Our PV system consists of sixteen roof-mounted Mitsubishi 185 watt panels, for a nominal 2.96 kW DC at the panels. For PV panels, the "rated power" is for 1000 W/m² solar intensity perpendicular to the panels at 20°C. In most of the U.S., that's roughly the intensity at midday on a very clear day from April through August. In our location, with very low humidity (no haze) and very clean air, we meet or exceed that for several hours a day through much of the summer.

We have a Fronius IG 3000 inverter with data logging. The stated efficiency is 95%, which should turn 2.96 kW DC into 2.8 kW AC at the grid. We never quite achieve this. Maximum AC power in summer sometimes reaches 2.7 kW, but 2.5 kW is a more typical AC power at solar noon. However, we did hit 2.8 kW in early April 2010. I think it's because the panels are much cooler in the spring – daily temperature max ≈60°F rather than ≈95°F – and the efficiency of the panels deteriorates with increasing temperature.

Visit the technical page for more details of our solar electric generation.

This is a question I asked when trying to decide what size system we needed, but I had a hard time finding an answer. Information on peak power is easy to come by, but – allowing for weather, changing day lengths, etc. – how much total energy is produced in one year? The answer is clearly location sensitive – better in locations that get more total hours of sunshine – so our results shouldn't be applied blindly to other locations. But in the coastal mountains of central California, where winters tend to be fairly cloudy but summers are virtually full sunshine every day, the energy production of our 2.96 kW system averages 5400 kWh/year, which is 1800 kWh per year per installed kW. That should be a reasonably good estimate for much of the American Southwest, excepting areas close enough to the coast to experience summer fog.

Even though not quite electrically neutral, our electric bill is zero! This is because in California solar customers have time-of-day pricing. We generate excess electricity midday in summer and sell it to PG&E at a high price. Then we mostly buy it back at night in winter when prices are lowest. Strictly speaking, PG&E should send us a check at the end of the year. But guess what?

With ten year's data, we can estimate the payback time of the system. Because of time-of-day pricing, I can't compute the exact amount we save, but I estimate the PV system saves us $1000/year in electric bills. The system cost $22,600 installed. (It would be less now; the price of panels has dropped significantly.) A California energy rebate gave a $6200 credit upfront. Then there's a 30% federal tax credit. Altogether, our net cost was $11,600. A straight-line payback calculation at $1000/year gives a 11.6 year payback time. Rising energy costs will likely lower this to, perhaps, 10 years. So the system has probably paid for itself already, and it should last another ≈15 years with little or no maintenance.

Electricity is not our only energy use. We use natural gas year-round for cooking and also for on-demand hot water. In addition, we have a natural gas fireplace downstairs for supplemental winter heat. (See the Passive Solar Design page.) Our monthly bill from the gas company shows out energy use in the truly obscure unit of therms (1 therm = 100,000 BTU, for those of you who really wanted to know), but they can be converted to kWh for comparison to electrical energy.