The Paradox Engine

[Tusk]

In this case we are getting the inertia B working for us for free? And it is equal 1/2
of the applied force?

Yes you are correct but as always there is more to it telecom, inertia B is half the applied force but it is the combined effort of inertia B and the 'spare' half of the applied force which work together not unlike two equal weights each side of a set of scales, their combined weight double that of each. In this way the secondary motion is equivalent to the primary, each motivated by the applied force or the equal of it. The total 'mass in motion' or momentum is twice what you might expect if you were unaware that simply shifting your point of applied force could cause such additional motion yet not be in breach of CoM.

[telecom]

So,
in relation to your apparatus, this lever with two equal masses represents a disc,
and the centre of mass represents the point of attachment to the rotating arm.
And the applied force represents an EM drive?

[Tusk]

Correct again telecom, I took a few liberties in the interest of simplification but in essence they are very similar thanks to the symmetries in the system; a curious serendipity, since what there is so far of the greater hypothesis has the asymmetry of the universe at it's core. But the PE apparatus challenges us sufficiently without the additional burden of the problem of grand unification 

[telecom]

It appears that this effect of the inertia becoming a driving force only shows up
during the acceleration, since when the rotational speed becomes constant, the inertia
stops playing the role as a force, it only acts as a mass.
I presume that the activating force will become close to 0.

[Tusk]

It appears that this effect of the inertia becoming a driving force only shows up
during the acceleration

If you also meant deceleration (which I believe is the common intent when referring to acceleration) then yes telecom, but I thought it best to make that clear since the device is cyclic and inertia works for us both ways. I have neglected to investigate if it might be possible to replace the force of inertia with our own resistance (e.g. driving a generator simultaneously) having been more concerned with promoting and explaining the basic concept; no shortage of engineers and inventors capable of taking the idea further so it seemed wise to run with it 'as is' in the hope of an ASAP result.

Inertia does provide us with a handy resistance 'bench mounted' to the universe so to speak. If we do substitute our own resistance the problem of where to mount it arises, but I would not be surprised to see a viable solution, the idea of an outer circle of coils was my first thought (bench mounted of course) but such things are better left to the professionals 

Lately I have been more curious about the disk energy issue. If we allow the condition where the disk rotates once in the observer FoR for each rotor arm rotation (which can occur with a suitable arrangement) then in the FoR of the rotor arm the disk rotates twice for each rotation of the arm.

Since Ek = ½ mv² this implies we might somehow recover the energy spent motivating the disk X4 but for the problem of rotor arm reversal which if allowed must reduce the rate back to 1 (per arm rotation). I am hesitant to trot out new and untested ideas due to the barely credible nature of the material presented thus far. But here I think is a promising line of investigation, I have a few potential 'quick draw' solutions but this deserves serious unhurried contemplation.