Solar panel payback and sizing, home batteries, and backup power — the evergreen math, no incentive chasing.
Bifacial solar panels collect light on both faces, adding roughly 5–20% output over a monofacial panel — but only when the surface beneath reflects. Over white gravel, a light membrane roof or snow they earn the premium; flush-mounted on dark shingles with a few inches of standoff, the rear face collects almost nothing.
Solar panels work by the photovoltaic effect: photons knock electrons loose in doped silicon cells, and the cell’s built-in electric field pushes them in one direction, producing DC. An inverter converts that DC to 240 V AC synchronised with the grid, which then feeds house loads first and exports or stores the surplus.
Monocrystalline panels reach 19–23% efficiency against polycrystalline’s 15–17%, and mono has effectively won the market — most manufacturers have discontinued poly as the price gap closed. The remaining question is efficiency per square foot, which only matters when roof area is the binding constraint.
Panel efficiency is watts per square metre, not quality or value — two 400 W panels produce the same energy, the more efficient one is simply smaller. Mainstream panels run 18–20% and premium 21–23%, and paying for efficiency only helps when roof area is what limits your system.
Five numbers on a panel datasheet drive the design: rated wattage (Pmax), open-circuit voltage (Voc), short-circuit current (Isc), efficiency, and the temperature coefficient. Voc rises as temperature falls, so the cold-weather Voc is what sets how many panels can go in a string without damaging the inverter.
Solar involves three separate warranties from three parties: the panel manufacturer’s product and performance warranties (typically 25 years, 80–90% output retained), the inverter manufacturer’s (10–25 years), and the installer’s workmanship warranty covering roof penetrations — the last is the one that most often fails.
A solar array is the complete set of interconnected solar panels on a site, wired into strings that feed one or more inverters. Array size is set by three limits at once: your energy use, the usable roof or ground area, and what your electrical panel can accept back.
A DC-to-AC ratio of 1.15–1.25 is the common design range: the array is deliberately oversized against the inverter, so a few peak hours get clipped in exchange for more production in the many hours either side. Annual clipping losses at that ratio are usually low single digits.
Solar panels work in winter, and cold actually makes cells more efficient. Winter production drops because of shorter days and a lower sun angle, not temperature — and snow blocks output only while it sits on the glass, sliding off tilted panels faster than most people expect.
The panel count comes from your annual kWh, not your square footage: annual kWh divided by (peak sun hours × 365 × 0.8) gives the kW you need, then divide by panel wattage. A home using 12,000 kWh a year at 4.5 peak sun hours needs roughly 9.1 kW — about 23 panels at 400 W each.
Solar panel output is rated wattage × peak sun hours × a derate of about 0.8. A 400 W panel at 4.5 peak sun hours makes roughly 1.4 kWh a day, or about 525 kWh a year — the derate covers inverter losses, heat, wiring, dust and angle.
Shading is non-linear on an unoptimised string: the shaded panel limits the current for every panel in series with it, so a small shadow can cost far more than that panel’s share. Power optimisers or microinverters isolate the loss to the affected panel — the whole argument for them on a complex roof.
Peak sun hours are daily solar energy expressed as equivalent hours at full intensity (1,000 W/m²), not hours of daylight. Most of the US sits between 3.5 and 6.0, and it is the figure that makes production arithmetic work: system kW × peak sun hours × 365 × 0.8 gives annual kWh.
South-facing maximises annual production; east or west gives roughly 80–87% of that. Tilt matters far less — anything from 15° to 40° lands within a few percent of ideal. Under time-of-use rates, west-facing can be worth more despite lower output, because it produces into the expensive evening window.
Panels are rated at 25°C cell temperature and rarely operate there — a roof array in summer commonly runs 45–65°C. With a typical temperature coefficient of −0.3% to −0.4% per °C, that is a 6–15% loss, which is why a cool bright spring day can out-produce a hot July afternoon.
A mid-latitude US array produces roughly 33% of its annual output in summer, 28% in spring, 23% in autumn and 16% in winter. Spring often beats summer per day, because cooler panels lose less to heat — and whether the seasonal imbalance matters depends entirely on what exported kWh are worth.
The costly sizing mistakes are ignoring the NEC 705.12 back-feed limit, which can fail interconnection review, and sizing to last year’s bill without accounting for an EV or heat pump arriving later. Forgetting the 0.8 derate overstates production by about 20%.
Compare installers on estimated annual kWh per dollar rather than on price per watt, and confirm each quote uses the same production assumptions. The other half of the decision is whether the company will still exist in year 12 to honour a 25-year workmanship warranty.
Ground mounted solar panels let you set perfect tilt and azimuth and keep the array reachable for cleaning and service — but you pay for foundations, a trench back to the house, and usually more permitting. Roof mount is cheaper when the roof orientation is decent.
Roof age is the first question: with under about 10 years of life left, re-roof before installing, because removing and reinstalling an array later commonly costs thousands. After that come orientation, pitch, shading through the year, structural capacity and available unobstructed area.
The physical install takes one to three days. The full timeline runs six weeks to six months, because permitting, utility interconnection review, inspection and permission to operate all happen around it — and the system must legally stay switched off until that last approval arrives.
Solar needs two separate approvals: a building and electrical permit from your AHJ, and an interconnection agreement from your utility. They run in parallel, and the utility one usually takes longer — it also caps system size through the NEC 705.12 120% busbar rule and local transformer capacity.
Racking attaches to the roof structure, and the flashing at each penetration is the part that must keep working for 25 years. Standing seam metal is the best case because clamps grip the seam with no penetration at all; tile is the most labour-intensive, and asphalt shingle the most common.
A string inverter converts DC for a whole series of panels at one central unit — cheaper and simpler, but the shaded panel limits the string. Microinverters convert at each panel, isolating shade losses and giving per-panel monitoring, at higher cost and with more units on the roof.
A power optimiser is per-panel electronics that condition DC output before it reaches a central string inverter — giving per-panel maximum power point tracking and monitoring, plus module-level rapid shutdown. It is the middle option between a plain string inverter and full microinverters.
Solar is worth it when your annual production times the value of each kWh pays back the net system cost inside the panels’ 25–30 year life. A 6 kW system at 4.5 peak sun hours makes about 7,900 kWh a year; at $0.18/kWh that is roughly $1,420 saved, so a $15,000 system pays back in about 10.5 years.
Simple payback is net system cost divided by annual saving, where the saving is production multiplied by what each kWh is actually worth to you — full retail under net metering, often far less for exports without it. Degradation of about 0.4–0.5% a year and any assumed rate escalation are the levers most often abused.
Without full net metering, exported kWh may earn only 25–40% of retail, which shifts the economics from producing to self-consuming. The design changes: a smaller array sized closer to daytime load, west-leaning orientation, load shifting into daylight, and a battery becomes much easier to justify.
Net metering credits exported solar at the full retail rate, so a kWh sent out offsets a kWh drawn later. Net billing and avoided-cost tariffs credit far less, often 25–40% of retail — and which arrangement your utility offers decides how large an array makes financial sense.
An owned, paid-off solar array generally adds value and conveys with the house. A leased array or PPA is a contract the buyer must qualify for and agree to assume, which is where sales stall — so start transfer paperwork early and keep production records and permits documented.
A loan or cash purchase means you own the system and any incentive; a lease or PPA means a third party owns it and you buy the power or rent the hardware. Ownership decides who captures the upside — and third-party contracts are the ones that complicate selling the house.
Time-of-use rates put peak pricing in the late afternoon and evening, exactly when a south-facing array stops producing. That shifts the design goal from most kWh to most valuable kWh — a west-facing array producing 10% less annually can be worth more, and a battery removes the problem entirely.
System size is the biggest single driver, with cost per watt falling as size rises. After that: roof complexity and pitch, racking type, and electrical panel work — the last being the most common cost surprise, when the panel cannot accept the back-feed under the 120% rule.
A DC-coupled battery connects on the DC side through a hybrid inverter, converting once and reaching round-trip efficiency around 95%. An AC-coupled battery has its own inverter and converts twice, closer to 90% — but it retrofits to existing solar without touching the original system.
Essential-loads backup powers a critical-loads subpanel — fridge, furnace fan, well pump, lights, internet — from a single 13.5 kWh unit. Whole-home backup means the battery must serve every circuit including HVAC and the range, which typically takes multiple units and a much larger inverter.
Usable capacity is nameplate multiplied by depth of discharge, and it is the only figure worth comparing. LFP allows 90–100% DoD, NMC lithium 80–90%, lead-acid just 50% — so a 10 kWh lead-acid bank gives 5 kWh while a 10 kWh LFP gives close to 10.
A home battery is worth it if you want silent instant backup, you are on a time-of-use rate worth arbitraging, or you have solar and no favourable net metering. A single common unit holds about 13.5 kWh usable — enough for essentials for a day or so, not for running a whole house indefinitely.
Home battery siting is governed by NEC Article 706 and UL 9540, with IRC R328 covering dwellings. The practical constraints are ambient temperature range, manufacturer clearances, and rules on installation in or near habitable spaces — location affects lifespan more than most buyers expect.
Battery warranties specify years, cycles and total energy throughput, and they end at whichever limit arrives first — typically 10 years, 6,000–10,000 cycles for LFP, with 60–70% capacity retained. Work out your expected cycles per year to see which limit actually binds for you.
A home battery system is cells storing energy chemically, a battery management system protecting and balancing them, an inverter converting DC to AC, and transfer equipment with islanding protection that safely disconnects from the grid. Without that last piece, a battery cannot back up a house at all.
Battery arbitrage is worth the peak-to-off-peak rate spread multiplied by cycles, minus 10–15% lost to round-trip efficiency. Pure financial payback is often long, so the honest framing is that a battery is bought for resilience with the arbitrage offsetting part of the cost.
Size a home battery from the job it has to do. Backup sizing comes from your critical loads times the hours you need them; self-consumption sizing from your evening usage after solar stops; arbitrage sizing from how much you can shift out of peak pricing. The three give different answers.
A critical-loads panel is a subpanel holding only the circuits a battery or generator will keep running during an outage — typically the fridge, furnace blower, well or sump pump, some lights and internet. Choosing what goes in it is the actual design work, because it sets the battery and inverter size.
An off-grid system is sized for the worst month, not the annual average — which is why arrays are heavily oversized against summer need. Expect 2–3 days of battery autonomy, an inverter that can start your largest motor, and a backup generator as part of the design rather than a fallback.
A grid-tied system uses the grid as its battery and shuts down in an outage; an off-grid system must supply every kWh itself, sized for the worst month with days of storage and a generator. They are different designs, not the same system with or without a wire.
A hybrid inverter handles solar and battery in one unit, converting DC once rather than twice, and it manages islanding during an outage. Specify one up front even if the battery comes later — retrofitting storage to a plain string inverter usually means AC coupling and a second inverter.
The design question that decides everything is whether the solar can recharge the battery while islanded. A system that cannot only has whatever the battery held when the grid failed; one that can keeps cycling through a multi-day outage — and that capability depends on the inverter, not the panels.
Grid-tied solar shuts down in an outage by design, under IEEE 1547 anti-islanding requirements — otherwise the array would keep energising the utility line and endanger the crew repairing it. Adding a battery with proper islanding capability is what allows the system to keep running.
Solar panels last 25–30 years or more, degrading roughly 0.4–0.5% per year rather than failing outright. Most manufacturers warrant around 80–90% of rated output at year 25. What fails first is usually the inverter, at 10–15 years, not the panels.
Solar panels usually need cleaning once or twice a year, and only when soiling is visible — rain handles most dust. Use water and a soft brush or squeegee from the ground, early morning while the glass is cool. Never use pressure washers, abrasives, detergents or hot water on cold glass; all four can void the warranty.
Start with the disconnects — an inverter is fed by both a DC and an AC disconnect, and either being open looks like a dead unit. Record the error code before resetting anything, and never open the enclosure: DC conductors from the array stay live whenever there is daylight.
Solar panels are tested to UL 61730 / IEC 61730, typically a 1-inch ice ball at around 50 mph, and usually survive hail that damages a roof. The greater risk in high wind is the attachment to the roof rather than the glass — and microcracks from impact are invisible but show up as lost output.
Monitoring reports production at system, string or per-panel level depending on your inverter architecture. It matters because a solar system can underperform silently for years — a failed optimiser, a dirty section or a shaded string produces less without ever throwing an error.
If a roof has less than about 10 years left, replace it before the panels go on. Removing and reinstalling an array to re-roof later commonly costs thousands — far more than doing the roof first, and it can complicate both warranties.
Before assuming a fault, compare against the same month last year rather than against the summer peak — season, weather and an optimistic sales estimate explain much of it. Real causes are soiling, new shading, a failed optimiser or microinverter, a down string, or an inverter fault.
