View Single Post
  #25   Spotlight this post!  
Unread 23-02-2012, 22:48
Wayne TenBrink's Avatar
Wayne TenBrink Wayne TenBrink is offline
<< (2008 Game Piece)
FRC #1918 (NC Gears)
Team Role: Engineer
 
Join Date: Apr 2007
Rookie Year: 2006
Location: Fremont, MI, USA
Posts: 528
Wayne TenBrink has a reputation beyond reputeWayne TenBrink has a reputation beyond reputeWayne TenBrink has a reputation beyond reputeWayne TenBrink has a reputation beyond reputeWayne TenBrink has a reputation beyond reputeWayne TenBrink has a reputation beyond reputeWayne TenBrink has a reputation beyond reputeWayne TenBrink has a reputation beyond reputeWayne TenBrink has a reputation beyond reputeWayne TenBrink has a reputation beyond reputeWayne TenBrink has a reputation beyond repute
Re: 1918 NC GEARS 2012 Robot

It is about 36" tall. (edit: already been answered)

The primary roller in the frame and the arm roller in the arm are both powered by the same FP motor. The roller motor/sprocket axis is in line with the arm pivot, so the urethane belt tension doesn't change with arm position. The arm roller does a decent job picking balls out of corners, etc., when the chassis isn't moving too fast, but for open field pickup we generally keep the arm up.

Our alignment is simple, but seems to work well. Mostly its just the drivers eyeballing it. We also use the camera when needed. We don't expect to win any Rockwell Automation Awards.

We have seen the same distance spread due to ball variation that most others have reported, and have taken steps to minimize its magnitude and impact. We did some studies and found that the "sweet spot" for scoring for a given wheel speed combination was quite deep. (We run the bottom wheel wide open and throttle the top wheel to control distance, which improves backspin). Shooting straight on, we could hit about 5/6 from within a distance tolerance of +/- 1 foot from the "ideal" distance thanks to backspin and the backboard. It dropped off to about 1/2 at 2 feet from "ideal". Once we figured this out, we changed our approach. We quit worrying about adjusting wheel speed as a direct function of distance and set up a few standard speeds for shots we expect to be taking often (short, medium, and long, etc.). Now, we just concentrate on the aim (azimuth) and get in the ballpark for distance, something the drivers can judge at a distance or quickly see from the camera. With a bit of practice, they were able to line up and shoot in a couple of seconds. We expect to miss a few, but that is likely to happen anyway because of ball variation. We'll just pick 'em up and shoot again.

We learned a lot about shooters this year (starting from nothing, it didn't take much!). The main thing we learned (thanks to some help from my brother) is that the motor doesn't really power the ball to the target. The kinetic energy in the ball comes from the stored energy in the spinning wheel (flywheel). A bigger motor (or more of them) doesn't throw the ball farther, it just improves the recovery time between shots. What you need is a good flywheel(s) with some decent grip (texture and compression) on the ball. We went from two 775's per wheel to one, and redesigned our wheels.

We chose early on to concentrate on shooting from the key and closer. Not only is the scoring percentage better from up close, but it lets us gear the shooter wheels for faster recovery between shots and keeps us in or near the protected zone. We also made ball collection and ball handling reliability a priority. Collecting from one side and shooting from the other allows for a straight ball flow. We also wanted a very low CG. This approach works well for that.

We saved a few new balls to recalibrate the wheel speeds prior to competition. The new ones go a lot farther than the well used ones.

Special thanks to whoever it was that came up with the trajectory calculator. It was a priceless tool in our design process. We used it to determine ball speed and launch angle, which drove amost every other detail of the ball feeder/shooter.

Edit: I hope the GDC offers their wisdom on the <R35> obscured number issue. This will come up again and again this year and in the future. IMHO, the intent of the rule is met if you can read the numbers while in the starting configuration. People don't come to events to check out the sleek lettering on the bumpers. They come to watch what the robots can do. It would be a bummer if this became a significant design constraint for manipulators. IMHO, putting numbers on the arm wouldn't satisfy a lawyer, either. Numbers get obscured in the course of playing the game by other robots, field elements, etc. I'm neither a lawyer or a GDC member, so I'll deal with whatever happens.

We will miss TC. We have many fond memories and hope to return soon. It just didn't work with our schedule this year.
__________________
NC Gears (Newaygo County Geeks Engineering Awesome Robotic Solutions)

FRC 1918 (Competing at St. Joseph and West MI in 2017)
FTC 6043 & 7911

Last edited by Wayne TenBrink : 23-02-2012 at 23:01. Reason: Spent so much time drafting, it was obsolete!