The idea is to lock the inner tumblers in place tightly enough that they do not come in contact with the outer tumbler elements after blocking (the orange sidebar allows less play than the yellow peg). Having the orange sidebar contact and transmit rotation to the outer plug also adds some variability to the max turning angle. so that measuring the precise angle of the keyway will not provide useful feedback. This is because the max turning angle is dependent on not only the depth of the main sidebar but also on the depth of the blocking sidebar.
This design however is not optimal. I think that it could be made simpler. I will post an image soon but before then will describe it in detail. I propose taking a normal disk tumbler design, a plug with disks and a sidebar , and adding a second plug within that. the outer disks have internal gear teeth that mesh loosely with external teeth on disks set, off center in the internal plug. The internal plug is shaped somewhat like a crescent moon. A blocking sidebar runs alongside these internal disk in a slot milled in the internal plug and is moved back and forth in this plug. the slot milled for it has tilted sections that cause it to move into the internal disks when pushed towards the front of the lock. Transmission of rotation from the internal to the external plug is accomplished using the blocking sidebar as in the original design. Said sidebar is L shaped and a groove is milled into the side of the lock cylinder in which it slides. this might be used to add core change functionality to the lock. the external plug will also be driven by said l shaped sidebar as in the original design. The blocking sidebar provides a weak point and will fail before the main sidebar rendering the lock broken but secure if the attacker tries brute torque. As an added feature to resist brute torque without damage the blocking sidebar could be spring loaded so that when excessive torque is applied it rotates slightly deforming a relatively strong spring. The inner plug , and outer disks attached to it, rotate but the outer plug stays in place. Aditionally because this design uses only a small amount of the outer disk's surface a master ring system might be possible. This might require modifications to the blocking sidebar torque transfer system by would enable an SFIC core change system or added resistance to master key decoding in a master keyed systems by eliminating ghost keys. I will post an image of said design ASAP.
