There would be a "unnatural" orbit where solar radiation pressure would cancel the mismatched orbit, I think.

So the shades would stay in equilibrium in a bit higher solar orbit than their velocity would permit without the light pressure.

I'm having trouble visualizing how this would be stable - from their reference frame (orbiting a Lagrange point), the pressure is coming unidirectionally. They'd be accelerated in one direction constantly, which seems like it couldn't possibly be stable.

Actually, you are right, I was talking as if the shades are just orbiting the Sun.

L1 orbits do require station keeping.

Perhaps you could cheat a bit by making the shades black or oscilating/slanting them a bit forward/rearwards as to do the station keeping with very little fuel? Or shaping them like a lense to direct most light away from Earth, but not to experience too much net force?

Or they are just designed to be expendable?

For what it's worth, Wikipedia suggests making them shaped like a lens.

Yes. They need serious stationkeeping.

Do they? It seems to me that it'd be a lot simpler to send up dumb sunshades, and then simply keep launching more as previous ones drift. The linked sunshades would only weigh a gram each. How much would a sunshade weigh if it had to be an entire spacecraft capable of stationkeeping? Many orders of magnitude more.

You can't station-keep indefinitely anyway (your fuel will run out), so just launching lots more dumb sunshades seem like the more effective and cheaper solution.

While I think x,y,z will be an issue, I think the bigger problem is orientation.

Any sort of station keeping is going to add significant mass, complexity, and cost.

Why does orientation matter? Just throw up enough dumb sunshades. It costs a lot more in station-keeping to get the orientation correct than to just let it be random.

And keep in mind that, absent active station-keeping, you need a mix of many different orientations to block out the Sun throughout the year. And you're not going to perfectly remove any spin anyway.

I agree that would be the smart way to do it, but that would at least double the scope, probably more, since orthogonal area is a cosine function. There's also oblique reflection to consider.

Why do you assume that the linked paper hasn't already taken this into consideration? They know that they're basically just structured aluminum foil, weighing in at 1 gram each, and completely dumb; why would they be assuming orientation at all? The reasonable thing to do would be to assume random orientation (especially over time, i.e. rotation, and over the orbit).

I actually agree(d) with that reasoning, I also initially assumed random placement. Others on this page suggested them being steerable.

But we're both very wrong- I read the whole paper. Some aspects are well thought out and highly technical, while others are glossed over as requirements tbd.

Briefly: He does intend to steer/ orient the discs using radiative pressure from a network of control satellites and gps-like navigation "beacons", but overlap is also required to take into account the required oblique angles and transparency of the discs.

I'm left being very impressed with the provided technical details of the disk constellation, but even more skeptical overall, mostly due to lack of detail on some major supporting elements. He offers no detail of the required deployment method, or even the scope of the required navigation and control network. He also states each disk will require a gps-like receiver, 2 cameras, some power and processing capability, and a method of communicating with said networks, all while still assuming 1g per disc and no added cost...

It's not conceptual level, so much as it's conceptual for maybe one out of dozens of required massively scaled elements.

Also, it addresses no risk management / mitigation concerns.

The carbon cost of manufacture and transport seems like it would absolutely dwarf any benefit to such a scheme.

20 million tons of plastic to manufacture is small comparing to world production - and we're talking about world effects here. Launch pollution is calculated...

Not as much as you'd think, because they're actually lenses. They just redirect the light away to miss Earth, rather than blocking it entirely.