Florian -- You are absolutely right! I did not believe it until I modified my file to use a free rotating trunion joint at the top of the armature. I have used only constraints that exist in the real world where where parts are constrained by physical contact with adjacent parts, where pins line up with bushings and pistons are free to rotate in the bore. There are no "artificial" constraints that are relative to construction planes or axis. Your method does provide control of the armature and keep the piston rod in a single plain. Normal trunion joint are easy to make but a rotating one is not so easy.
One thing worth noting is that this method of restraint works only on single cylinder or double cylinder designs when the second cylinder is at 180 degrees from the first. Three, four or more cylinder designs would have to have spherical joints at both ends of the connecting rod on all cylinders that are not set at 180 degrees as they will all exhibit the figure 8 motion at their rod ends.
[youtube1]https://youtu.be/70ODsZQP3FE[/youtube1]
Allen -- The side forces in this type of mechanism are not as heavy as you might expect within a range of angles of the "Z" part of the shaft. Offset angles between 20 and 40 degrees seem to be the most efficient. Greater or smaller angles increase the forces and obviously 90 or zero degrees don't work at all. I have built and run quite a few "Z" or wobble plate engines with as many as six cylinders using only brass or cast iron bushing on the shaft and standard ball bearings for the rotor or armature. I'll post a few if you haven't already seen them.