Earlier quoted context omitted.
I think the largest problem in orbital manufacturing is actually rejection of waste heat. Otherwise, being able to run robots in an oxygen-free environment and move around large masses without conveyors sounds pretty good.
It's also an environment full of cosmic rays, which requires hardening of all your electronics, an environment where fluids don't flow through pipes, an environment where dust and powders don't settle, an environment where conveyor belts don't work because things float off them... Almost every single industrial process that we have relies on a constant, freely available, predictable, unidirectional acceleration of 9.…
There is no 20 Mg steel smelter in orbit. The use cases are very different. First off, you're not starting with coke and iron ores and ending with flat carbon steel. Whatever you make is staying in orbit, to replace something that would otherwise have to be launched. You probably need to make 17-4 PH stainless steel, with synthetic thin-film silica surface coatings to prevent vacuum welding. And where is the raw material--iron, nickel, tantalum, chromium, etc.--coming from?
You can't just replace an existing industry with vertical conveyor belts 5 miles high and expect it to work.
Firstly, the raw materials: https://en.wikipedia.org/wiki/Mond_process , https://en.wikipedia.org/wiki/Carbonyl_metallurgy , https://en.wikipedia.org/wiki/Chemical_vapor_deposition .
You would likely be reacting powdered asteroid dust with CO + H2 (syngas), distilling the carbonyl vapors, and depositing nearly pure metals at higher temperature. The purified metals do not subsequently react with oxygen, because that is all kept bottled up for other purposes. Now you don't need the Bessemer process or oxygen converter process, because the iron is already pure, and also you don't have enough oxygen to waste on it. Since you're already using CVD in high vacuum to build up your ingot, you might as well just add the alloying elements and silica coating right there, on the asteroid, before you fling it to whatever orbit you want it to be in.
Metal carbonyl chemistry works at relatively low temperatures, below 250 degC. You don't have to worry about toxicity if you don't even have humans on site at the facility.