Yup, lots can be 3D printed, I'll be looking into how to do the most with the smallest number of machines soon. Combined with various CNC and possible metal printing technologies, an yes it'd take an entire workshop, that's kind of the point - giving it the ability to make pretty much anything. Yes lots will be custom, almost all of it in fact. But as this will also be open source then other can improve upon it and add to it rapidly before we even print the parts for the seed. Building things terrestrially first will give us manufacturing capacities in our hands and make it trivial to build what we throw into orbit.
Not everything can be printed, however, and a CNC rig and robot arms decked with interchangable tools can do quite a lot. Enough to make the tools to do anything else, almost. Ofc this will all be designed with microgravity in mind, but we could just centrafuge the lot to simulate gravity.
There certainly will be consumables. Some solar panels might still have service life in them - but probably not - same with batteries. Some of this could potentially be recycled. But the goal is to use as much as possible to build - if this means 0% consumables, perfect, but that may take some cunning. Minimally we can definitely reduce the launch from tonnes to a few hundred KG, building more that will have more long term function. Harvesting scrap can be accelerated be printing off more tugs. It's not a meagre stash, either, pretty much everything that's gone out of service life since 1960 that hasn't been decayed or de-orbited(most of it) is still there. Then each time they launch a rocket they leave the upper stage floating - there's quite a few of them, and they're prime for panelwork or melting into blocks to feed into the CNC. Considering there's already 3D printers that can print most of their own parts, and the CNC machine I'm making later this year should be able to make all of it's own parts when I'm done, that's an easy 2/3 of self replicating right there. Once that's done I'll be automating my reflow oven, and making some pick n place machine.
A coating of non-conductant should suffice, doens't matter much what it is as long as it does't conduct. What matters most is function, not form - if we can do this with something we don't have to lift then bonus. And it's coils precisely I was thinking of, this would reduce required lift down to magnets for being able to build motors and such - unless used electromagnets, but I'd not imagine much ferrites laying about. Much titanium and aluminium tho.
Yes making chips is a "rediculously complicated process" - but it's not as if the Bridgman–Stockbarger technique is a well kept secret, or the lithographic process involving UV laser through a sapphire lense grown from a single crystal to the required shape - these are pretty easy things, the difficulty is in the details and a little practice and a bit of math previous and this shouldn't be hard to nail. I'd not suggest this be starting functionality, but something it'd add to itself over time. It'd be better to start with it ofc, but then we'd have issues with feeding it, initially. Might be able to reprocess old chips, not overly confident.
Regarding the manipulation of junk, I'd personally use three possibly six units with what could be best descsribed as a wind sock strung between them - they can incercept and position themselves in front, then just slow down a tiny bit and allow it to put itself in the hole - via a draw-string type mechinism the apature can be closed to prevent it's escape as they steer it.
Regarding the moon, 0.000001% will add up over a few thousand orbits. Over a few decades it'll become noticable. There's also when you remove the mass, too, that'll impact. And you'd need to mine those sorts of numbers to get a tonne or two of helium. It's really not a well thought out plan at all. Mining was cleared by who, precisely? Because I was under the impression that the outer space treaty forbids the claiming of celestial bodies, such as the moon. Ergo it is not theirs to provide permission for such. Actually looking for evidence of permission (yes, forum uses encapsulation of underscores as formattitng. if you encapsulate your link: < link > without the spaces it'll become clickable: https://en.wikipedia.org/wiki/Outer_Space_Treaty) I did manage to come across a nice little extract from somewhere else wikipedia which seems to validate my standpoints:
Since 1980 there has been major progress in miniaturization, nanotechnology, materials science, and additive manufacturing, so it may be possible to achieve 100% "closure" with a reasonably small mass of hardware, although these technology advancements are themselves enabled on Earth by expansion of the supply chain so it needs further study. A NASA study in 2012 proposed a "bootstrapping" approach to establish an in-space supply chain with 100% closure, suggesting it could be achieved in only two to four decades with low annual cost. A study in 2016 again claimed it is possible to complete in just a few decades because of ongoing advances in robotics, and it argued it will provide benefits back to the Earth including economic growth, environmental protection, and provision of clean energy while also providing humanity protection against existential threats.
So, NASA think it's viable, too. although their talking more asteroid mining there I feel, that was what I was initially thinking when I first thunked this madness - it's just possible to use it to clear up LEO too.