Part 1 · Hardware
The system
Installed on 4 June 2026 on an east/west pitched roof in Hampshire. The quote said 21 panels, but the installers found room for a 22nd on the day, so we got an extra 475W for free.
Specification
| Panels | 22 × Aiko 475W N-type ABC (the proposal said 21), east/west split roof |
| System size | 10.45 kWp |
| Inverter | Sigenergy EC 8.0 SP (8 kW hybrid, EMS built in, Modbus TCP over the LAN). The installer first fitted a 6 kW EC 6.0 SP by mistake, and it got swapped for the right unit. The full story → |
| Battery | SigenStor BAT 6.0 + BAT 10.0 = 15.06 kWh LiFePO₄ nominal / 14.60 kWh usable, 10,000-cycle rated. Combined continuous ~7.6 kW, capped by the inverter. |
| Orientation | East group: azimuth 98°, tilt 30° · West group: azimuth 278°, tilt 32° |
| Est. annual generation | ~7,500 kWh (pre-loss, scaled for 22 panels; ~11.5% annual shading loss per survey) |
| Grid connection | G99 approved with no export limit, so the inverter's full output can go to the grid once export certification comes through |
| Price per kWp | £1,148 |
| Tariff | Octopus Agile import with Outgoing fixed 12p export. Why, and the Flux story → |
An east/west split is better than it sounds. Peak output is lower than a south-facing array, but the generation spreads across the day: morning sun on the east face, evening sun on the west. With a battery and Agile pricing that suits us nicely. The generation curve lines up better with when we actually use power, and there's less midday clipping against the inverter's limit.
Warranties
| Component | Cover |
|---|---|
| Panels (product) | 25 years |
| Panels (performance) | 30 years, ≥87% output |
| Inverter | 10 years |
| Battery | 10 years |
| Workmanship | 10 years |
The original business case
The pre-install modelling, based on our actual smart meter data from January to March 2026 (we use about 3,667 kWh a year), reckoned on roughly £1,900 to £2,060 of value in year one against the £11,999 cost, with payback somewhere around year six and a half. This site exists to test that projection in public, every day, on the stats page.
| Revenue stream | Year 1 est. | How |
|---|---|---|
| Self-consumption | £1,100 | Avoid importing at Agile rates (evening peaks can hit the high 30s in winter) |
| Battery peak-shifting | £260 | Discharge 4–7 pm to dodge the Agile peak |
| Cheap-slot charging | £120 | Charge in the cheapest overnight half-hours |
| Axle VPP events | £150 | £1/kWh grid-stress dispatches |
| Saving Sessions | £120 | Auto-join + force discharge |
| Weather optimisation | £89 | Solcast-driven charge planning (Predbat) |
| Load shifting | £75 | Appliances on solar surplus |
| Total | ~£1,914 |
Projections assume 4% energy inflation and 5% system losses, and were modelled before the 22nd panel went on. Treat them as the hypothesis. The stats page is the experiment.