For example, one person in a 50 m² apartment, divided by 4 floors, with 20% efficient panels, 25% duty cycle for night/latitude, still implies 625 W/person. That won’t power everything all the time, but it would reduce the required size of the external electricity source — be it PV, nuclear, fossil, whatever — by a decent amount.
This whole rooftop solar movement reminds me of 'solar roads', a snakeoil scheme a couple of years ago. It has no rational basis, and no compelling engineering support. Its just a 'roofs are flat; solar panels are flat; hey! lets put solar panels on rooftops!'
E.g. does the rooftop have an appropriate southern exposure? Is it in direct sunlight? Does it have the carrying capacity for tons of equipment? Is it a safe place to do construction? Will the roof's normal function be unaffected by solar installation?
The answer to all these things is, No.
Second, even if all the things you’ve just said “no” about are true for 100% of existing property, if this is part of the planning regulation for new buildings, then the designs can be required to be such that the answers are “yes”. I know this is technically achievable by way of having at least one example of PV on a roof.
And again, one roof for many families doesn't scale. A solar farm has every advantage over this scheme.
Do you think that isn’t obvious? So keep new apartments away from houses and vice versa. So what, no big deal.
The idea is called “planning permission” and why sewage plants are not normally next to residential areas.
Problem solved a very long time ago — which is convenient, because this is the exact same problem you’d have to deal with when planning the solar farm in the first place, because the apartments can shade solar farms too if they’re built too close.
> And again, one roof for many families doesn't scale.
Not claiming it scales with arbitrary height. It’s clearly constant with arbitrary height. Claim is that it’s still useful for ~4 stories with one family per story.
> A solar farm has every advantage over this scheme.
Except for land cost, grid capacity including connecting to the farm to the grid in the first place, and the money rooftop solar saves on the cost of the roofing material, and the way local power generation at least enables the possibility of power when the grid is down like happens sometimes in CA fire season, and the way this puts more money in the hands of normal people whereas a traditional solar farm or power plant will put it in the hand of shareholders, and the fact that large land use changes have their own ecological issues and NIMBYism (remember, if we suppose that the rooftops don’t cover enough area, then the solar farm you want to build is larger than the combined area of every building in the city it’s associated with), and the way rooftops get built every time a residential building gets built so in the case of houses it will scale perfectly with population when a PV farm or an apartment won’t. (Apart from that, what have the Romans ever done for us? Etc.)
Edit:
Just to be clear, if you were to say “American building codes make rooftop PV needlessly expensive”, I’d agree with you. My in-laws and parents both had PV installed in the UK, and even a decade ago it cost them less than half of the prices I see quoted on HN for American rooftop PV today.
However, this an entirely separate argument to the one you appear to be making.
If a city wants more solar, then they could put up a solar farm in some rational place with a good economy of scale, and connect it to the grid.