[seg9585]

As a 2006 FRC participant (Aim High) who built a robot using the pitching machine launcher design, I have a few comments about it:

First off, calibrating the wheels ended up being a little more annoying than expected. Even though we used encoders to measure and even out the speed between each wheel, we found the release angle was still inconsistent -- the balls needed to be squeezed by the wheels a bit to get the traction needed to accelerate the ball, and the balls would end up flinging a bit to the right or left depending on which wheel made contact with the ball first and how the ball re-expanded.
For Aim High this wasn't a big deal because the target was relatively large, but with baskets this year we need a lot more precision.

Also, the ball originally came out without spin, so the balls flew like a knuckleball, making scoring unpredictable. We changed the speed of one wheel to add spin, but then had to apply a directional bias since the spin did induce lift on one side of the ball (still better than a knuckleball)!

Lastly, we had to run those large softball wheels with CIM motors to get them spinning fast enough, but when they were on all match it quickly drained our battery. In later matches we had to turn off the spinner during collection periods just to save power, then wait for the wheels to spin up to speed again.

I did see effective single-wheel launchers against a static surface that took out a lot of the multi-wheel variable issues we had, and they seemed to work well.