Here is post by Keith Henson from 2021-11-22 on the “Scale of the problem” topic about the difficulties in using lunar material as originally envisioned by Gerard O’Neill in the 1970s:
… I was deep into moon mining technology. The mass driver turned out to be much more difficult than originally thought, though now it might be relatively easy.
The problem is velocity scatter from the ejection end of the mass driver. I don’t remember the exact numbers, but they would not be hard to find in old papers by Tom Heppenheimer. The source of the problem was the switching speed and jitter of the SCRs that controlled the mass driver coils. Over the last 40 years, this has become perhaps 1000 times better with things like GaN fast power transistors.
In the days when Heppenheimer was working on the problem, the only solution was “achromatic” orbits where (by analogy to lenses that bring different wavelengths to the same focus) mass driver launches from a particular point on the moon with a little variation in velocity wind up (with some scatter) at the same place, the “catcher” out near L2.
I spent close to ten years looking into power satellites built from the earth simply because the complications of a moon mine, mass driver, and all the processing needed to fabricate SPS parts were beyond me.