To be honest, I wouldn't know how to assemble that using Assembly4. I did a classical cardan joint, and you can see that it's not a constant velocity transmission. In this case, it's mathematically exact, so you can even set the deflection angle to 89° and it'll still work. Now, how to do the same with that complex setup .... ? Generally speaking I have experienced that you need to be able to describe mathematically the model for Asm4 to be able to represent it. In this CV joint case, I don't understand how it works, which parts are fixed, which move ...
Now, as with the hexapod, if you accept to "cheat", meaning that you suppose that it is a constant velocity joint, and you impose the exit arm to have the same rotation angle as the input arm, and the middle part to have a deflection angle 1/2 of that of the exit arm (which is given by the gears), it might be easier. But then, you won't be able to check that it is indeed a constant velocity joint.