.5 PSH.
• Humid tropical and equatorial — Amazon, Southeast Asia, Congo basin, Caribbean coasts: 4 to 4.5 PSH; heavy afternoon cloud.
• Continental temperate — central and eastern Europe, US Midwest, northern China, Japan: 3 to 4 PSH.
• Maritime temperate — UK, Ireland, Netherlands, Scandinavia, Pacific Northwest: 2.5 to 3.5 PSH.

The calculator proposes a value from your browser’s time zone, but it is worth confirming: the Global Solar Atlas, from the World Bank, gives the exact figure for any point on the planet and it is free. And if you want the system to carry the cloudy season without falling back on the grid, size it on the worst month rather than the annual average: the array comes out 25 to 40 % larger, and that extra cost is exactly the price of not sitting in the dark in the bad month.

Fourth: nothing arrives intact from panel to battery

Between roof and battery, energy leaks along the way. The system efficiency factor gathers all those losses into one number, and for a well-executed residential install it is around 0.8. It is worth knowing where that 20 % goes:

• Temperature: 8 to 12 %. A panel is rated at 25 °C. On a metal roof under strong sun the cell easily passes 55 °C, and every degree above cuts output. It is the largest loss and the most ignored.
• Conversion and charging: 4 to 8 %. The MPPT controller and the inverter take their cut transforming the current.
• Wiring: 2 to 3 %. Proportional to distance and gauge. Thin cable to a far roof eats more than it looks.
• Dirt and dust: 2 to 5 %. In the dry season, with dust and no rain to wash it off, it shows. A rinse every two months recovers it.
• Tolerance and degradation: 1 to 3 %. Panels lose about 0.5 % of capacity per year.

If someone quotes you without applying any loss factor, the system will come out 20 to 25 % short — and you will find out in the first cloudy month.

The complete formula

Rounding up is not optional. Half panels do not exist, and falling short means the bank never finishes charging, which kills the battery over time. If the result is 1.2 panels, you buy 2.

Worked example

A house with six-hour evening outages wants to back up the essentials: fridge, six LED bulbs, two fans, the router and the chargers.

1. Energy to back up: 1,530 Wh a day.
2. Recommended bank: 1,530 × 1.2 = 1,836 Wh, that is 1.8 kWh.
3. Required array: 1,836 ÷ (5 × 0.8) = 459 W.
4. Panels: 459 ÷ 550 = 0.83 → 1 panel of 550 W, which once installed generates around 2,200 Wh a day.

A single panel. It surprises people, and it is correct: backing up the essentials for six hours is a small problem. The battery is the expense, not the roof. Now add a 12,000 BTU air conditioner running half the outage: 1,200 W × 3 h = 3,600 Wh more. The total jumps to 5,130 Wh, the bank to 6.2 kWh and the array to 1,540 W, which means 3 panels. One appliance tripled the entire system.

Why the hour of the outage changes the answer

Two homes with identical consumption can need different systems depending on when they lose power, and this rarely shows up in a quote.

If the outage runs 18:00 to midnight, panels generate nothing during the event. Every watt-hour has to be stored in advance and the bank works at full stretch every day. If the outage runs 10:00 to 16:00, panels can feed the house directly while refilling the bank: same battery, considerably more runtime, and in some hours the bank does not discharge at all.

Timing does not change the panel count (the array is sized to refill the bank over a full day), but it does change how much bank you truly need and how much margin you have on a bad day. That is why the calculator asks you to mark hours on a 24-hour strip instead of just typing a number.

Five mistakes that keep repeating

1. Sizing from the electricity bill. The bill includes the iron, the microwave and the pump: things you will not back up. Sizing that way can double the budget without buying you one extra minute of backup.
2. Ignoring the temperature factor. On hot roofs it is the largest loss and the one almost nobody mentions in a quote.
3. Comparing batteries by nominal capacity. Without knowing the allowable depth of discharge, two identical figures can mean twice the usable energy in one over the other.
4. Choosing the inverter from average consumption. The inverter is chosen from startup surge. A 750 W pump can demand 2,250 W for a second, and a 1,000 W inverter shuts down in that instant.
5. Forgetting the system’s own draw. A hybrid inverter burns 20 to 50 W just being on. Over 24 hours that is 0.5 to 1.2 kWh: possibly more than your router and lights combined.

What to ask whoever quotes you

• The load table, appliance by appliance, with power and hours. If they do not have one, they did not size it: they guessed.
• The peak sun hours they used and why. If they say “ten hours of sun”, the calculation is wrong by definition.
• The loss factor applied. It should sit between 0.75 and 0.85.
• Usable bank capacity in kWh, not nominal amp-hours.
• Continuous and surge inverter ratings, and the transfer time in milliseconds.

With those five data points you can genuinely compare two quotes. To arrive with your own numbers, run the calculator and take the printed result with you: it does not replace a professional design, but it completely changes the conversation.