Thread: Ball Compression Data
View Single Post
  #1   Spotlight this post!  
Unread 03-04-2012, 22:31
Bruce Newendorp Bruce Newendorp is offline
Engineer
FRC #0525 (Swartdogs)
Team Role: Engineer
  
Join Date: Jul 2004
Rookie Year: 2001
Location: Cedar Falls, IA
Posts: 74
Bruce Newendorp is a splendid one to beholdBruce Newendorp is a splendid one to beholdBruce Newendorp is a splendid one to beholdBruce Newendorp is a splendid one to beholdBruce Newendorp is a splendid one to beholdBruce Newendorp is a splendid one to beholdBruce Newendorp is a splendid one to behold
Ball Compression Data
We installed a ball compression transducer on Thursday morning of the MN 10,000 Lakes Regional. It is a 10 lb. load cell made by Measurement Specialties Inc. (FC2231-0000-0010-L) purchased from DigiKey. Thanks to Team 2169 for the help. It is opposite one of the rollers in the ball handler. There is approximately 1 1/4" of ball compression where it contacts the load cell. There is a 4" X 4" 1/8" thick polycarbonate plate between the ball and the load cell to compress a large area of the ball and transfer the force to the load cell. The plate is held in place at one end with double stick tape. We made an aluminum cup with a flange around the upper rim to hold the load cell. We drilled a 1 1/4" hole to drop the cup/load cell assembly into the ball guide. Sorry but we got rushed and did not get pictures - you know how it is on Thursday morning...
We ran 5 balls through 5 times each to find out how much variation there is for a given ball. The variation for each ball ranged from 2.6 to 12.5% of the average value and the difference between balls is larger than that so we concluded we could differentiate between hard and soft balls as they passed through.
We saved the peak value for every ball that went through the robot. The data is in the attached Excel file. The data for about 120 of those balls is in the file. The remaining data from semi finals and finals is still in the cRIO. In the rush at the end of finals we failed to download the last 4 match files and now it is sealed in a bag in the crate on the way to St. Louis.
Initially we gathered data only. As you can see in the data, the new balls are easy to pick out. It appears we were maxing out the load cell at a 918 output value when the hardest balls went through. Over time, we started triggering a shooter speed reduction for hard balls and and hood angle change for soft balls. By Saturday afternoon were calibrated pretty well.
Our next task is to match up the force data and shooter speed data to each shot in our match videos to determine if we can improve our algorithm.
Last edited by Bruce Newendorp : 03-11-2012 at 22:16.
Reply With Quote