[Madison]

On the two tread-based designs I've had a hand in, we used two different methods of making certain the belt had as much contact with the floor as possible.

In 2001, 271's robot used a design similar to what Gui mentioned. A teflon coated running board sat just above the belt, always keeping it in contact with the floor. There was one small portion with a wannabe-idler pulley that kept the belt from walking off.

In 2002, 810's used a massive tread system that exerted incredible force to the ground. While the belts were considerably wider, at 6" apiece, the real key was in how they made contact with the floor.

They were deployed, as part of our anti-transmission, using 2, 2" bore pneumatic cylinders. Nearly the entire weight of the robot rested on these cylinders, which transfered that weight across the belts. Because the cylinders had some give to them, they also helped to evenly distribute the weight.

To do this without our anti-transmission, as it was designed, might violate FIRST's 'pneumatics as springs' rules, however, though we never ran into any problems.