Let's go back to the overall picture we were discussing earlier:
Step by step the following happens (well in simple terms. You can check the patent for detail!)
1. Tip of 11 sits on 21
2. 42 hits 22 on its anticlockwise swing.
3. 32 rotates to block 22.
4. 42 returns clockwise to hit 32.
5. 32 drops into 22.
6. 10 drives 20 and 22 around.
7. Impulse from 11 to 41.
Back to 1. Tip of 11 sits on 21.
What is clear is that impulse is only imparted to the balance wheel once per cycle. (11 onto 41).
I’m not sure but it may also not be self starting; by its nature it locks the wheel (11) until released.
What I like is that the impulse is imparted somewhat radially (like a coaxial) from 11 to 41 – but it could possibly be improved. There is also somewhat radial movement when 42 goes clockwise and hits 32.
What I don’t like is the sliding contact between 11 and 21; 32 also slides across 22. Also 42 (anticlockwise) and 22 has apparent sliding but it might be improved by careful geometry optimisation. The degree of importance of the sliding will depend on the forces involved. These could be improved by geometry development; and mesh timing of the gears.
At this point it must be remembered that the patent is not a toolkit to allow others to make the system in its ideal form. It is only describing the principle of operation sufficient to gain protection. Thus I am sure there are developments and details to work out that would improve the system compared to its description here that will remain secret within Patek for the moment.