Workaround works!
Though I figured out that it is enough to use just two cubes for defining the workingfield of physics. The two cubes are positioned at the start and the end of the longest diagonal through the workingfield-cube.

Cheers, chrisu.