Unless I am mistaken, when you constrain the 2nd connecting rod, the constraints become superabundant ... and the system hangs!
It would be necessary to remove the left coupled wheel (eg) from its kinematic equivalence class and then put it alone in a sub-assembly. This would then have to be constrained with the shaft of the “coupled wheels” sub-assembly by a “PlaneCoincident” constraint which would allow it to rotate.
This is correct. Rotation was the only remaining DOF, so when you lock the angle to drive it, then DOF = 0.
If I remember correctly, every time you change the value of the angle; it is necessary to recalculate to obtain the displacement. Be careful, the angular displacement must be done in small steps (around 5 °, sometimes a little more also works).
Edit : I just tested. Just press Enter.