We did lots of prototyping of our shooter.
Our shooter:
CIM motor with wheel mounted directly to it.
Wheel is custom made out of HDPE
-8" diameter x 1" tall
-Wedgetop tread mounted to the HDPE wheel with rivets. Ends attached with AM clamps from performance wheels.
Shooter deck is made of HDPE sheet that is 1/2" thick. We had to brace the bottom of the deck to prevent sag which caused more compression on the frisbee.
Rails of shooter are made of HDPE with weather striping foam on inside edge. The rails are attached by machine screws about every 5" to prevent warping/bending when the frisbee moves around it.
The spacing from the shooter rail to the wheel is 11 1/16 inches. The foam is 1/4" inch thick so we get some compression on the frisbee but not much.
We also have a lexan cover over the top of the entire shooter.
You can see a picture of it here:
http://www.chiefdelphi.com/media/img...ed21b3a4_l.jpg Notice that the wheel here is a performance wheel that has been replaced.
Here is an image with the shooter on the robot:
http://www.flickr.com/photos/chaprob...57632928507774
Compression and surface on the wheel and rails were were the factors we spent the most time on. The diameter of the rails need to be determined by the size of your wheel so it is consistent all the way around.
Our shooter actually shoots better without full voltage. We have experimented with 9.2 - full voltage. I do not have any specific data but we mark the frisbees less with lower voltage and get more consistent distance and accuracy we think with the 9.4-9.5 voltage range. We competed at our first tournament and our encoder was not giving us the feedback we needed to control the shooter speed. We ran either full voltage or 80% for all the matches.
We can shoot from full court consistently at the two point goal.
Hope this helps.