The final design is shown in the figure above. Each of the major components are listed from 1-7. Number one is the torque application. As it can be seen, instead of using the motor, a torque application system is used. For motion detection, inferred circuits were used. For force detection there is a force plate which goes on number 3, the vertical sliding mechanism. Number 5 displays the grippers which hold the ball and number 6 is the horizontal sliding mechanism which slides the right gripper left and right to account for different sized balls.

Pictures and Video

Visit our Youtube to view all of our videos.

Results and Discussion

The FT device did not consist of data acquisition system due to time constraints and a few other concerns. The device used is shown below:

Two similar types of soccer balls were tested using the device. They had the same pattern and were made by the same company. The circumference of the ball was found before testing. By knowing the circumference, the diameter of the ball was found. Also the values for normal force and torque applied were found. This was done for a total of six trials. Each trial consists for four different runs done.

1) The plate which held the scale wasn’t fully balanced thus a new system was implemented from trials 4-6
2) Contact point between scale and ball was on a seam section of the ball
3) The torque arm wasn’t fully in equilibrium
4) The no load weight for the torque arm wasn’t accounted for

These errors were fixed for trials 4-6. For trial 4 and trial 5 the sensitivity of the scale was tested. In trial 4, the normal force being applied to the ball varied from a scale reading of 50-200 grams. This gave very consistent results and a standard deviation of 0.02. While in trial 5, the vertical adjustable platform was increased so higher normal forces were being applied on the ball with a scale reading from 600-800 grams. While using this range the standard deviation rose to 0.07. Thus the sensitivity of the system is affected once in high ranges of 800+ grams scale reading. Trial 6 was done using an increment value for the grams being applied. It had a total of five runs. The first run was done with a scale reading of close to 100 grams. Then runs 2-5 were increased by an increment of 100 grams. So run 2 was at 200 grams scale reading, run 3 at 300 grams, run 4 at 400 grams and run 5 at 500 grams. The standard deviation was obtained at 0.03. Thus for better consistency and optimum performance from the FT, the normal force being applied to the ball should be less than the normal force corresponding at 800 grams scale reading for soccer balls and preferably as low as possible due to the scale sensitivity.

The same testing was done for two types of basketballs: 1) NBA official leather basketball, 2) NCAA Final Four basketball. The results are shown in the table below:

Table 5 shows the values for the basketball testing. Three trials were done for testing: 7) NBA Leather Basketball initial testing, 8) NBA Leather Ball optimized testing, 9) NCAA Basketball. Trial 7 had the largest standard deviation for coefficient of friction. This is due to many errors which occurred during testing:

1) The ball was over gripped which may have formed deformation
2) Table which the FT was put on, wasn’t very sturdy
3) Inconsistent results
4) Interference in the system
5) Grippers were designed for soccer balls

Thus the errors were fixed and Trial 8 was performed on the same NBA leather basketball. One of the things implemented was a new surface for the ball to be in contact with. A plexiglass was put on top of the plate which was in contact with the basketball. By doing so and recalculating the no load weight, the experimental values were more consistent and the coefficient of friction was found to be 0.702 with a standard deviation of 0.066. For trial 9 (NCAA basketball) the same setup as trial 8 was used. The coefficient of friction was found to be 1.402 with a standard deviation of 0.173. During this trial an error did occur. While adding mass to the torque arm, the arm wouldn’t spin as freely at the moment of slip. Thus it made it difficult to know the instant moment of slip. One way to fix this error is by implementing the data acquisition system. However due to time constrains and a few other factors the data acquisition couldn’t be used since it wasn’t available for use. Thus, for the basketballs the data acquisition system would be necessary since the moment of slip isn’t as freely visible due to the torque arm not moving as freely as it would when testing a soccer ball.