Best Direction for Solar Panels: N, S, E or W?

The best direction for a solar array depends on the decision being optimized: annual energy, useful production at particular times, total roof output or financial value. An equator-facing fixed array is a useful starting candidate in many locations, but shade, local weather, roof geometry, tilt and tariffs can change the preferred design.

Last updated: November 2026.

This guide explains how to compare orientations. It does not provide a universal production penalty or claim that the same roof layout works in every hemisphere. Use the azimuth and tilt guide for terminology and the shading guide for obstruction analysis.

Define direction and tilt before comparing results

Azimuth describes the direction a tilted array faces; tilt describes its angle above horizontal. In the PVWatts convention, cardinal directions are:

The SAM PVWatts system-design documentation uses that convention regardless of hemisphere. Some tools use a different convention, so retain the software’s definition. At exactly horizontal tilt, changing azimuth alone does not change the physical orientation of the module surface.

Use geographic, or true, direction when the model requires it. A magnetic compass needs the appropriate local correction; NOAA’s geomagnetic calculator provides location- and date-dependent declination. A phone reading can also be affected by calibration or nearby metal. A preliminary measurement is not a substitute for the project’s verified geometry.

Is south always best in the Northern Hemisphere?

For many unshaded fixed-array sites outside the tropics, facing toward the equator, south in the Northern Hemisphere, north in the Southern Hemisphere, is a sensible candidate for annual production. It is a starting hypothesis, not a guarantee of the highest modeled output. Near the tropics, seasonal sun paths and local conditions make a simple hemisphere rule especially incomplete.

The EIA orientation and tilt explanation shows that direction can change both annual output and time-of-day production. Its historical modeled examples are location-specific; they do not prove universal percentages for a different roof.

Keep the objective explicit:

A design can win on total roof energy while producing less energy per module. A design can also produce more annual energy while offering less financial value under a particular tariff. Those are different comparisons.

What Modeled Data Shows for Three Sites at One Tilt

To show how much results vary, we ran the European Commission’s PVGIS tool for 1 kWp of fixed panels at 30° tilt in three cities. Each run used the ERA5 weather database and a 14% system loss assumption, and only the azimuth changed. In the Southern Hemisphere city, the equator-facing direction is north.

Source: PVGIS 5.3 runs on 4 November 2026. Percentages are relative to the equator-facing result for the same city. This is modeled data for three sites at one tilt, not a measured roof result or a rule for other locations.

Three things stand out. First, the equator-facing roof produced the most in all three cities. Second, east and west landed close together and well above the pole-facing roof, so the choice between them is usually decided by load timing and shade, not annual energy. Third, a pole-facing surface at this 30° tilt lost far more than an east or west one. A shallower pitch changes that, so check the real pitch before you rule a face out.

Southeast, southwest, east and west roofs

An east-facing surface tends to weight exposure toward morning; a west-facing surface tends to weight it toward afternoon. Exact production timing depends on latitude, season, weather, shade and daylight. Do not assign a universal clock window or promise that west captures a utility’s evening peak after sunset.

Southeast and southwest candidates may provide a useful compromise between annual yield and timing, but there is no defensible universal “95–98%” rule for every roof. Nor is west inherently more productive than east. Morning cloud, afternoon storms, temperature and obstructions can change the comparison.

For an existing pitched roof, model the actual faces before proposing special racking. A new tilt or orientation can change wind loads, mounting requirements, cost, appearance and access. Responsible design review must assess those consequences alongside modeled energy.

Can a north-facing roof work?

“North-facing” is not a universal disadvantage: it is generally equator-facing in the Southern Hemisphere. For a Northern Hemisphere site, compare a pole-facing roof with feasible alternatives using its actual pitch, shade, weather and equipment. A shallow surface differs substantially from a steep surface, but neither has one fixed penalty.

Use a comparison record rather than a categorical cutoff:

• Exact roof-face azimuth and tilt, with a measurement source.
• Verified usable area and obstruction geometry.
• Feasible module count and mounting arrangement.
• Same-capacity yield comparison with an appropriate reference.
• Total-roof energy, costs and relevant seasonal output.
• Electrical architecture and current equipment constraints.

Premium modules, optimizers, microinverters or bifacial modules do not automatically cure unfavorable geometry. Higher module efficiency can change capacity within a limited area; bifacial gain depends on rear irradiance and mounting conditions. Evaluate specific equipment and installation evidence rather than promising a percentage improvement.

East-west layouts on flat roofs

An east-west arrangement can change module density, row shading and the production profile. Whether it produces more total roof energy depends on the actual feasible layout, not a universal claim that it fits 25–30% more modules.

Compare two questions separately:

1. With equal DC capacity and comparable equipment, what is the energy yield of each orientation?
2. Within the actual roof constraints, how much capacity can each approved layout accommodate and what total energy does it produce?

Keep manufacturer mounting requirements, roof loads, drainage, fire access, maintenance paths, setbacks and local approvals in the layout review. Model row shading with the appropriate method. The SAM documentation distinguishes fixed roof-mount assumptions from fixed open-rack self-shading treatment; a generic roof option does not prove that dense rows have no shading losses.

Match direction to the actual tariff and load

Time-of-use pricing may make later production valuable, but compare the overlap with the actual charge periods and customer load. Do not treat an avoided retail import and an export credit as interchangeable. A flat export rate and a time-varying import rate can favor a different layout than uniform compensation.

For commercial demand charges, recompute the defined billing peak after adding each PV profile. West-facing PV does not guarantee a lower demand charge if the relevant peak occurs at night, on a cloudy day or in another billing window. See behind-the-meter optimization for that meter-and-tariff comparison.

A reproducible orientation test

For a quick check of how an array behaves under shade, see shading analysis and the guide to optimal tilt angle. Use PVWatts or an appropriate project model for preliminary alternatives. Record the version, location, weather source, module and inverter assumptions, array type, losses, DC/AC ratio, shade method and output boundary. The result is a model estimate, not measured roof performance.

1. Enter the actual roof face and feasible capacity as the baseline.
2. Hold other inputs constant for an equal-capacity orientation test.
3. Export annual and monthly results; obtain finer time resolution when the decision depends on timing.
4. Compare feasible complete layouts separately if module count, mounting or equipment changes.
5. Match output to load and tariff only when those inputs support the analysis.
6. Retain unresolved geometry and equipment questions for responsible review.

Hypothetical calculation check: If two equal-capacity alternatives produced modeled yields of 1,000 and 900 kWh/kWp-year, the second would have 900 ÷ 1,000 = 90% relative yield, or a 10% reduction against the first. For assumed capacities of 6 and 7 kWp respectively, total outputs would be 6,000 and 6,300 kWh/year. The second wins on total energy but not specific yield. These figures are arithmetic examples, not PVWatts runs, orientation benchmarks or evidence for a larger installation.

Tilt belongs in the comparison too. Latitude is a rough starting point sometimes used for fixed arrays; the SAM documentation recommends testing the actual project objective. Do not transfer seasonal tilt adjustments or a minimum mounting angle into universal code or product rules.

Mixing orientations requires an electrical design

Different roof faces can share a project without sharing every electrical constraint. Separate MPPT inputs, string arrangements and module-level electronics offer different architectures. Compatibility depends on the exact equipment, string lengths, current, voltage, operating conditions and manufacturer instructions. Using a string inverter does not itself mean that multiple orientations are prohibited.

Do not assume an optimizer or microinverter guarantees independent behavior or removes all mismatch. Ask the responsible designer to verify the proposed topology against current documentation; the auto-stringing review checklist helps organize the evidence.

For a quote, request the actual roof-face inputs, equal-capacity results, complete-layout results, shade assumptions, tariff method and review status. Ask for a demo of the solar design workflow on one of your own roofs, and book it here. This article does not verify satellite measurement accuracy or automatic financial comparison for every orientation.

Frequently Asked Questions

What is the best direction for solar panels?

Equator-facing fixed arrays are useful annual-energy candidates in many locations, especially outside the tropics. The best project choice depends on local weather, shade, tilt, feasible layout, load timing and tariff. Compare the actual alternatives rather than applying one universal orientation percentage.

How much output do east- or west-facing panels lose?

There is no universal penalty. Run an equal-capacity comparison with the same weather, equipment, losses, tilt and output boundary, changing the relevant orientation. Then compare feasible complete layouts separately if capacity, mounting or other inputs differ. East does not inherently yield less than west.

Should a model use true or magnetic direction?

Use the direction convention required by the model, normally geographic azimuth for PVWatts. Its cardinal values are north 0°, east 90°, south 180° and west 270°. Correct a magnetic measurement using the current local declination and check measurement quality; do not assume a phone reading verifies roof geometry.

Can a north-facing roof support solar?

Evaluate its actual geometry and location. North faces toward the equator in the Southern Hemisphere. For Northern Hemisphere sites, a pole-facing roof can still be an option, depending on pitch, weather, shade, costs and constraints. No universal roof-pitch cutoff or fixed production percentage establishes suitability.

Can different orientations use the same inverter?

The project may include multiple orientations, but the proposed stringing and MPPT architecture must follow the exact equipment’s voltage, current, module and manufacturer conditions. Separate trackers or module-level electronics offer different options. A string inverter does not universally prohibit multiple orientations, and additional electronics do not remove every compatibility check.