Solar in California (ca 2024)
Here are some long-running notes on our solar system.
2024: Design
I recently spec’ed out a solar system. It’s not installed yet, but the contract is signed.
Our situation is:
- ~16 kWh daily consumption
- Mostly electric appliances
- Gas furnace
- 1 EV
I spent a lot of time thinking about the design, and the key principles were:
- Solve today’s problems
- Maximize future optionality
- Scale horizontally, not vertically
In California in 2024, you need a battery today since you’ll be on NEM 3.0, and it doesn’t pencil out without a battery. On the other hand, batteries will easily dominate the cost of a new installation. On the third hand, the energy storage landscape is changing extremely rapidly.
- Battery chemistry improving rapidly over time
- Evolving standards like V2H would unlock huge kWh of stranded storage
- Emerging trend of “appliances with batteries”
So I opted for “just enough” battery today and the end result is:
- 13.5 kWh battery (FranklinWH, Franklin aGate + optional generator module)
- an SS2-50 generator inlet port
- SPAN smart panel
- SPAN Drive (EV charger)
- Backup by Biolite
The Franklin system is price competitive with the Tesla PW3, but the key feature to me was the optional native support to connect a generator. Bonus of not not giving money to Elon.
The SS2-50 is IMO the best way to preserve optionality. Since it’s a standard, you can do things like hook up a “battery generator” to it, or even a car, if your vehicle supports it. You could use your external power source as a true emergency backup, or just as a way to add more storage to your system without having to get an electrician to do any expensive rewiring of stuff.
- https://youtu.be/WQljBFxqp4k?t=154 is the Anker F3800, which is far cheaper and more expandable than a typical Franklin / Enphase / Tesla whole house battery.
- https://youtu.be/hGaftao5E7U?t=54 is a Chevy EV Silverado which has a massive 200 kWh battery which is a literal order of magnitude larger than the Franklin.
The SPAN is important for my home, since we are stuck with 125A service, and having software configurable load-shedding seems useful to keep overall load within limits. Better yet, it also integrates with the Franklin ecosystem and the SPAN Drive, so that you can do things like, “only charge the EV after I’m done charging the home battery”. The Drive doesn’t support V2H but that’s why the generator inlet is so key. We’re partially insulated from the standards evolving.
Last is the Backup by Biolite. Although this device isn’t shipping yet, I’m a backer on Kickstarter. I opted for the 3 kWh version to power my fridge. It’s cheaper $ / kWh vs the FranklinWH and allows me to underspec the house battery by 3 kWh. Also, it embodies the idea of horizontal scaling rather than vertical scaling.
This is all stuff that is under contract.
But also, there are two other appliances I’m strongly considering for the future.
They both implement the horizontal scaling idea, by bringing their own energy storage, and again, allow me to continue growing the overall storage capacity without too much additional, expensive, electrician work.
I hope to have real world usage data in a few months and will update here
2025: Installation
The install took about 2 weeks, mostly because the final design required about a 20 foot hand-dug trench from the house to the carport. Other than that, it was pretty standard.
From left to right, that is:
- Enphase combiner box, that houses the connections from the roof-top panels.
- New “dumb panel” that I needed to bring me up to code. It only has a single giant breaker inside for the 125A service.
- SPAN smart panel. So pretty!
- Franklin aGate, which controls the battery.
Local code required that there could be nothing flammable (ie, wood) over the battery. Since my house’s roof overhangs the perimeter of my house by about 5 feet all around, that meant I could not place the battery closer to my house unless I was willing to install a very expensive fire suppressant system. I was not willing to do this, so the next best place was by the carport.
This was the 20 foot hand-dug trench, plus building this mounting structure with a concrete base and enough space for 2 batteries. Although I only installed one.
2026: First Year Retrospective
This was the first year that we had solar!
I learned quite a lot. But first the stats.
Generated 8.8 MWh off a 7.4kW array.
- Directly consumed 2.4 MWh
- Stored 3.4 MWh in our battery
- Exported 2 MWh back to the grid
Our actual usage was 6.4 MWh, which doesn’t align exactly with the numbers above, because our battery is not big enough to cover 100% of our needs 100% of the time, so we ended up importing from the grid during the winter months.
- Direct solar consumption: 2.4 MWh
- Pulling from the battery: 2.9 MWh
- Grid import: 1.0 MWh
The app estimates we saved about $2200 in energy costs, which feels approximately correct. The payback period for this installation is unfortunately still extremly long, but hey, we avoided creating 3.4 tons of CO2. In California, that seems to be worth about $25 per ton at cap-and-trade auctions, so I’m going to claim an additional $85 of value.
So what did I learn in my first year of solar?
I was (predictably) obsessed with the stats generated by the system for the first 6 months, constantly checking in on the various apps to see what was going on. But after a while, I mostly stopped worrying about it.
I ran the battery down to 5% reserve for most of the year to maximize cost savings, but at times, when major storm warnings occurred, I bumped the reserve up to 70% to hedge against a grid outage. We got pretty lucky here, never actually losing grid power, but it was nice peace of mind.
Our usage is fairly consistent the entire year. We don’t have an A/C, and we have a gas furnace, whose electric blower consumes about 1 kW when running. Our system isn’t sized to cover the winter months – our 7.4 kW array isn’t big enough to fully charge our 13 kWh battery, and a 13 kWh battery wouldn’t cover our daily usage.
If we really wanted to be fully self-sufficient during the winter months, we’d need to double our storage capacity and probably triple our generation capacity, both of which would be extremely cost-prohibitive in 2025 prices. Plus, I’m still waiting for the EV industry to standardize their V2H implementations because it seems criminal to pay $20K for a 13 kWh battery when it’s possible to buy an F-150 Lightning with a 100+ kWh battery for $60K and you can haul mulch with it.
I have not used the generator input at all, although I guess I’m glad to have it.
The Biolite Backup hasn’t shipped because of normal Kickstarter engineering growing pains, but also because of the tariffs. Fuck Donald Trump.
We haven’t done any remodels so battery-powered induction cooktops are still a pipe dream.
Cars. Ugh.
In 2025, we drove:
- 6985 miles in our Chevy Bolt, which is our “daily” driver (more like semi-weekly).
- Averaged 3.9 mi/kWh, so 1791 kWh or 1.79 MWh (!!)
These numbers reveal how utterly wasteful driving is. My car commute is ~20 miles round trip, which requires 5.1 kWh of electricity, while the entire house consumes between 10 – 15 kWh per day.
It’s depressing to see how energy intensive it is to move that much mass around. Anyway…
Figuring out where this electricity came from is a bit of an exercise, since I charge at home and at work. When charging at home, my SPAN smart panel is typically set up so that it will only charge the car after the house battery is full, and I typically do not pull from the grid.
This strategy works most of the year, but in the winter time, there are stretches where I do need to import from the grid specifically to charge the car, so I reconfigure SPAN to pull from the grid.
The net is that almost all of my car charging actually comes from my own solar panels, thus making our Bolt extremely clean to operate.
Anyway…
- SPAN tells me this circuit used 1.31 MWh in 2025, but it doesn’t give me a breakdown of grid import vs self-generation.
- My best guess is that we imported ~347 kWh (Jan, Feb, Dec), since those are the months where we don’t generate enough to both charge the house battery and the car. At $0.31 / kWh (PGE’s E-ELEC winter off-peak rates), that works out to $108.
- The missing 480 kWh came from charging at work, but even in December, the California grid is still pretty darn clean.