Solar + EV off-grid: how I sized my system to charge my car entirely on sunshine
By KilowattKarl·4 replies·533 views
I've been running a solar + battery home with a Chevy Bolt for two years. The goal from the start was to drive entirely on sunlight — no grid power for the car, ever. Here's how I sized the system and what the reality has been.
The design constraint
My Bolt uses about 9kWh/day on my typical commute (28 miles each way). I wanted to charge from solar-stored energy — no grid pull for the car. So the system needed to generate that 9kWh as net excess after covering the house's own consumption.
House load first
Before I sized a single panel, I did a detailed energy audit. My house uses about 24kWh/day average (high because I have a well pump, electric heat pump, and a teenage son who apparently thinks the refrigerator door is optional). So total system need: 24kWh house + 9kWh car = 33kWh/day.
Panel sizing for Denver
Denver averages 4.9 peak sun hours across the year (lower in winter, higher in summer). With my 6.4kW array and an 80% system efficiency factor: 6.4kW × 4.9 hours × 0.80 = 25.1 kWh/day average annual production. That's actually less than my 33kWh need. But my battery (20kWh) allows banking sunny days against cloudy ones, and my usage is lower in summer (no heat pump) and higher in winter (heat pump + less sun). It balances.
Reality check: winter
January is hard. Average production drops to 13–14kWh/day and my heat pump consumption goes up. During extended cloud cover I pull from the grid for house power and let the solar handle the car charging when possible. I hit 73% grid-free for the car in January (worst month), and 99%+ from April through October.
Annual numbers
Total estimated solar production: 8,400 kWh/year. Grid import: 1,100 kWh/year (primarily January–February). Net metering export: 1,800 kWh/year (summer surplus). My annualized grid cost: about $95/year including the utility's fixed connection fee. My annualized energy cost before solar: $2,400/year. I won't give a payback number because everyone's situation is different, but the math works well in Colorado.
Anyone else tracking their solar-to-EV ratio? I'd love to compare notes on winter performance specifically.
My off-grid setup doesn't export so the comparison is a bit different, but I track solar-to-Bolt percentage. January in central Idaho: 68%. October: 98%. July–September: 100% (actually have excess to spare). The seasonal variation is significant but the annual average is right around 90% solar-charged.
The thing that helps most in winter: parking in a garage when possible. A cold battery (below 15°F) charges much more slowly and accepts less power. A garage-kept car at 40°F charges at full speed. That 20°F difference in battery temperature meaningfully affects how much solar charging you can actually absorb on a winter day.
The heat pump interaction with solar is an interesting optimization problem. Heat pumps are significantly more efficient than resistive heating but draw more peak power. If you run the heat pump to pre-heat the house during peak solar hours and let the well-insulated house coast through the evening, you shift a large load to solar-friendly hours. This is the kind of thing that's hard to do manually but some smart home systems (and the Span panel) can automate.
Real question: how are you handling net metering with your utility? A lot of utilities have been cutting NEM rates or adding fixed charges for solar customers. In California, NEM 3.0 cut export rates dramatically and the payback on new systems got much longer. Colorado's regulatory environment has been better, but I'm curious if you're seeing any changes at your utility.
@sarah I'm on Xcel Energy in Colorado. They're still running a reasonable NEM program — I get near-retail credit (~$0.10/kWh) for exports. The fixed "renewable energy credits" fee is $9/month on my plan, which is annoying but not system-breaking. There have been whispers that Xcel will revise the program but nothing's happened yet. I'd definitely size differently today if I were starting from scratch under a bad NEM scenario — less focus on exporting, more focus on self-consumption with battery storage.