The top half of this machine came together rather quickly.  The left side uses purification to turn 2 rust into clean copper sulfate minus one water.  The right side uses projection to turn 2 rust and 1 reactive water into dirty copper sulfate with one water excess.  It's a match made in heaven!  Arrange a simple interface that moves the spare water from one machine to the other, and we're off to the races.

...But that machine is only Rate 4.  Optimum is 10/3.  It is pulling one of the rust inputs at only half speed.

In order to optimize the rate, not only do we have to pull those missing rusts, we also need to cannibalize one of the dirty cooper sulfates without interrupting the steady transfer of water singletons. In the end, we need the odd-out machinery to create 1 clean and 2 dirty copper sulfates via purification, all which must be constructed slightly differently from one another. 

Bring on the duct tape and zip ties!

The first working version of this solve lineage weighed in at 895g.  Most notable savings came from doing all the odd-out construction on a single multibonder, and the quirky idea of letting one of the odd-out quicksilvers pass through the steady-state machine to be plucked from the output (this made it much simpler to maintain the water transfer).  

But my favorite 10g savings was finding a way to reuse the steady-state calcifier.  Check out the ridiculous movement pattern on arm 16, delivering the odd-out clean cooper sulfate.



