I just thought I would share some info about the air trajectory device that I built for this season. This was probably my favorite event (even though it caused me endless stress and frustration) and it's sad to see that it won't be replaced with another trajectory event :/
So this device uses pneumatic cylinders to launch the ball, but it doesn't use a swinging arm. Instead, the cylinder directly transfers its linear motion to the ball. The cylinders and pistons are made from common aluminum extrusions that were purchased online and at metal shops. It was actually surprisingly cheap! The grooves for the pistons were cut on a drill press with a hacksaw (because I didn't have access to a lathe, and there was too much chatter trying to cut it with a chisel). The ends of the cylinders were capped with a small chunk of aluminum extrusion and then brazed together.
The hardest part of building this was getting the device to be efficient enough to launch a small foam golf ball the full distance. The piston assembly that pushes the ball weighed quite a bit in comparison and probably 85-90% of the energy from the falling mass was used just to accelerate that assembly forward. I messed around a lot with different types of lubrication, cutting the grooves for the o rings deeper, and sanding down the pistons so they would make minimal contact with the cylinder. Also, the fact that I was trying to produce the pistons on a drill press instead of an actual lathe meant that they were not perfectly round, and resulted in some air leakages and extra friction. In the end, I found that getting a perfect air seal was not necessary, and that the lowest friction set up was the best.
The rest of the device is constructed of 1/8 - 1/4" extruded aluminum that I got for free from a cut off bin at a metal shop, a bunch of 80/20, and various brackets/connectors.
The reason I decided to go with this design instead of the traditional design of a bellow and a barrel is because this creates a much more consistent seal. Depending on the orientation of the ball in the barrel it could drastically affect the quality of the seal and thus the distance it would shoot. In addition, I noticed that you also had to win a sort of ball "lottery" because many of the foam golf balls and racquetballs I purchased were not perfectly spherical and some would have way worse consistency than others. You could use competition grade ping pong balls to get a perfectly spherical projectile, but I wasn't a fan of how susceptible they were to wind currents. I also found that sometimes with the traditional design the ball would leave the barrel with some spin, causing some accuracy issues at long range. This device seemed to solve all those issues but it still isn't perfect. I had tons of issues troubleshooting it in the beginning and building it took way too many hours. Also, it's very high maintenance. The cylinders need to be constantly lubricated, and the force of the piston slamming against the device after each shot meant that it got a lot of wear and tear. Enough to the point that the plywood base on the "cup" that holds the ball began to split (the veneers of the plywood split apart). There were also some times when the entire cup would just fly off.
By nats, I had tuned this device to get around +/-7.5cm at 8m and +/- 1cm at 2m. I think if I had some more time I could have gotten the accuracy down to +/-3-4cm at 8m. There was still some play inside the cylinders and the mass release mechanism was just a simple pin so the way you pulled it affected the consistency a bit.
At the national competition, the device got 0cm at the close target, 5.2 cm at the far target, and landed 2 bucket shots. To the best of my knowledge, this was the best performing device at the competition but we also got a 100 point penalty for being over the line when activating the device so we ended up in 4th.
There should be some higher quality images of the device on the best of 2016 wiki page if you're interested.