Disc Projectile Testing!

Part of an a task assigned on Coop was to improve the performance of a disc launcher product used in a produc. I was told that the current performance was unacceptable and in order to keep customers happy it must perform as well as an existing product.

I decided to use the highspeed camera I had access to in order to measure the "muzzle velocity" of both the batwing and current toy in order to compare them. I would then make modifications to the new toy based on my observations and record the results. Hopefully I could change the shooting in some manufacturable way to get results close to the existing product.

Using the high speed footage and the known framerate of the camera, I could solve the speed the discs were launching out at by counting the frames it takes for a disc to cross the 6" scale I had made. My results from the testing were as follows:

I tested two different Disc sizes and a number of different modifications that were thought of by myself and another engineer during a brainstorm. They ranged from changing the thickness of the launcher clip, to changing the clip geometry to adding support walls to help stiffen the clip. Each of them improved performance slightly and when two of them were used in combination (Extended arm + Back) it resulted in an average muzzle velocity of 15.41 ft/s which exceeded the goal of 15.0 ft/s. 

Disc Launch from 4ft at 15 degrees

Graphically it is easy to see the predicted performance gains from my modifications. 

Lessons Learned:  I struggled a lot in the beginning because many of the past attempts to make the mechanism more efficient would result in the toy misfiring or stalling. This was obviously a problem and I had to be very precise with my modifications to make sure they were as accurate as possible and represented the "ideal" case of a molded part. I also learned that documenting my results consistently was a very efficient way of comparing. Because of random variation, some mechanisms would get good results every once in a while others would perform poorly randomly. Had i just examined a small sample and not kept track of all the testing of each I could easily have been misinformed about which results were which