3/12/14: Bonus Meeting
Things we did/Why we did it: Modified the tower rollers in hopes to grab the large balls at full speed; Tried drilling into sprockets to make a chain bar on the funnels system. Uploaded the standard middle zone autonomous with encoder values and tuned it for better reliability and consistency. Apply preset encoder heights. Applied an elastic backing for the 3rd bucky ball
Problems: Sometimes the robot would get stuck into an infinite loop of the code. Conflictions occur with the funnels. The preset lift heights for pushing the large balls were too low.
Results: The 18 point autonomous has consistent paths, increasing reliability and aims for the large balls to speed up human interaction. The robot can grab large balls without driving slowly. The robot now has preset lifting heights. The large ball storage is sturdier
Misc. Notes: Changing logic from “else”s to “if else”s can cause the code to get stuck in a loop. We realized that the 3rd bucky ball can pose problems in the new autonomous
Future Plans: We need to finish up work with the funnels: make sure they fit in dimension, deploy appropriately, and grab floor buckies. In addition to this, we need to develop the new autonomous, further tune the bucky ball intake, practice driving, and retension the lift, and reinforce the the lift.
Today was a preparatory meeting for FRC. However, we were able to squeeze some vex time. Because of this, our goals were quite scattered. However, we were motivated with our work, and got a lot of things done.
First off, our programmer returned the robot to the shop. During break, he rewrote the middlezone autonomous we used in states with encoder values. Thanks to this, the autonomous is much more consistent in its paths. Before he could start coding, he had to finish up mechanical changes. He reattached the drive train motors, rewired the robot, and mounted encoders onto the lift and funnels system. We decided to go with encoders simply because we do not have potentiometers. In addition to this, he attached a bracing below the drivetrain to increase stability.
At the start of the meeting, we developed a new set of rollers to grab the large balls. We realized that the tank tread on UVM’s rollers allowed the intake to dig into the large ball. So, we decided to mimic this effect with additional rubber bands. Instead of 4 rubber bands padded out like last time, we added 8 rubber bands which crossed over each other, creating a large bulge about a quarter inch deep. We only replaced 1 side of the intake and we could immediately grab without a wall at full speed. We replaced the other side just for consistency. We then filmed the robot while grabbing. Looking at the footage, we noticed that the robot grabs the large balls well, with very little wheel slip. Because of this, we believe we aren’t grabbing as quickly as UVM because our intake isn’t geared for speed.
After this our programmer took the robot to adjust the code. We had numerous disconnection issues, so we hope the vexnet 2.0s coming in will solve our problems. When it did run, distances were astonishingly consistent. As for mistakes, we realized that the funnels did not close enough. Because of this, the funnels would run into the wall. To fix this, we decided to retract the funnels slightly. Using the extra time, the programmer decided to add a sweeping action. Rather than manually turning the robot, the robot would turn, facing the large ball. This would shave the time needed for repositioning. All that the human player would do is realign slightly and trigger the limit switch to push the large ball. When we got to the large ball sequences, we noticed that the lift was very low, which increased the chances of running into the barrier. So he then raised the encoder values slightly to increase the arm height, adding more tolerance.
As the end of the meeting approached, we made very minute changes. For one, we doubled up the zipties on our large ball storage. We manually put a large ball and found out it supported the large ball with moderately violent pushing. However, when we grabbed a large ball, and did a test where we spun the drivetrain around violently, the ball came off. We realized that elongating our storage system led to moving back the ziptie supports. Because of this, the zipties do not fully cover the first large ball, making it susceptible to falling out.