View Single Post
  #4   Spotlight this post!  
Unread 20-01-2017, 00:59
Ekcrbe's Avatar
Ekcrbe Ekcrbe is online now
When can I watch Einstein again?
AKA: Erik Boyle
FRC #4640 (Metallic Panthers)
Team Role: College Student
 
Join Date: Jan 2012
Rookie Year: 2012
Location: New York, NY
Posts: 333
Ekcrbe has a reputation beyond reputeEkcrbe has a reputation beyond reputeEkcrbe has a reputation beyond reputeEkcrbe has a reputation beyond reputeEkcrbe has a reputation beyond reputeEkcrbe has a reputation beyond reputeEkcrbe has a reputation beyond reputeEkcrbe has a reputation beyond reputeEkcrbe has a reputation beyond reputeEkcrbe has a reputation beyond reputeEkcrbe has a reputation beyond repute
Re: Accuracy of single axis shooter

I made a post about shooters in a different thread that I think covers some of the general questions about what to consider when designing one.

Basically, there are no quick-fix answers, and it does involve a lot of trial and error, but from some FRC experience, I will attest that it can be done very well.

Try to diagnose the source of your inconsistency. Are your shots inconsistent left/right, long/short, and/or high/low? I know it may not be immediately apparent, but try to tell whether your shots are coming out at different speeds or different directions if you are missing the goal from front to back. I would hypothesize that some likely culprits are inconsistent positioning of the balls coming into the shooter (for left/right inaccuracy), poor speed control on the wheels (long/short), or too-sudden decompression of the balls upon release.

I've seen all three issues in shooters I had a hand in during my years, and I would say a few things.
  • It's important to have your feeder send the balls into the shooter in the same position each time, not try to make the shooter align the balls into its preferred track. This is a recipe for shots to come out crooked. (edit: To elaborate on this, consider that if your feeder does put balls into the shooter in slightly different places, you want them to come out perfectly parallel, so that their total inaccuracy is only that very small difference, not something much larger. I like having a consistent profile across the width of the shooter to help with this, although others will say two wheels with a space in between can work just fine as well.)
  • The importance of closed-loop speed control cannot be overstated. We had a lot of success in prior years with instituting a "speed check" which would disable the feeder into the shooter unless the wheel was within a small error range of the requested set point. You'll have to balance your required accuracy versus your desired throughput rate to determine what that acceptable range is.
  • And I've found it nice, although other teams probably have made good shooters without caring for this point, to have the hood gradually decompress the ball in the last ~30 degrees or so and end with a short section facing tangent to its arc to help the balls return to their natural shape after being compressed by the wheels. If your hood cuts off suddenly and the balls leave contact with the wheel and hood at the same time, the balls can "pop" back into shape in unpredictable directions, causing vertical inaccuracy. This is likely the point that the most people will scoff at because they think it doesn't matter, and maybe it doesn't, but I've never regretted paying attention to this detail.
__________________
Four years of FRC 68 Truck Town Thunder
2012 Championship Newton Division Finalists—Thanks 330 and 639 for the Full Court Press!
2013 Kettering University District Chairman's Award Winners!
2014 Great Lakes Bay Region District Winners—Thanks 288, 4819, and 5166! | 2014 Waterford District Chairman's Award Winners! | 2014 Michigan FRC State Championship Chairman's Award Winners—Congratulations 33 and 503!
2015 Center Line District Chairman's Award Winners! | 2015 Lansing District Winners—Thanks 314 and 1684! | 2015 FIRST in Michigan District Championship Chairman's Award Winners—Congratulations 503 and 2137!



Last edited by Ekcrbe : 20-01-2017 at 01:05.
Reply With Quote