Build Season
After our initial strategy and brainstorming sessions, we placed shooting as our number 2 priority below drive, thus most of our prototyping/testing went towards the shooter. We chose to use a wheeled shooter design based on our own experiences with them, as well as their proven success in FRC as a whole. From the beginning we knew a lot of backspin would be important, and initially used
HOT’s 2006 robot as an inspiration.
Based on this inspiration, we built our first prototype with a long hood. We wanted to minimize the amount of rotational velocity our wheel lost with each shot. To do this we needed a large wheel, with most of its weight distributed at the edge. For packaging, 8” was as large as we were comfortable with. The 8” KoP wheel met these requirements and we had a lot of them in our shop, so we started prototyping with them. Our initial testing proved we were heading in the right direction, so we continued to refine that design.
Basically, all of our prototypes looked like this in some form.
Even after minimal testing we found the tread on the KoP wheels would slip which caused less consistent and less powerful shots. We then tried a slew of options available in our shop, from high voltage electrical tape to the common conveyor treads used on wheels. Eventually we settled on using the paint on urethane we used in 2011 on our rollers. This gave a smooth, consistent surface, with considerable more traction, and remained attached at high speeds (if applied correctly).
As we moved towards our final prototype, we observed that the ball was not using the full length of the hood. As a result, we shorted it down so it was just longer than the exit. This ended up being the length we used on the final design. In the early stages of prototyping, our designs were not well built, or easy to adjust which lead to inconsistent results. In addition, we wanted to be able to shoot from both the fender and the key, and testing proved simply adjusting the shoot speed did not give us enough range to be able to accomplish this task; we needed to adjust our hood angle. With each iteration, we added more adjustment (in both hood angle and compression) and used more precision in fabrication and assembly. Our final design incorporated pneumatic cylinders, providing two hood angles, one for key shots and one for fender shots. While our final design also had adjustable compression, we found about 2” of ball compression was ideal, and had it set there for most of the season.
Final Prototype
We never really questioned the need for a turret, as we assumed it to be essential for targeting quickly. As well, a turreted shooter allowed us to have a forward facing shooter for driver control and a rear facing shooter for autonomous so we could retrieve balls from the bridge without turning the robot around.
The other aspect, along with backspin, that we thought was key to consistency was feeding the balls into the shooter itself. For this, not much prototyping was done, but we decided to load the shooter with a pneumatic cylinder directly below the ball. The advantage here is that the air pressure is regulated so the force would be consistent (compared to the variability of a motor) and applied to in the same location on the ball relative to the shooter wheel regardless of where the turret was facing.
Due to space constraints and the amount of stroke we required, a telescoping pneumatic cylinder was needed. After completing minimal research, we abandoned the telescoping cylinder approach and attached two standard cylinders together with the rods extending from opposite ends of the assembly. We used a cup (“the flower” as we called it) on the end of the cylinder rod to push the balls into the shooter. We surrounded the ball with pink polystyrene insulation, with a surface of thin polycarbonate (~.01”) and Teflon.
Pokey Pokey (the shooter loader)
The shooter was initially designed to use 2 BaneBots RS-550 motors as they are the most powerful motors available other than the 775’s (which we no longer use after many issues with them) and the CIM's. Prototyping proved we needed a maximum wheel velocity of about 5000 rpm to achieve the range of shots we desired. With the 550 has a free speed of 19300, we chose to reduce the speed by a ratio of about 4:1 using gears from the Fisher-Price transmission.
Approaching ship day, we were overweight. The second motor on the shooter was an easy couple of pounds between the motor, the Victor, and wiring. After removing the second motor, we found that the shooter still performed well, but lengthened our spin-up time by about 1 s.
Simbot Jordan's Shooter
Greater Toronto East Regional
We were unimpressed with our shooting at this event. We were shooting 50% in drive control, far from the 90+% we desired. No design changes were made to the shooter at this event, as we gave any and all free time to our programmers to continuously improve our software (mainly tune our PID constants).
Waterloo Regional
In an attempt to save weight at this event (we were preparing to add a dingus for balancing for our next event), we tried switching our 1/8” polycarbonate hood for 1/16”. In theory, we thought this would not only reduce our weight, but it could possibly improve our shooting as well. The theory was that the polycarb would flex differently for different ball densities, allowing variable compression and more consistent shots. On the field during a practice match, we found that the new hood completely failed, as balls flexed the hood too much, and basically dribbled out of the shooter. The hood was promptly
switched back for the rest of our practice matches. With the PID constants more tuned, and an extra two weeks of driver practice, our shooting percentage increased to 76%.
Greater Toronto West Regional
Before this event we finally went through with attempting to linearize the Victor output. This gave a huge improvement to the stability of our shooter speed.
Now preparing to add the dingus to our robot, there was still some weight we needed to remove. We removed the adjustability from our hood angle, by replacing the pneumatic cylinders with plastic rod. This choice was obvious for us as our fender shooting was significantly worse than our key shooting so we rarely used the fender. We noticed a slight improvement in consistency, which we attribute to the pneumatic cylinders giving a little when a ball is shot, further confirming that flexibility in the hood results in poor consistency. Our shooting percentage continued to increase at this event, with an accuracy of 83%.
Removed pneumatic cylinders for hood adjustment
Championships
To improve targeting accuracy, we had decided to add a laser to point at the top basket. The operator was now able to confirm the turret was aligned after camera tracking, or manual aim if the low basket was out of sight of the camera (i.e., blocked by another robot).
Laser Mounted (it can be seen mounted on the side in pick foam).
During our qualification matches we were having electrical issues, and pulled the laser during troubleshooting, as it may have been part of the issues.
Our shooting percentage did not improve much at this event, with an accuracy of 84%.
IRI
For IRI we decided to reattach our second shooter motor with the extra weight allotment given. This was to improve the shooters spin-up time. This probably had negligible effect on shooting accuracy.
Note: It’s been a crazy season, so some details may not be exact.