Yes, it wasn't a complete abstract, there was no intent to be misdirectional, and I had previously acknowledged the "supply chain" issue and that the link specificly mentioned asteriod mining as opposed to LEO debris. The LEO debris is just a starting point - again, because it is close, easy to reach, and needs clearing up anyway. If using this to mostly make things, it reduces total cost of deployment of these things, realistically sillicon chip production is likely to be one of the later things deployed, for a variety of reasons. But there's no reason it can't evenually be made to happen. Ofc, once mass harvesting sillicon this won't be much of a concern overall, with facilities in place to be building almost anything else it's possilbe to account for this "supply chain". The earlier we can produce this, though, the better.
Near zero is about all we can manage, realistically, until we've amassed more resources "up there" via the likes of mining(say by throwing one of these machines out into the asteroid belt past Mars) and then it really can be self replicating. Near zero is still a damn sight better than "you're going to have to lift everything". One of these concepts makes things a little more realistic to persue. You're not going to find a large cube of titanium, but you can find plenty that will be able to be rendered into a cube - or a more suitable shape - readily. You'd definitely not need a couple of tonnes of CNC machinery - a few six axis arms are almost essential but not for retrieving crap floating thorugh space, you'd use something more like a net - maybe more like a wind sock would be more sensible. The arms would hold and or mainpulate parts and toosl. All this could be run from solar power, but there might be some better tricks to deploy. Possibly hydrogen generators, with an on-demand system splitting water harvested from comets. Making a perfect copy of itself, sans fuel, is still useful, as it reduces the "cost" to make it operational to fuel. Once this can be sourced externally then this is even better. The goal is even for fuel, but we'll have to work with what we can wrangle.
One way for dealing with "fluids in microgravity" and their behavioural adjustments is potentially Tesla's valve system, or something based on. I was under the impression the extruder head on printer nozzles operated under pressure other than gravity feed, commonly, it shouldn't be an issue. Another is to just centrafuge the lot and make artificial gravity. Or you could 3D print the moulds in plastic, disolve the metal as a hyperconcentrate solution, insert a seed crystal into the mould, inser the solution into the mould, very slowly cool the solution and then lend the seed crystal a little kinetic energy and watch as physics moves all the atoms in the metal to form a single crystal the shape of the mould. Discard mould and finish part.
Coolant on drill bits and the like was a concern, With artificial gravity Selective Laser Sintering might be possible, but otherwise carving from a block using a laser sounds plausible to me. Otherwise coolant can largely be recycled and the refuse lumps of shaven metal it sweeps can be harvested.
Yes, the "tooling" involved in CPU fabrication is precision affairs, you may consider me "dismissive" but I'm aware it's possible. That's all I need. Once I have machines that can build their own parts, they can build themselves again, mostly, with subtle differences that result in more precise output - a few cycles of that and it can produce something suitable, surely. The supporting infrastructre may indeed be extensive, but it's not as if there's any shortage of room in space, so we're good there. You avoid sending it sillicon all the time by stealing asteroids from problematical orbits, and from resources thrown back by the first seed you throw out past Mars to start on that asteroid field.
The mass on the moon is relatively effected, it's redistributed - then it's absent 5 million tonnes. Five million laps later, it's orbit is now noticably different.
Doing it solo isn't likely what's going to be happening, I can already draw on a lot of research in a lot of directions to springboard the entire affair, and hopefully I'll encounter others with more vision - I know they exist, we just need to cross. Considering 97% of what I'm talking about already exists, it just needs to be combined together in the "right" way then I don't really see why it shouldn't be possible. You might well lable it madness, I see plenty talking greater insanity. The end result shouldn't require to be massive, either. What I have in my shed might be, but what we lift can possibly unfold itself and assemble once it's up there automated, using LEO debris for the last bits of panelwork etc. There's far more material in LEO than you give credit. Look into it. More than enough to pop off a few dozen tugs a copy of itself and a few centrafugal launchers - yes, it's likely to require additional lift to make these things functional, but it reduces the cost of deployment, significantly. I wager that there's enough afterwards to build much of the facilities to process resources when they get thrown back from past Mars. Selling 40% of that to earth should pay for getting the "missing" equipments and parts to the facilities. If we can score contracts to remove LEO debris, then this can even be made to pay for itself somewhat. Even if things like making the bulk of EM-Drives is unsuitable due to it being hype, we can do things like make the Xenon tanks and much of the ion propulsion systems for the tugs up there - reducing lift to electronics packages and fuels. Once we've some facilities in place, collecting more scrap will literally cost fuel - add 10%-15% on the cost of getting that fuel there and I'd wager most others offering a similar service wouldn't be able to compete. At the end of the day, this initative doesn't need to be made ot pay for itself, but aiming for it can't hurt. It's the mining between Mars and Jupiter that comes with the payback. Recyclling LEO debris is good practice for being able to build additional units and spare parts remotely. Yes, there will still be a requirement for "external" goods, this is unlikley to be completely eliminated in the early phases, but we can reduce it significantly and that will matter. Eventually as resources and capacities expand, this will solve itself.