Spigor, as you say, the power from the motor comes to a fork in the road and takes it. Most of it goes to the low-speed shaft of the gearbox driving the planet carrier. The rest of it goes to the hydrodynamic coupling, or speed variator, and then to the high-speed shaft and sun gear. The gearbox combines these two power inputs into the total power output which appears at the gear case, or cage, and then at the output shaft of the unit after passing through the parallel shaft gears, the ratio of which is fixed.
To determine torques we need only the fundamental ratio of the gearbox. By fundamental I mean when we observe the gearbox in isolation, with the cage fixed. In the Vorecon illustration, the sun gear and planet gears look to be about the same diameter. If they are the same, and because the gearbox increases speed, then the fundamental gearbox speed ratio is 1:4, going from input to output. The torque ratio is the inverse, 4:1.
Keeping things simple, for a torque of 4 on the low-speed input shaft of the gearbox, the torque on the sun gear and high-speed shaft is 1. Then the torque on the cage must be the difference, so it is 3. Torque direction is by inspection.
It is critical to understand that the ratios of these torques are not affected by any changes in the speeds of the various elements. The magnitudes of the torques will change as the torque imposed by the load changes, but the ratios of the torques are fixed by the fundamental ratio of the gearbox, which does not change. That is the difference between the Vorecon gearbox, distinguished from the overall unit, and one in which the speed and torque ratios can change, such as a three-speed gearbox in an automobile. Yes, I am an old fogy. The fact that these ratios do change in the overall Vorecon unit causes confusion on this point.
If the prime mover is an induction motor running directly across the line, then the motor speed is fixed. This is fine and good. The purpose of the Vorecon is to eliminate the need for a variable speed prime mover such as a VFD. But the speed of an induction motor would not be exactly fixed. It would change a little as the load changes. It would be fixed in a synchronous motor, but enough quibbling.
The fixed speed of the motor sets the speed of the low-speed shaft of the gearbox driving the planet carrier. The speed variator varies the speed of the high-speed shaft driving the sun gear. The combination of these speeds sets the output speed of the gearbox, which appears at the cage. The output speed of the unit is set by the combined speed at the cage and the final stage in the drive train, the parallel shaft gears, which increases the speed of the cage at a fixed ratio.
During a fixed speed ratio for the Vorecon and a fixed load, the speed variator does not increase the torque on the sun gear nor does it extract any. It only meets the torque that is appearing at the sun gear. Otherwise it would be accelerating or decelerating the unit.
But the Vorecon must be able to accelerate or decelerate. Then the speed variator does increase or extract torque from the sun gear, although it is only partially responsible for the power increase or sink. Once the new output speed is reached and the new load is fixed, the speed variator goes back to only meeting the torque on the sun gear. But even during acceleration or deceleration, the torque ratios in the gearbox do not change.
I suppose there could be applications in which accelerations and decelerations occur almost continuously. In fact Voith designed a unit for those conditions. In principle it is the same as the Vorecon, except that it has a more sophisticated means for controlling the speed variator. See the Voith WinDrive for wind turbines.
Going back to the beginning and looking at the load split in the opposite direction, the load on the motor is the combination of the load on the low-speed input shaft of the gearbox and the load on the speed variator.