Competition? What competition? The truth was, we were so ambitious with our project that we didn't finish it in time to join the robot shootout on Monday March 10th. Initially, we divided up the team such that Dave and Lewis would construct the mechanical robot and Jon and Ryan would take care of the electrical and control side of the challenge. We set the deadline for ourselves to finish the mechanical and electrical subsystems by Monday March 3rd, one week before the competition, with rough integration being complete by Wednesday the 5th. However, we didn't hold very well to that schedule. We scrapped the first lasercamm run and made another on Friday the 7th. The electrical circuits also weren't built in final form at that point because we weren't sure where on the robot they were going to fit.
Prior to final integration, the electrical team had tested all individual motors and sensors and was able to utilize them successfully. On Saturday afternoon, Dave and Lewis completed the darn robot at last and the electrical team got their first chance to test out their circuitry. The result was a disaster. Nothing worked! I mean NOTHING! The turrets didn't turn, the catapult couldn't shoot far enough, the wheels wouldn't turn the chassis, it was pathetic.
Jon and Ryan quickly pulled themselves together and started debugging one electrical component at a time while Dave and Lewis were busy throwing things at the wall after realizing that the beautiful robot they constructed was completely brainless. The main difficulties lay in the fact that we tried to integrate everything simultaneously and the circuitry was so complex it was almost impossible to debug. We had constructed a circuit that was supposed to control 6 steppers and 2 DC motors simultaneously. Instead of completing the whole system separately and combining them in the end, we would probably be better off building and integrating each part progressively. At the same time, our original solution was a bit too ambitious given the time span of the assignment and as the Sunday 7 PM deadline got closer, we started to reduce certain functions and simplify the robot. Since the maximum difference in distances to different baskets was about 1 foot, we figured it's not necessary to pin point our exact position and got rid of the arm idea. After that, we found out that the IR sensors on top of the robot were quite accurate given the size of the basket and we gave up the catapult turning platform as well. We simplified our robot so much that our final code was only around 100 lines.
Another big obstacle was power, when we tested out the Maxon motor that powered the catapult, we connected it directly to the power source and the output matched our expectation. This was a big oversight since the power exiting the driver board modules was definitely not the same as the power produced by the power source. As a result, the catapult didn't have enough power to shoot 7 feet. We resolved this problem later by creating a switch device and powered our catapult with two 9-volt battery in series and a 7.5 ohm power resistor.
We finally had a working robot on Wednesday 5:30 AM and we did a demo for Tom at 6:30 PM on Wed. The robot shot 13 balls, made 11 of them. Out of the two shots we missed, one hit the outer rim and the other hit the IR beacon. But our IR sensor worked so well that the direction of ball was dead on every time. We did another demo for Matt at 9 PM and the robot made all 13 shots! Had we finished the robot in time, we would have done well in the shootout.
Some of the important lessons we learned include:
You can see a movie of the GWH in action here: