I'd imagine any combination of {duplication and mutation on the blue cone gene which isn't on X, mosaicism or Klinefelter syndrome, transposition from X or repl

I'd imagine any combination of {duplication and mutation on the blue cone gene which isn't on X, mosaicism or Klinefelter syndrome, transposition from X or replication on X with mutation} would work for you. Maybe not Klinefelter because of X-inactivation, I don't know how those two things interact offhand. I _think_ the key point is, as you said, that the different gene copies need to have different response ranges, or you end up functionally trichromatic even though the pigments are from different places.