[apalrd]

We used servos this year, and it was fine (once we calibrated them right).

I set it up with an Upshift and Downshift button (btns Right and Down respectively on a Logitech DualAction, yes those are face buttons), which then set the input state (high,low), which drove a state-machine through four possible servo positions (high,low,high-peak,low-peak), and that state would index a table to find the servo pwm values. It would peak and hold the servo, forcing it into gear then backing off to prevent servo burnout (we did burn out a pair of servos during the build season). I believe we were peaking for around 5 seconds.

Before doing peak and hold, we calibrated the high and low servo positions to the optimal steady-state position. The shift force immediately after requesting a shift was quite low, so I would have to jiggle the throttle stick to get both sides to shift. With the peak and hold, it usually shifted very quickly. We found that we rarely shifted, so it really didn't matter.

During autonomous, the script engine could directly set the input state and the peak and hold state machine would run. We ran autonomous in high gear, it was fast.


All of that said, we shifted pneumatically in 2011 and 2010 with no issues. We used the AM standard piston in 2010 and shifted at 60psi, and used a similar spec pancake-stype piston in 2011, also at 60psi. We heard about the breaking screws in 2011 and replaced the screw with a roll pin at CMP. I believe our 2010 robot already had roll pins. During the 2011 season, I also worked on an automatic trans shifting algorithm, and it always assumed it could shift with little lag during shifting (including during acceleration).