R is 735. So is I.
This could be given a 6P treatment to save four instructions on each loop, but I'm not doing that to save 20 on tertiary (and I have a much better rate solve).

Solve entirely created in stock OM except that at one point I realized I was building it backwards and used FTSIGCTU to flip the machine over. Then I went back to unmodded OM. Title edited in emacs.

Anyway.

This is really more TG than A, but whatever - it doesn't have waste, so the area can't be that bad, right?

The most annoying part of this solve is those last two Fe-Fe bonds. I had to rip out a bunch of instructions when I discovered they would get debonded when I made the final two Hg-Hg bonds. Ugh.


EDIT: I have realized that this is non-optimal - I could do better by swapping to the tin amalgam, at least at tertiary; I doubt I'd end up with a better primary, and I can't improve my secondary.