Initially, it was a struggle to even get a 4B solve working at all, let alone one that didn't involve very long periods. Many cycles were used trying to awkwardly manoeuvre the output shape in the limited space to get the right bonds. Eventually found this glyph layout which allows for attaching water and projecting silver simultaneously, while the bonders and debonder on the right cleanly rectify the bonds for the output with two pivots. Started with a period 18 1p solve using this layout, and later on found enough conditional processes to convert to 1/3p for a much shorter period.