Solar Panel Direction Calculator: Which Way to Face (2026)
Published July 16, 2026 · Updated July 22, 2026
Facing your panels east or west isn't the disaster people fear — in New York, at a typical roof tilt, an east-facing roof still delivers 83% of a south-facing one, west 81%. What most people get wrong isn't the direction — it's that their phone compass points at magnetic south, not true south, which in Seattle is off by nearly 15°. Enter your ZIP for your exact figure, and the compass heading that actually points true.
Solar Panel Direction Calculator
Real NREL PVWatts output for your roof's facing — plus the compass heading that points at true south.
Assumptions: output figures come from NREL PVWatts v8 (NSRDB) for a fixed roof mount, 14 % system losses (~0.83 derate), and a 1 kW reference array; azimuth curves are sampled at 10°, 25°, and 40° tilt and interpolated to your exact tilt. Declination is computed locally from the NOAA/NCEI World Magnetic Model (WMM2025, epoch 2025.0). ZIP-to-coordinates via Zippopotam.us (public-domain U.S. Census data).
What Direction Should Solar Panels Face?
In the Northern Hemisphere the answer is true south (180° azimuth) — that maximizes annual production. But "south" on your phone compass is magnetic south, which can be off true south by up to about 15° depending on where you live. The calculator above resolves both: your yield at the direction you actually face, and the compass heading that points at true south.
The key idea is that the penalty for facing off-south is a curve, not a cliff. Southeast and southwest are nearly as good as south; east and west give up a slice; only as you swing toward north does the loss become large — and even then it depends heavily on your tilt.
The 6 kW array on my own roof faces south-southwest — about 15–20° off true south, at 46° latitude. Run through PVWatts at my exact tilt, that facing costs me around 1% a year versus dead south: lost in the noise. The part that could actually have tripped me up was true vs. magnetic south — aim by compass without correcting for declination and you stack the full local offset on top of whatever your roof already gives up. The calculator's "aim at X°" heading erases that, for free.
East-Facing And West-Facing Roofs
This is where expectations are usually wrong. An east- or west-facing array at a normal tilt keeps about 80–83% of true-south output — a loss of roughly 17–20%, not 50% (as little as 8–9% on a low-slope roof). The calculator gives your exact figure. West can even beat east if your utility uses time-of-use rates that pay more for late-afternoon energy — and on the Pacific coast, where mornings are cloudier than afternoons, west quietly out-produces east even in raw kWh.
The full picture across roof directions and tilts — the same numbers behind the calculator — is:
| Roof direction | Low tilt (10°) | Mid tilt (25°) | High tilt (40°) |
|---|---|---|---|
| South toward the equator | 100% | 100% | 100% |
| SE / SW ±45° off south | 98% | 95% | 94% |
| East / West ±90° off south | 92% | 83% | 78% |
| NE / NW ±135° off south | 86% | 69% | 57% |
| North 180° off south | 83% | 62% | 47% |
These are contiguous-US averages; the calculator above returns the exact figure for your location, which varies with latitude and climate — so a slightly different number there is expected, not a bug.
How Much Do East, West, And North Cost?
Direction matters more than tilt, but "matters" still rarely means a dealbreaker. East and west trade a modest slice of output; north is the only facing that gets genuinely expensive — and even north eases sharply as the roof flattens.
A worked example: a 10 kW system in New York
Here is what a fixed 10 kW array in New York — on a typical 25° (≈6/12) roof, 1,305 kWh/kW·yr south-facing from PVWatts — produces at each facing, and what the shortfall costs at an illustrative $0.16/kWh (about the U.S. residential average, EIA):
| Roof faces | Annual output | Loss vs south | Per day | Cost/year* |
|---|---|---|---|---|
| South (180°) | 13,050 kWh | — | — | — |
| East (90°) | 10,845 kWh | 2,205 kWh (17 %) | 6.0 kWh | ~$350 |
| West (270°) | 10,610 kWh | 2,440 kWh (19 %) | 6.7 kWh | ~$390 |
| Split E/W | 10,727 kWh | 2,323 kWh (18 %) | 6.4 kWh | ~$370 |
| North (0°) | 7,921 kWh | 5,129 kWh (39 %) | 14.1 kWh | ~$820 |
* Loss vs true south, priced at an illustrative $0.16/kWh. Output = New York's PVWatts v8 south-facing yield at 25° scaled by each direction's azimuth factor.
The shape is the point: east or west costs about $350–390 a year — a sixth of output, not the half people picture. Split east/west lands in between and buys a flatter daily curve. Only north, at nearly 40% and ~$820, is in genuinely costly territory — and on a low-slope roof even that climbs back into the 80s. Direction sets the ceiling; a south-facing roof simply starts higher.
Solar Panel Direction & True-South Aim By U.S. City
Real NREL PVWatts v8 orientation losses and the NOAA WMM-2025 true-south compass heading for a spread of U.S. cities. The aim column is the heading to point a phone compass at so the panel faces true south — the correction nobody else gives you.
| City | Latitude | Declination → true-south aim | East / West | North |
|---|---|---|---|---|
| Honolulu, HI | 21° | 9° E → 171° | 91 % / 91 % | 79 % |
| Miami, FL | 26° | 7° W → 187° | 89 % / 86 % | 71 % |
| Houston, TX | 30° | 2° E → 178° | 87 % / 87 % | 70 % |
| El Paso, TX | 32° | 7° E → 173° | 85 % / 83 % | 63 % |
| Atlanta, GA | 34° | 6° W → 186° | 85 % / 84 % | 65 % |
| Los Angeles, CA | 34° | 11° E → 169° | 81 % / 86 % | 63 % |
| Albuquerque, NM | 35° | 8° E → 172° | 84 % / 81 % | 60 % |
| Nashville, TN | 36° | 4° W → 184° | 85 % / 83 % | 64 % |
| San Francisco, CA | 38° | 13° E → 167° | 81 % / 86 % | 62 % |
| Denver, CO | 40° | 8° E → 173° | 83 % / 79 % | 57 % |
| New York, NY | 41° | 12.5° W → 193° | 83 % / 81 % | 61 % |
| Chicago, IL | 42° | 4° W → 184° | 82 % / 83 % | 61 % |
| Boston, MA | 42° | 14° W → 194° | 82 % / 81 % | 59 % |
| Minneapolis, MN | 45° | 0° → 180° | 81 % / 80 % | 57 % |
| Bismarck, ND | 47° | 5° E → 175° | 79 % / 80 % | 55 % |
| Seattle, WA | 48° | 15° E → 165° | 81 % / 84 % | 62 % |
| Anchorage, AK | 61° | 14° E → 166° | 76 % / 82 % | 56 % |
| Fairbanks, AK | 65° | 15° E → 165° | 75 % / 79 % | 52 % |
Orientation losses: NREL PVWatts v8 (NSRDB) at a 25° tilt, as % of true-south output. Declination and true-south aim: NOAA/NCEI World Magnetic Model (WMM-2025, epoch 2025.0), computed at each city.
Three things worth noticing:
- The correction flips coast to coast. In New York you aim your compass at 193° (12.5° west of the compass's "south") to face true south; in Seattle you aim at 165° (15° east). Same true south, opposite corrections — which is exactly why a single rule of thumb can't work.
- Minneapolis needs no correction at all. It sits almost exactly on the agonic line (declination ≈ 0°), where magnetic south is true south. The correction is real, but it isn't universal.
- East/west barely moves; north is all about latitude. East and west sit in the low 80s almost everywhere — a ~15–20% haircut. North swings far more, from ~79% in sunny Honolulu down to ~52% in Fairbanks, because the farther north you go, the lower the sun rides and the more a north face is left in shadow. On the Pacific coast (LA, SF, Seattle, Anchorage) west even beats east, as clearer afternoons than mornings tip the balance westward.
North-Facing Panels
North is the worst case in the Northern Hemisphere, but how bad depends almost entirely on tilt and latitude: a north-facing array keeps only about 46% of true south on a steep (40°) roof at mid-latitudes — as little as ~37% in the far north — but roughly 82% on a low-slope (10°) one. So it is not automatically a dealbreaker — in high-rate areas, or when it's the only viable roof plane, a shallow north-facing array can still pay back. Read your exact north-facing percentage from the calculator and weigh it against your local rates and peak sun hours.
Split East/West Arrays
Splitting panels across east- and west-facing roof planes spreads production across the morning and afternoon. Total annual energy lands between a pure-east and pure-west array (the calculator reports the split-E/W figure), and the flatter daily curve can pair well with batteries or time-of-use billing.
Magnetic South vs True South — The Phone-Compass Trap
Most people aim panels with a phone, which reads magnetic north/south. The gap between magnetic and true is the magnetic declination. Aim naively at magnetic south and your array is misaligned by the full declination — but the honest news is that the annual cost of that error is usually small (the calculator computes the exact percentage for your ZIP). Use the computed "aim your compass at X°" heading and you remove the error entirely for free.
How Tilt Changes The Azimuth Penalty
Azimuth and tilt interact: the flatter your array, the less direction matters. A near-flat array is close to omnidirectional, while a steep array is much more sensitive to facing the wrong way. Drag the tilt slider in the calculator to watch the whole curve flatten, and use the tilt angle calculator to find your best angle. For whole-system sizing, the solar panel calculator ties it together.