Finalize design concept for ball capture, release & score

[rhauch]

At the Jan 23 meeting, we evaluated several designs for:

• capturing the ball - getting control of the ball on the ground and pulling it into the robot
• releasing the ball - moving the ball from inside the robot (after capture or catch) onto the floor and propelling it forward (e.g., passing or expelling)
• scoring the ball in the low goal - moving the ball from inside the robot (after capture or catch) into the low goal

Concepts included:

1. Side-mounted arms that rotated inwards, with a roller mechanism on one arm that pulls the ball into the robot. Still likely requires a separate ramp (mostly for scoring). Still separate propelling mechanism required.
2. Single drop-down capture arm (hinged low) with roller mechanism that pulls the ball into the robot. Still likely requires a separate ramp (mostly for scoring). Still separate propelling mechanism required.
3. Various forks that could pick up a ball on the floor and raise it into the robot, and for scoring. Still separate propelling mechanism required.
4. Some sort of mechanism similar to the Boom Done design (see the video, photo, CAD drawings, and a capabilities document.
5. Single drop-down capture arm (hinged high) with roller mechanism that pulls the ball into the robot. The high-hinge means that the arm could actually serve as a door (for blocking), with the roller nicely on the interior of the robot when the arm is in the down (stored) position. (When in this lower position, the arm is wide enough to sit in front of the superstructure, providing a physical support in case of a force pushing on the arm; e.g., blocking.) No ramp is required, since we can just drive up to the goal for scoring, and we can capture the ball on the floor (and optionally move it into the robot). We talked about different ways to release the ball:
   1. Position the arm so that when in the lowest position the roller is under the ball. The roller can can push the ball out.
   2. Position the arm so that when in the lowest position the roller is in front of the ball, and use roller(s) on the interior floor of the robot to push the ball out (after the arm is raised).
   3. Position the arm so that when in the lowest position the roller is in front of the ball, and simply slope the interior floor of the robot to let the ball roll out onto the floor after the arm is raised enough. The roller can also propel the ball forward once it's on the floor. (The sloped interior floor also gives us more volume for mechanics.)

The last design (5.iii) seemed to hold the most promise, largely because it needs no ramp or internal forks/rollers and thus is very simple. Also, because we really only need two positions for the arm (up and down; the ball can be captured by moving down), we can use two pneumatic cylinders (one on each side of the arm, just below the arm's hinge line that is just above the opening of the robot). The cylinders would be compact enough to be out of the way (after catching, the ball needs to fall down to the interior floor).

Now that we're again considering pneumatics, we're going to consider using them for extending the net arms. Still concerned about the amount of rotation; we'll probably need about 135º, which we should be able to get with some careful design.

[rhauch]

One consequence of most of the designs (including 5.iii) described above is that the arm and capture opening may need to be on the front side of the chassis, so that we can drive forward to grab a ball that is just out of reach of the roller. IOW, strafe mode might not be ideal for capturing.

Luckily our superstructure design and square chassis make it easy to rotate the superstructure on the chassis. We can actually try both modes to see which works. We might even be able to switch for a particular match and alliance, depending upon how much time is available.

Also, our superstructure design is indeed flexible -- all of the above capture designs are all feasible with it.

(This robot design would be great with a swerve or mecanum drive.)

[rhauch]

We should consider calling this robot "Chewbacca". It's going to be pretty big.

[rhauch]

This is a sketch of design 5.iii:

Who is going to CAD it up?

[rhauch]

ISSUE: We need to check that we can score in the low goal with this configuration. The arm is just narrow enough to fit within the goal box, but when extended it might hit the upper edge of the goal. (We need to CAD up the design to see for sure, as well as to size the length of the arm.)

Worst case we might be able to drop it on the floor, back up a bit, drop the arm, and lift it into the goal.

[rhauch]

See #34 for the CAD task.

[rhauch]

We pretty much have finalized the design to be mostly like the above drawing. There are a few minor differences (more 80/20 supports for net arms and motors and for roller arm cylinders), but the biggest difference is that the roller arm is hinged at the bottom rather than the top.

There are several reasons for this change. The first is simple geometry: to capture the ball setting on the floor, a top-hinged arm must move down slightly, then up (as the ball hits the bumpers and starts rotating up), and then down as the ball goes into the robot structure. This is really complicated movement that we can't really effectively control. By moving the hinge to the bottom, the roller arm can drop down (via pneumatic pressure w/ gravity assist) around the ball (to a maximum distance from perimeter of 20") and, as the roller is turned on, the ball will move through the arm and end up resting on the arm that is roughly in the same position. Then as the arm is raised, the ball is lifted into the robot. The roller also keeps the ball inside the robot.

Secondly, a bottom-hinged arm keeps the ball in the robot: the distance between the bottom of the roller and the arm (at the hinge) is just smaller than the ball diameter -- big enough to squeeze the ball through, but small enough that the ball will not fit on its own. When the ball is inside the robot (after capturing or catching) and the arm is in the up position, to expel, pass, or even score the ball (into the low goal) we simply start the roller in the opposite direction: the ball is squeezed and expelled at a rate proportional to the roller speed.

Thirdly, the same arm design that keeps the ball inside the robot will keep balls out. And even if a ball were to squeeze in unexpectedly, all we have to do is expel it by turning the roller on.

Fourthly, we don't need a ramp or additional structure. All we need is the arm, so this simplifies the design.