Solar Panel Sizing Calculator

Solar panels are rated under Standard Test Conditions: irradiance of 1,000 watts per square metre, a cell temperature of 25°C, and a defined atmospheric spectrum. A 400 W panel produces 400 W under exactly those conditions and less at any other time, which is nearly always.

A peak sun hour is the accounting device that makes this tractable. It is one hour at 1,000 W/m², so the day’s total solar energy divided by 1,000 gives the number of peak sun hours it delivered. A location with 4.7 peak sun hours does not get 4.7 hours of daylight — it might get fourteen in June and nine in December — but the total energy arriving over the year averages out to 4.7 full-strength hours a day.

The convenience of the unit is that it turns a complicated integral into a multiplication. Array kilowatts times peak sun hours times 365 gives annual kilowatt-hours before losses, and that is the entire sizing calculation.

The single most common sizing mistake is dividing annual consumption by peak sun hours and 365 and stopping there. That ignores everything between the panel’s glass and the utility meter, which removes about a seventh of the output.

A south-facing roof at a tilt near the site’s latitude is the reference case, and everything else is measured against it. East or west facing costs roughly 15% of annual output — less than most people expect, and one reason east-west installations have become common where the roof gives no choice. North facing in the northern hemisphere costs around 35% and is rarely worth doing.

Shading is different in kind, not just degree. Because panels in a series string carry the same current, one shaded panel can drag down the entire string far out of proportion to the area shaded — the classic demonstration is a single chimney shadow costing a third of an array’s output on a plain string inverter. Module-level electronics, either DC optimisers or microinverters, largely solve this by letting each panel operate independently, and they are close to standard on US residential installations partly for this reason and partly because the National Electrical Code’s rapid-shutdown requirements effectively demand module-level control.

The practical point is that a shading estimate made by eye in July is worthless. The sun sits far lower in winter and shadows are much longer, so a roof that looks clear in summer can be shaded for hours a day in December. Any installer worth hiring will measure this with a shade analysis tool rather than guess at it.

Section 25D of the tax code, the Residential Clean Energy Credit, gave homeowners 30% of the total installed cost of a solar system back as a federal tax credit — with no cap, making it by far the largest incentive in US residential solar. The One Big Beautiful Bill Act, enacted 4 July 2025, terminated it for property placed in service after 31 December 2025.

A system installed and commissioned in 2026 therefore receives no federal residential credit, regardless of when the contract was signed or the deposit paid. On the $25,200 system in the example above, the credit was worth $7,560, and its removal changes the economics of residential solar more than any market movement in the past decade.

One route remains open. The Section 48E commercial credit still applies to solar owned by a third party, which covers leases, power purchase agreements and prepaid arrangements — so a homeowner who does not own the system can still see the benefit passed through in the pricing. Whether that is a good deal is a separate question with a long history of mis-selling attached to it, and it deserves the same scrutiny as any twenty-year contract.

State credits, utility rebates and property-tax exemptions are untouched, and remain substantial in several states. Enter your cost net of those.

Sizing an array to cover exactly a year’s consumption feels like the natural answer, and under old-style net metering it usually was: every exported kilowatt-hour rolled back the meter at full retail value, so generation and consumption were interchangeable and only the annual total mattered.

That arrangement has been narrowing across the US. Successor tariffs — California’s NEM 3.0 being the most consequential example — pay export at a wholesale-derived rate that can be a fifth of the retail import price. Under those rules a kilowatt-hour you use as it is generated is worth several times one you export, and the economics shift from "how much do I generate" to "how much do I self-consume". That favours a somewhat smaller array, or the same array paired with a battery to shift midday generation into the evening.

The argument for oversizing runs the other way and is about the future rather than the tariff. Adding panels to an existing system is disproportionately expensive: the scaffolding, the permit, the interconnection application and the site visit all recur, and the inverter may not have headroom. A household likely to add an electric car, a heat pump or a heat pump water heater within a few years is usually better off building that capacity in at the start, when the marginal cost is a panel and a rail rather than a whole second project.