It is easiest to think of it as the pedal controling the displacement of the variable displacement pump. For each revolution of the input shaft, it can push a variable amount of fluid through the wheel-drive half of the system. At the neutral pedal position, the pump has zero displacement - no fluid is actually pumped in either direction.
As the pedal is depressed, the pump increases its displacement. Assuming constant input RPM's (mechanical power source not lugging-down yet), the amount of fluid per rev increases. Conversely, the torque required on the input shaft for a given output hydraulic pressure also increases. To get maximum output torque on the wheels, the minimum displacement position of the pump will produce the most output pressure for the least input torque from the engine. This is the lowest "gear".
In a perfect HST pump, this holds true all the way down to zero displacement. At you approach that point, you would have max output pressure with zero flow.
In the real world, the pump is not perfect. Some of the fluid slips by the pistons (or wahtever mechanism it uses) and avoids being pumped. At low output flow rates (wheels not turning) and high pressures (trying to push an immovable object), the leakage becomes more than the output flow of the pump. This limits the lowest ratio the pump can operate at for a given RPM. This leads to a maximum in the torque curve similar to the graph posted.
Additionally, there could very well be a pressure relief in the system, limiting the maximum output pressure. Also, there could be pilot operated valves where a minimum flow is required before full pressure is developed or some other mechanism as was described to help produce a dead zone so as to prevent the tractor from creeping when the pedal is released.
- Rick