The lead input assembly can produce tin in 5 cycles, or output lead and tin in 8 cycles.
The quicksilver input assembly can output the three higher metals in 26 cycles from 3 tin. It can switch modes in 11 cycles (skip to the start of arm 2s instructions to see this happen). In the other mode, it can output 1 iron from 1 tin every 5 cycles. 
Using the central purifier, the tin purifijection can be balanced such that both inputs have no downtime.
Making the solution and period longer makes the 11 cycle cost of switching modes less significant.
The theoretical minimum rate for this design is 29.5r. Doing the algebra, i found the input periods to align every 590 cycles, starting at p94, 31.33r. I thought that to demonstrate the design properly, i needed a longer period than that, so i chose the next lowest option, period 684, 29.74r.

One interesting bit of tech here is the 5r quicksilver transport chain for the iron projection mode. Normally, single atom transport can at best happen at 6r in 2w. This is because you usually want to gain advantage on the atom, which the 5r mode cannot do. Take arms 20 to 22. These grab something behind them and gain advantage on it so as to deposit in front of them, enableing arm 23. In the case of the quicksilver mode switch however, the purpose of transport is to cover distance, not to gain advantage, so the 5r mode can be used.