:smoggers: (ok maybe say it)

Explanation of how a 16 is possible:
You need 18 air, so 18 inputs.  After you've pulled 17 inputs that's 17 air and 51 qs.  You've already outputted 5 outputs (15 air, 5 qs) so you're left with 2 air and 46 qs.  Each lead can hold 5 qs, so that's 2 air, 1 qs, and 9 gold on the board, for 12 total atoms.  Then the 18th input spawns which makes 14 atoms on the board.  To put the extra qs into a lead you'd need to let the lead spawn, so temporarily there'd be 17 atoms on the board.  So, for a 16 there's no projection after the last input is spawned.

Tricks needed:
You can move atoms in a loop (and full debond is possible with this, as I show).  Because the glyph area is just 15 you can keep an atom on track, allowing for a back-and-forth shift which you see a lot of in my solution after the last input spawns.  You need both these tools for the last move because the final product must shift off the bonder in one move (if it didn't, a gold would bond to it) and there's only enough room for a loop on one side.

Why this layout:
I had a symmetric 16 which was pretty but slower.  This glyph layout doesn't stick an arm in the middle of things, is easy to feed a new lead into, and after debonding the air atom can pretty easily find its way to the bonder.  The bottom track loops speed up common tasks.  The perimeter track has some weird kinks so that the perimeter arms can shift around in either direction sometimes, allowing for some perimeter arms to be used in those moves.

Comments on the puzzle:
I haven't played in two years, maybe, but I'd be surprised if more than 10 people got min Amarms primary.  I think every trick I described above is necessary for a 16 solution.  My guess is that no playtesters got 16, but thought it was probably possible.