Solar marketing photos always show the same house: one big, clean, south-facing rectangle of roof. Almost nobody actually owns that house. Real roofs have dormers, chimneys, plumbing vents, skylights, valleys, and hips, and every one of those features eats space, casts shade, or both.
Our own array is a good example — 40 panels feeding two Tesla inverters, spread across multiple roof planes because no single plane could hold everything. Getting there took real layout work, and most of it happened before anyone stood on the roof. Here’s the process I’d recommend to any homeowner facing a complicated roof, using tools you can run yourself for free or nearly free.
Why complex roofs punish lazy layouts
Two things make an obstructed roof harder than the simple case:
Lost area. Every chimney and skylight removes panel positions, and the required clearances around them remove more. On a busy roof, the usable area can be dramatically smaller than the total area.
Shade. A chimney doesn’t just occupy its own footprint — it drags a shadow across neighboring panels, and that shadow moves through the day and stretches much longer in winter when the sun sits low. A panel that’s clear at noon in June may be shaded half the morning in December.
Shade matters more than intuition suggests because panels are wired in series strings. Depending on your inverter architecture, hard shade on part of one panel can pull down more than just that panel. This is exactly why complex roofs so often end up with microinverters or DC optimizers — they contain shading losses to the affected module instead of letting them spread across a string.
Step one: inventory every obstruction and its required clearance
Before modeling anything, walk your roof plan (from the ground, from satellite view, or from attic drawings) and list every feature. Each one carries a clearance requirement — some from fire code, some from practical shading and maintenance logic. The table below covers the usual suspects. Important: setback rules come from your local adopted fire code and vary by jurisdiction; treat these as planning considerations to verify, not final numbers.
| Roof feature | Why it constrains layout | Typical planning consideration |
|---|---|---|
| Ridge | Fire code commonly requires a clear pathway/setback near the ridge for firefighter access and venting | Often 18–36 in. depending on jurisdiction and how much of the roof is covered — verify locally |
| Roof edges / eaves | Access pathways along edges are commonly required | Frequently around 36 in. on some edges; your AHJ (local building authority) has the final word |
| Chimney | Occupies area and casts a moving shadow, longest in winter | Keep panels out of the winter shadow path, not just outside the brick |
| Dormer | Blocks area on its own faces and shades the main plane beside it, especially east/west sides morning and evening | Model its shadow across the day, not just at noon |
| Skylight | Can’t be covered; usually needs working clearance | Leave room to service and flash it; check code for required clearance |
| Plumbing vents / flues | Small footprint, but panels can’t overlap them | Vents can sometimes be rerouted a few feet by a plumber — often cheaper than losing a panel position |
| Valleys | Water concentration zone; mounting there is a leak risk | Keep panels off valleys and respect drainage paths |
| Hips | Break planes into triangles that fit rectangles poorly | Expect orphaned triangular areas you simply can’t use |
That last column hides a genuinely useful trick: vent relocation. On a tight roof, paying a plumber to move a vent pipe two feet can open up a full panel position — usually a better deal than giving up the production.
Step two: build a shade model in SketchUp
SketchUp (the free web version is enough for this) has a built-in solar shadow study: geolocate your model, set a date and time, and it renders shadows accurately for your latitude. That makes it ideal for the question that matters most on a complex roof — where do the chimney and dormer shadows actually fall in December?
You don’t need an architectural masterpiece. Model the main roof planes at the correct pitch, then block in each obstruction at roughly the right height and position. Accuracy on heights matters more than looks — a chimney modeled a foot too short will lie to you about its winter shadow.
Then run the shadow study at a few key times: mid-morning, noon, and mid-afternoon, on the winter solstice, the equinox, and the summer solstice. Areas that stay clear across all of those are your premium panel real estate. Areas shaded only on winter mornings are usable but discounted. Areas shaded through the middle of winter days should be last-resort positions, and only with module-level electronics.
Step three: compare roof planes with PVWatts
PVWatts is NREL’s free production calculator, and it’s the right tool for the next decision: which planes deserve panels, and in what order? Run a separate simulation for each candidate plane — same system size each time, with that plane’s actual tilt (your roof pitch) and azimuth (compass direction) — and compare annual output.
Labeled example (illustrative — run your own address): on a mid-latitude U.S. roof, a simulation might show a south-facing plane producing the most annually, a west plane coming in perhaps 10–20% lower, and an east plane similar to the west. Numbers vary meaningfully with location and pitch, which is exactly why you run it per-plane instead of guessing.
Two notes on interpreting the results:
- PVWatts doesn’t know about your chimney. It models an unshaded plane. You apply the shade knowledge from your SketchUp study on top of it — a plane that PVWatts loves but that sits in dormer shadow all winter is worse than the raw number suggests.
- West-facing planes are worth more than their raw kWh if you’re on time-of-use rates, because they produce into the expensive late-afternoon window.
Google Project Sunroof is a decent gut-check at the start of all this — it estimates roof solar potential from aerial imagery — but it’s a screening tool, not a design tool.
Step four: match the layout to the electrical design
Once you know which planes get panels, the wiring plan should respect the shade map. The principles are simple even if the installer handles the details:
- Don’t mix planes with different orientations on the same series string with a plain string inverter.
- Put known-shade-compromised positions on their own optimizer or microinverter channel so their bad hours don’t tax clean panels.
- If the design spans several planes — ours does — confirm the inverter(s) have enough independent inputs (MPPTs) to treat each differently-oriented group separately.
Step five: sanity-check before you sign
When installer proposals arrive, compare them against your own work. A good proposal will include a shade analysis (often from an on-roof measurement tool or drone survey) — check that it agrees with your SketchUp shadow study. If a layout drawing shows panels inside a chimney’s winter shadow path or crowding a setback you know your jurisdiction enforces, ask about it. Installers are usually happy to explain; the ones who can’t are telling you something.
A complex roof will never match the marketing-photo house, and that’s fine. Ours doesn’t either. The goal isn’t a perfect roof — it’s making sure every panel you pay for goes in a position that earns its keep, and the free tools above get you most of the way there before the first quote lands.