I made an adapter so I could try this in my existing frame,, this frame has no feedback system so it is a test between the input drive gear and the pivot shaft.
My system has new bearings still filled with grease so they are stiff and I have some friction within the system so there is some force that will transfer.
First test,, spin the drive gear, the upper gear turns, the lower gear turns and the whole system turns,, very little effort to spin.
Second test, I stuck my finger in and stopped the lower gear from turning,, I then spun the input drive gear and it was a little stiffer to turn and the upper gear rotated but nothing else turned.
Third test, I stuck a TT motor on the pivot shaft, it can spin but has a force required, the input drive gear felt the same as the finger test and the only thing that moved was the upper gears.
Fourth test, I used my finger to let the pivot shaft "slip" past my finger with varying amount of force,, the input drive gear changed with the amount I was pushing on the pivot shaft but only went up to the point where it was the same as the stopped test.
It would appear that the physical distance relationship has been decoupled and the "magnetic coupling" then would be a constraint force only,, that coupling will be 3 TT motors running in series and across a 50 ohm resistor.
