For example, I've talked with one shop about producing new (yes, brand new) Model M backplates in carbon-steel using an '87 as a template, stamped with tighter thickness tolerances than the original. If you wanted them in Ti or Al? Also possible. All using the exact same method IBM did. If I wanted a fully machined Model M chassis? Believe me, they could do it - easily.
... hell, I should do an IC on that. Though the templating costs would be pretty high.
Ooh, I'd have interest in that. Though what you should really do is rework the Unicomp Ultra Classic with a new plate and enclosure. That way we'd have the Windows key that's really necessary these days, and build quality that meets or exceeds a vintage Model M, with all brand-new parts.
This is assuming that Unicomp is going to continue to produce keyboards, of course. How long have they been out of stock on everything now?
AFAIK Unicomp is out of stock due to tooling overhaul, but I've heard nothing about a lack of parts availability except for the chassis. The problem is that their version of the barrel plate is not a perfect match for the '87-'88, so I'd need a Unicomp for parts plus making some modifications. Really I'd like a barrel plate that's slightly thicker and designed for bolt-in installation (basically turning the melt-rivet points to tunnels) but that would be VERY expensive. It would however, make repairs and modifications a lot easier.
If I ever go to all the expense and trouble to make a keyboard case (chances are somewhere between slim and none)... I'm going with tungsten carbide all the way.
I have always wanted a cermamic keyboard case (and keycaps!) I would probably go for alumina over tungsten carbide, but both are good. Contrary to some opinions, making it would be quite simple. Instead of (say) buying a brick of it and then cutting out the excess (which would likely require diamond machining, not to mention the added material expense) one could simply buy a ton of powder, and use powder processing on it. add solvent, and a binder to the WC powder (I recommend a narrow PSD and two particle sizes tuned for better packing factor) form it, then sinter it. Unlike metal powder sintering ceramic sintering doesn't usually require a vacuum (or nitrogen atmosphere, etc) furnace. The same process would apply to other ceramics ofcourse.
And it's entirely possible to make ceramic keycaps, especially for easier stems like hall-effect or stackpole.
Ceramics of that size are very hard from what I understand. Not necessarily hard as much as
prohibitively expensive. Keycaps on the other hand? Yeah. M two-pieces would be pretty damn easy.
You don't mill tungsten-carbide. You grind it. Though more accurately you don't do either of those things - you pre-form it to the maximum extent possible
before sintering. Blocks of tungten-carbide are actually extremely brittle - incredibly so - and even stay that way after sintering. This is why the bits shatter in interesting ways instead of wearing down. Practical Machinists has this come up as a topic pretty regularly, too. You also don't want to EDM it because EDM will go badly. So will waterjet and laser. You want to cast tungsten carbide - which isn't a bad thing. Might be sand cast possible, which is cheap and easy and do it yourself at home.
But really if you wanted to go full out crazy? You'd demand an aluminum reproduction of a Model M chassis, both halves, at +-0.001". (By the way, the M chassis isn't designed for thermal expansion...)