Solar panels can work on a four-bedroom house with a north-west and south-east split roof. For a 5.46 kWp reference array in central England, I calculate about 4,253 kWh a year. That is roughly 83% of the output from the same panel area facing south, before allowing for site-specific shade.
The important part is not the bedroom count. It is how much panel capacity fits on each face, when your home uses electricity, and whether the north-west face is shaded. I show every input below so you can replace the reference figures with your postcode, roof and tariff.
What does the reference NW/SE system produce?
The reference design uses twelve 455 W panels: seven south-east and five north-west. Each face has its own input on a dual-MPPT inverter. The model uses the European Commission’s PVGIS calculator rather than a national rule of thumb.
| Input | Reference value | Why it matters |
|---|---|---|
| Location | 52.308° N, 0.717° W | Central England reference point, not a UK-wide promise |
| PVGIS query | API v5.2, crystalline silicon, building-mounted | Keeps the estimate reproducible |
| Roof slope | 28° on both faces | Changing pitch changes the yield |
| System loss | 14% | Already included in the PVGIS yield figures |
| South-east face | 7 × 455 W = 3.185 kWp, aspect -45° | PVGIS yield: 931.46 kWh/kWp/year |
| North-west face | 5 × 455 W = 2.275 kWp, aspect +135° | PVGIS yield: 643.55 kWh/kWp/year |
| Conservative sensitivity allowance | 0.96 multiplier | A deliberate 4% deduction for planning. It is not a clipping result derived from hourly PVGIS and inverter data. |
The arithmetic is:
- South-east: 3.185 kWp × 931.46 kWh/kWp = 2,967 kWh/year.
- North-west: 2.275 kWp × 643.55 kWh/kWp = 1,464 kWh/year.
- Combined before the sensitivity allowance: 2,967 + 1,464 = 4,431 kWh/year.
- Reference result: 4,431 × 0.96 = 4,253 kWh/year, rounded down.
For comparison, the same 5.46 kWp facing south models at 5,126 kWh after the same deliberately conservative 0.96 sensitivity allowance. The weighted NW/SE layout therefore produces about 83% of that south-facing reference. An equal split would do slightly worse because more capacity sits on the weaker north-west face.
How does that compare with a measured split roof?
My own 28° garage array has five 430 W panels south-east and four 430 W panels north-west. From 9 March 2025 to 9 March 2026, the south-east string recorded 2,066 kWh and the north-west string recorded 1,143 kWh. That works out at 961 and 664 kWh per installed kWp respectively.
| Face | Capacity | Measured energy | Measured specific yield | PVGIS reference |
|---|---|---|---|---|
| South-east | 2.15 kWp | 2,066 kWh | 961 kWh/kWp | 931 kWh/kWp |
| North-west | 1.72 kWp | 1,143 kWh | 664 kWh/kWp | 644 kWh/kWp |
The measured relationship is close to the model: the north-west string produced 69% as much energy per kWp as the south-east string. It still supplied 1,143 kWh over the year, but most of its useful contribution arrived in the brighter months. Your shade line can change that result more than a small difference in panel efficiency.
Why is a NW/SE roof different from a standard roof guide?
A single south-facing array has one dominant midday generation curve. A NW/SE array combines two unequal faces. The south-east panels start earlier and produce more annual energy. The north-west panels contribute later in the day, especially in summer, but lose more output in winter and can be more exposed to shade from ridges, trees and nearby buildings.
That changes the design in three practical ways:
- Model each face separately. Multiply the installed kWp on each face by its own PVGIS yield, then add the results. Do not apply one south-facing kWh/kWp figure to the whole roof.
- Give each face independent tracking. Put the SE and NW strings on separate MPPT inputs. Check the panel count, operating voltage, cold-weather open-circuit voltage and current against the chosen inverter’s current datasheet.
- Check the north-west shade before filling it. A roof survey should show horizon and near-object shading by month. Extra NW panels can be worthwhile, but the unshaded PVGIS figure is not enough to prove it.
If roof space allows, putting more capacity on the south-east face raises annual output. The 7/5 split used here is not a universal optimum. It is a transparent example that keeps both strings long enough for a suitable inverter while weighting the stronger face.
What household usage does this guide assume?
I use 3,500 kWh of annual household electricity as a planning case, with no heat pump or regular home EV charging. It is not presented as the usage of every four-bedroom home. Your last twelve bills are the correct input for your house.
For the no-battery case, I assume 25% of solar generation is used in the home as it is produced. This is a conservative default for a household that is often out during working hours. It means 1,063 kWh is used directly and 3,190 kWh is exported. Annual consumption limits how much solar can be valuable, but it does not by itself tell us when that consumption occurs.
What is the reference system worth each year?
This is a tariff illustration, not a current supplier quote. I use a flat 25p/kWh import price and 5p/kWh export price so the calculation is easy to reproduce. Replace both with the rates you can actually obtain. Export eligibility and rates depend on the supplier and evidence it requires, so do not count export income until that is confirmed.
| Step | Formula | Annual value |
|---|---|---|
| Generation | PVGIS face calculations × 0.96 sensitivity allowance | 4,253 kWh |
| Directly used solar | 4,253 × 25% | 1,063 kWh |
| Exported solar | 4,253 – 1,063 | 3,190 kWh |
| Avoided imports | 1,063 × £0.25 | £265 |
| Export income | 3,190 × £0.05 | £159 |
| Illustrative annual value | £265 + £159 | £425, rounded down |
The formula to reuse is: (generation × self-consumption rate × import price) + (generation × export rate × (1 – self-consumption rate)). A higher export tariff narrows the value of shifting energy into the home. A higher import tariff widens it.
Does this split roof need a battery?
No. A battery can increase self-consumption, but the roof orientation does not make one compulsory. Storage is worth considering when you regularly export solar and then import electricity later at a much higher price. It is less compelling when somebody is home during the day, an immersion heater or EV can absorb surplus, or your export rate is close to your import rate.
As a sensitivity check, raising self-consumption from 25% to 60% in the tariff illustration changes the annual value from about £425 to £723. The gross difference is about £298 a year: 1,489 kWh shifted from 5p export to 25p avoided import. That is an upper-level planning figure before battery conversion losses, standby use, degradation, finance and eventual replacement.
Do not turn that £298 into a payback claim until you have half-hourly import and export data. A battery cannot shift more energy than the household exports, the battery can accept, and the household later consumes. The battery size calculator is a tariff-arbitrage screening estimate. It can help test battery capacity against your usage and tariff, but it does not model when this split-roof array produces solar, so do not use it as a solar self-consumption forecast.
What system would I ask installers to quote?
For this reference roof, I would ask for a quote around twelve modern panels with seven SE and five NW, connected as two independent strings. I would ask for the exact inverter model and a written string calculation, not just “dual MPPT” in the sales description.
A 3.68 kW export-limited design can suit a 5.46 kWp split array because both faces do not reach their peak at the same time. However, the inverter manufacturer must permit the proposed DC capacity, string voltage and current. The DNO connection route depends on the equipment and export arrangement, so the installer should confirm whether the design follows G98 or needs prior approval under G99.
I would price the battery separately. That makes the extra hardware cost visible and lets you compare it with the value of the additional solar you expect to keep. If you may add a battery later, ask what metering, inverter and warranty constraints that choice creates now.
What should you check before accepting a quote?
- Use your postcode in PVGIS and request the annual output for the SE and NW faces separately.
- Ask for the installed kWp, pitch, aspect, system loss and shading allowance used in the quote.
- Check that each roof face has independent MPPT tracking and that both strings remain inside the inverter’s voltage and current limits.
- Use twelve months of bills for annual demand and smart-meter data for when you use electricity.
- Run the cash flow with your actual import and obtainable export tariffs. Keep standing charges out of the claimed saving.
- Ask for solar-only and solar-plus-battery prices so storage has to justify its own cost.
A NW/SE roof is not a failed south-facing roof. It is a two-face design that needs two yield calculations and a realistic usage model. If a quote shows those inputs clearly, you can judge the weaker NW face on its own merits instead of accepting or rejecting the whole roof by rule of thumb.
