Update: The Calutron

     My -- admittedly cursory -- reading had left me with the impression that nobody bothered with mass spectrometry to produce fissionables these days.  The process was still good for turning out research-level amounts of obscure isotopes, and maybe some medical ones, and that was it.  Real baddies went in for centrifugal separation, since it was relatively simple* and produced the greatest volume for a given amount of hardware and energy to run the system.

     Yeah, guess what?  Wrong.  At least one rogue state homegrown member of the nuke club picked up bargain-priced calutrons, apparently of Chinese origin, and used them to crank out some measurable quantity of hot stuff.

     So there's still a degree of "security through obscurity" for these tricky widgets, and my curiosity will have to go unsatisfied.  Any would-be start-up is going to have to work that stuff out the hard way, which takes time and leaves traces.
_____________________
* None of this is "Khyber Pass gunsmith" simple, thank heavens, requiring specialized, high-precision technology and skilled technicians working with chemical and radiological hazards that would curdle your hair -- in some cases, literally.  Even at that level, there are degrees of difficulty, and some processes are more efficient than others.  But they're all insanely inefficient.  The Manhattan Project tried every method they could manage (you'd never guess which one they couldn't make work) and ended up running them in sequence, gaseous and thermal diffusion plants feeding (via chemical processes of hair-raising danger) the calutrons, a system of vast industrial works employing tens of thousands of people that had, by the end of the war, produced slightly more than enough material for four bombs; they'd used three and didn't have the fourth ready to go.  This would be an amount of fissionable material that a moderately-good weightlifter could raise overhead.  It's a wheelbarrow load.