Eureka!

Please bear with me.

We are working on motion distribution NOT torque distribution.
We should be concentrating on torque distribution via the electromechanical clutch.

There is no axle disconnect on the VX, and the transfer case has no allowance for any 'freewheeling' of the front axle. The design of the transfer case is such that the front axle (and wheels) are ALWAYS driven at the same speed as the rear driving wheels. Rotational speed is NOT torque. Torque only exists when there is resistance to this rotational movement. This resistance is applied by two things: the electromechanical clutch, and the ball/ramp mechanism (see figure 3 and 4, as well as page 16-17 of the patent). The ball/ramp design of the coupling mechanism is NOT self locking UNLESS there is a control input. Resistance to movement (at the front wheels) provides this control input just as well as activation of the e-clutch.

This explains Tone's snapped driveshaft. The solenoid was deactivated, yet the speed differential between the front and rear axle was great enough to allow the ball/ramp clutch to force the rotor away from the connecting sleeve -this action transferred (almost) 100% of the available torque to the front axle and BOOM. Tone, did your dyno graph suddenly fall off just before your driveshaft went?

The E-clutch provides a means of modulating torque by applying measured resistance to motion according to a predetermined pattern.

Assuming that I haven't completely lost my mind, there is NO WAY that we will ever have true 2wd without an axle disconnect feature.
So, that leaves us with modulation of the e-clutch's holding power via PWM. This can be done only as long as both driveshafts are turning. If the front driveshaft should stop suddenly (like it is dropped into a hole under power), expect the ball/ramp clutch to do its work, regardless of what setting the manual TOD switch is in.