Every failed off-grid build we’ve audited announces itself the same way: a generator coughing awake at three in the morning, in November, again. The owners bought good batteries. They bought a serious inverter. What they didn’t buy was enough December.
Solar is the spine of an off-grid system — not because it’s fashionable, but because nothing else delivers energy this cheaply, this quietly, with no moving parts, no fuel runs and no oil changes. Every other subsystem on this site — storage, heating, water pumping, the whole stack — is shaped by what the array does and when it does it. Get generation right and the rest of the design is arithmetic. Get it wrong and no battery bank on earth can save you; a battery stores energy, it doesn’t invent it.
This page is the map of the whole generation chain, at the altitude a serious buyer needs: what a real site and load assessment looks like, how arrays are actually sized, what the charge controller does all day, how to buy panels in 2026, and how generation gets matched to the way a household really lives. Each stop links to a deep guide when you’re ready to go further.
On this page
Why Solar Carries the Whole System
Here’s the inversion most buyers haven’t caught up with. In 2010, panels were the system: modules cost over $2 a watt and swallowed half the budget, so “going solar” meant agonising over panels. In 2026, quality mono TOPCon modules trade like a commodity — routinely under $0.35 a watt retail in the US and under €0.20 in much of Europe. On a whole-home off-grid build, the panels are often the cheapest major component — frequently less than the labor to mount them.
Where does a $30k–$60k (or €25k–€55k) budget actually go? Roughly like this:
| Where the money goes | Share of budget | What it actually buys |
|---|---|---|
| Solar array (panels, racking, array wiring) | 15–20% | The energy source itself — now the commodity end of the build |
| Battery storage | 30–35% | Every kilowatt-hour you use after dark |
| Power electronics (MPPT + inverter/charger) | 12–18% | Voltage, charge logic, clean AC — the brains |
| Balance of system (copper, protection, disconnects, grounding) | ~10% | The part inspectors and insurers care about |
| Design, labor, commissioning | 20–25% | The difference between components and a system |
Follow the money and your attention should invert too. The brochures still sell panels; the outcomes are decided by design, storage and power electronics. Panels are a commodity. Systems are not. That is the single most useful thing to know before you talk to anyone holding a quote.
Start With the Load, Not the Roof
You don’t size an off-grid system from the roof down. You size it from the fridge up. The first deliverable of any competent designer isn’t a panel count — it’s a load profile: what you draw, when you draw it, and how that changes between June and January.
The spread is enormous. A frugal cabin runs on 3–6 kWh a day. An efficient family home lands at 10–20 kWh. An all-electric house with a heat pump, well pump and workshop can pull 25–40 kWh a day in winter — a completely different machine. Two households in identical houses can differ by a factor of three on habits alone, which is why generic “systems by bedroom count” fail so reliably.
Surge matters as much as consumption. A 1 hp well pump sips roughly a kilowatt running — and demands 3–5 kW for the half-second it takes to start. That number sizes your inverter, not your array, and it’s the spec most cheap quotes quietly skip.
This is also where geography sneaks into the electrical design. In North America, homes run 120/240 V split-phase — well pumps and dryers sit on 240 V, so the inverter (or a stacked pair) must serve both legs. In Europe, most homes are 230 V single-phase, but Germany, Austria and the Netherlands commonly wire 400 V three-phase for heat pumps and workshops, which pushes the design toward a three-inverter cluster. The load list picks the inverter topology — never the other way round.
A proper site and load assessment nails down:
- Every load over ~100 W, with its duty cycle — measured where possible, not guessed
- The single largest surge (usually a pump or compressor) and what else runs at the same moment
- Daily kWh in the worst month you’ll occupy, not the annual average
- Which loads can move into the middle of the day — free capacity, as we’ll see below
- Shade, roof pitch and azimuth, or the ground-mount option that ignores all three
- The honest conversation: what stays on during a dark week, and what doesn’t
Sizing the Array: You’re Buying December, Not July
Grid-tied solar gets to think in annual averages, because the grid absorbs summer surplus and covers winter deficit. Off-grid gets no such luxury. You live month by month, and the month that sizes your array is the darkest one you’ll occupy. July is a liar; December writes the spec.
The working currency is peak sun hours (PSH): a site with 3 PSH delivers the equivalent of three hours of full-rated output per day, so a 1 kW array yields about 3 kWh there before losses. And losses are real — hot panels, wiring, charge control, battery round-trip, inverter conversion — so a well-built system delivers roughly 70% of the textbook number to your actual loads. The whole sizing method fits in one line:
array kW = daily kWh ÷ (December PSH × 0.70)
Worked example: a 10 kWh/day home near Denver sees about 4.4 PSH even in December (high, cold, sunny) — 10 ÷ (4.4 × 0.7) ≈ 3.2 kW, call it eight 430 W panels. The identical home near Munich, at 1.4 December PSH, needs ~10 kW — twenty-four panels — to live the same life. Same house, same fridge, triple the array:
| Location | Typical December sun (PSH/day) | Array for a 10 kWh/day home |
|---|---|---|
| Phoenix, Arizona | ~5.5 | ≈ 2.6 kW |
| Denver, Colorado | ~4.4 | ≈ 3.2 kW |
| Seville, Spain | ~3.4 | ≈ 4.2 kW |
| Seattle, Washington | ~1.2 | ≈ 11.9 kW |
| Munich, Germany | ~1.4 | ≈ 10.2 kW |
| Amsterdam, Netherlands | ~0.9 | ≈ 15.9 kW |
| Stockholm, Sweden | ~0.5 | ≈ 28.6 kW — which is why nobody does this with panels alone |
Tilted arrays, 70% system efficiency. The Stockholm row isn’t a dare: past roughly 55°N, sane designs cover the darkest six weeks with wind or a generator instead of chasing panels.
Key number
≈4× — the array difference between Seville and Amsterdam for the same December load. Latitude is the biggest single line item in your quote, and it’s the one thing no installer can discount.
Your December number is the quote
Once you know the month that writes your spec, the next step is a price against it. Send your loads and site — the spec comes back written, and vetted pros quote against that, not against a blank check.
Tilt, orientation and snow
Because December writes the spec, off-grid arrays lean steeper than grid-tied ones: latitude plus 10–15° biases harvest toward the low winter sun and sheds snow instead of storing it. In the far north, near-vertical wall mounts trade summer yield you didn’t need for winter yield you desperately do — and bifacial panels over snow claw back double-digit gains from reflected light. An east–west split stretches generation into morning and evening where a battery is small. None of this is exotic; it’s just designing for the month that matters.
MPPT Charge Control: the Quiet Workhorse
Between the array and the battery bank sits the least glamorous, hardest-working box in the system: the MPPT charge controller. Its job is two-fold — hold the array at its maximum power point (the constantly moving voltage where the panels give their best), and convert high-voltage strings, typically 150–450 V, down to what a 48 V bank wants, while running the battery’s charge stages. Against the older PWM approach it harvests 20–30% more in exactly the conditions you care about: cold, low-sun winter days.
Two design points separate professionals from parts-swappers. First, string voltage windows: panel voltage rises as temperature falls, roughly 10% higher on a −15°C clear dawn than on the datasheet. Strings must be sized so the coldest morning of the decade stays under the controller’s maximum — get it wrong and the magic smoke escapes on the prettiest day of the year.
Second, deliberate oversizing. Good controllers current-limit gracefully, so hanging 1.3–1.5× the controller’s rating in DC panels is not a mistake — it’s the cheapest winter energy you can buy. You give up a flat-topped noon in May, when the batteries were full by ten anyway, and gain the December harvest that keeps the generator asleep. At commodity panel prices, clipping is a feature.
One architectural note: off-grid defaults to DC-coupling (array → MPPT → battery) because charging keeps working regardless of what the AC side is doing. AC-coupled arrays have their place in hybrid systems — the architecture guide draws that line properly.
Choosing Panels in 2026
The panel aisle in 2026 is refreshingly boring. N-type TOPCon is the mainstream default — 430–500 W residential formats at 21–23% efficiency, with better low-light response and a gentler temperature coefficient (around −0.30%/°C) than the PERC generation it replaced. Heterojunction (HJT) sits above it as the premium option; leftover PERC is clearance stock. Brand-to-brand differences within a tier are smaller than the marketing implies.
What actually deserves your attention: the temperature coefficient (a dark roof drives cells to 60–65°C and shaves 10–12% off rated output — the gentler the slope, the more of your paper watts survive summer); the degradation warranty (87–89% output at year 25–30 is now table stakes — hold suppliers to it); mechanical ratings for hail, wind and snow load if your site is exposed; and buying 10–15% spares from the same batch, because matching a panel three years later is a headache electrically and cosmetically.
What doesn’t deserve your attention: paying a premium for the last half-percent of efficiency when you’re not space-constrained. If you have ground, a bigger array of ordinary panels beats a smaller array of exotic ones on every metric that matters — including the only one that pays the bills, kWh per dollar in December.
One note on incentives: in the US, the 30% federal residential credit (Section 25D) expired for systems installed after December 31, 2025, so 2026 buyers price systems on raw economics — while several European countries (Germany and the Netherlands among them) still apply 0% VAT to residential solar. Check what’s live locally before budgeting off anyone else’s number.
Matching Generation to the Way You Actually Live
An array doesn’t power a house; it powers a schedule. Solar production is a midday plateau. Domestic life peaks at 7 a.m. and 7 p.m. Everything in off-grid design is about closing that gap, and there are exactly three tools for it.
Storage bridges hours. The battery bank carries the evening peak and the night — its size, chemistry and depth-of-discharge rules are a discipline of their own, covered on the storage page. Load-shifting bridges habits. Every kilowatt-hour you move into the solar window is a kilowatt-hour of battery you don’t buy: pump the well into a cistern at noon, run laundry at one, heat water as thermal storage in the afternoon, pre-heat or pre-cool the house with the mini-split while the sun pays for it (see heating & cooling). On most builds, disciplined load-shifting is worth $3k–$5k of storage.
And the generator bridges weeks. Here’s the reframe that saves buyers real money: a generator in a well-designed system isn’t an admission of failure — it’s a designed component with a fuel budget and a duty cycle, typically 40–80 hours a year, covering the one dark stretch that would otherwise demand five figures of extra array and battery. Two to three days of battery autonomy plus a small generator beats five days of battery on cost almost everywhere humans actually live. The 3 a.m. November generator from our opening wasn’t a scandal because it ran — it was a scandal because nobody had planned for it to run, so it ran unserviced, unfueled and at the worst possible hour.
Run your own numbers
Ten minutes with your load list gets you a defensible array, controller and bank size — and a completed sizing is the first thing any good installer will ask for.
Get matched
Have this designed around your site, not a brochure
Tell us roughly where you are and what you want to run. We’ll come back with next steps and match you with vetted off-grid installers who quote against a real load profile.
Two ways forward
Get the gear shortlist — and a place on the installer waitlist
Tell us roughly where you are and what you want to run. We’ll come back with the exact CE-marked, 230 V array, inverter and battery that fit your load, and where to buy them (affiliate links that keep this guide free). We’re vetting European installers country by country and aren’t there yet, so drop your country below and we’ll notify you the moment a vetted pro can quote your build.
Go Deeper on Off-Grid Solar
This page is the map. These are the territories — each one a full working guide:
Payback Period, Honestly
When off-grid beats a grid connection — and when it doesn’t.
Off-Grid vs Grid-Tied vs Hybrid
Which architecture your situation actually calls for.
How many solar panels does an off-grid house actually need?
For a 10–15 kWh/day home, anywhere from eight panels in the sunny latitudes to thirty-plus in northern Europe — December sun, not house size, is the driver. Run the sizing formula above with your own December PSH, or use the calculator for a defensible number in minutes.
Do panels still work in winter and under cloud?
Yes — cold actually improves panel voltage and efficiency. The problem is hours, not temperature: heavy overcast delivers 10–25% of clear-sky output, and winter days are short. That’s exactly why off-grid arrays are sized on December production and tilted steeply, rather than sized on annual averages like grid-tied systems.
How long does the equipment last?
Panels are the long-lived part: 25–30-year output warranties, degrading around 0.3–0.5% a year. Power electronics realistically last 10–15 years and batteries 10–15 depending on chemistry and discipline — a well-planned budget treats those as scheduled replacements, not surprises.
Can I start small and expand the array later?
Yes, if it’s designed in from day one: controller and wiring headroom, bus capacity, and space on the mount. The catch is panel matching — modules from a different batch or generation don’t string cleanly with the old ones, which is why designers order spares up front or plan expansion as a separate string on its own MPPT.