Your analysis is way too conservative. An analysis by Tesla Motors some time ago calculated that if you drive the Roadster an average of 35 miles a day, you need about 105ft^2 of solar panels to produce all the electricity for it, assuming 18%-efficient solar panels. [1]

The Tesla Roadster has an area, when viewed from the top, of 72ft^2. [2] That means if you were able to cover it with solar panels that were 27% efficient, you could power it just from the power it produced.

As far as your last point, I agree: I'd rather put the solar panels on the garage roof, where I can somewhat control how much sunlight it gathers, than on the car.

1. http://www.teslamotors.com/blog/electric-cars-and-photovolta...

2. http://www.netcarshow.com/tesla/2008-roadster/

Good numbers in the Tesla Roadster wikipedia introduction.

wikipedia: The tesla roadster use 490 kiljoules per kilometer.

compute: Our 30 megajoules above will take us 38 miles.

Of course I'm still using mythical 100% efficient solar cells on a biggish flat roof and all sunny days, and Tesla used mythical shape hugging solar cells on all of their exposed surfaces.

I still think the best solution for a solar powered car is to place your solar cells the south western deserts near LA and pull what power you need from the grid. That way you get a lot more power delivered from your cells.

> The only function for a roof mounted solar panel is to look really cool in concept art.

Not true. The Prius has a solar panel option to power cabin ventilation when parked in warm climates. This makes a lot of sense because the aircon load required to cool down a hot car quickly is huge, but the power required to prevent the heat building up is not.

Perhaps solar panels could be tied to a battery that powers the dash electronics, so that they don't have to siphon power indirectly via gasoline? Small potatoes compared to actually moving the weight of the car, but it would still be cool.