The Paradox Engine

[Tusk]

The drive motor goes On and Off making the assembly to cycle:
When its On, generators produce  power to cover most of the drive motor requirements
When Off - generators keep giving and the energy is being stored in the battery,
which in turn is connected to the drive motor

Only one major change I would make to that telecom:

You have the generators 'on' all the time, whereas I intended they be open circuit or 'off' during the power half of each cycle. Otherwise the resistance from the generators would interfere with the creation of the secondary reaction. There is no loss incurred by simply allowing a mass to 'spool up' since we can recover the KE after turning the power off. And in this instance because we will probably be getting a second free rotation (or near to it) for every induced rotation of the disks (in the FoR of the rotor arm) it is more than a little advantageous to do so; providing we recover the energy in the same FoR without significantly reducing rotor arm motion, which we can achieve by this method.

I would have liked to develop a system which runs at a constant velocity/rate but realistically with this asymmetric phenomenon it should be no surprise that the device needs to be cyclic. 

[telecom]

I think I understand exactly what you are trying to achieve, and as somebody liked to say:
"Practice is the final criteria of the truth"
So we just have to wait until someone will invest some time and, perhaps, money
into building this machine to find out if it actually works as you envisioned.
Regards.

[Tusk]

we just have to wait until someone will invest some time and, perhaps, money
into building this machine to find out if it actually works as you envisioned

Belay that telecom, I just found an error in the design of this 'quick and dirty' version; I'm blaming you, always trying to simplify the build lol (just kidding, entirely my mistake).

You got me thinking about torque reactions from a conventional point of view, just as well; I missed something important with all this 'rough and ready' build simplification. Unfortunately we can't mitigate adverse torque effects on the rotor arm by recovering E with generators at the disk axes; it looks like the rotor arm E recovery (both forward and reverse) is the only way to go, so I'm falling back on my original design. The dynamics are a little convoluted but the physics is sound. The aforementioned 'simple build' would just erode rotor arm motion with the result that the additional rotation of the disks would be lost in the recovery half of the cycle.

Not surprising really, if it was that easy someone would have already stumbled on it by chance. But I think that throwing these alternate ideas around doesn't hurt, provided we don't forget to check our sums before changing direction.

The secondary motion manifest in the rotation of the rotor arm is key; best to let it 'do it's thing' and recover the E as originally described, then recover from the disk/s which induces a second (reverse) rotor arm rotation, which is then recovered to finally end the cycle ready to go around again.  I would definitely go for the twin disks though, no point in wasting energy accelerating a counterweight.

[telecom]

Hi Tusk,
in this case, can we replace the E drive with the DC motor/generator, located at the same spot,
but transmitting the torque by the gear drive?
It is much more convenient to get an off the shelf component rather than a self-made one.

[Tusk]

can we replace the E drive with the DC motor/generator, located at the same spot,
but transmitting the torque by the gear drive?

I have been considering this option for some time and can uncover no serious problems with it. I would opt for a drive wheel of similar diameter to that of the motor's rotor, and once again recommend a twin disk arrangement.  Some sort of clutch mechanism might be advisable, to avoid drag on the disks during recovery of rotor arm energy.

With both a generator and motor situated at system centre the engineering challenge increases; but not beyond precedent, we could assume.