Quasi-open System, or Dual Closed System (Thermodynamics Addendum?)

[philandy]

Hypothesis: If a closed energy system interacts with only one other closed system, there can be positive work.

My test uses a hollow ball, less dense than water but denser than air/vacuum. A second test using a ball less dense than air but denser than a vacuum. It also uses dual vertical tubes (columns) of the previously specified controlled density environments. The chambers are connected, valved, suction controlled, and osmosis resetting (meaning if an environment splashes it can be put back or replaced within tolerance). The test must also be noble and shown to work with materials that show little effect with outside influence, for example an iron ball would be movable with a magnet.

On the "left" or first column where such a ball is introduced into the environment at the bottom, it would float, hit a mechanical switch controlling the valves appropriately, hit a mechanical switch to osmosis reset, buoy, then propel above the environment and transition through the above valve into the "right" or last column. The ball would then fall, hit a mechanical switch controlling the valves appropriately, hit a mechanical switch to osmosis reset, settle, then propel below the environment through the below valve into the "left" or first column. A cycle has thus happened and beyond polish maintenance it will recycle.

Height controls total kinetic energy.

The problems are costs associated with osmosis resetting, valve sealing, and polish maintenance.

Would anyone else help me explore this?