Launcher SYstem

The launcher subsystem consists of the flywheels, shafts, bearing blocks, bearing block mounts, motors, motor mounts, and the mounting rail. These components required careful design and analysis to ensure they would be able to resist the loads of launching and spinning up. 

When choosing the wheels we would use for the football launcher, the two most important factors were the size and cost of the wheels. A larger wheel would allow us to apply a larger moment to the football when launching it. We were also hoping to have a bolt pattern to use for mounting, so we selected a wheel that had one despite not using it in our final design. 

We chose 13-inch pneumatic tires from Harbor Freight due to the balance of cost and size. These wheels had bearings that we had to remove to be able to drive the wheels.

The wheel hubs have a clearance fit to the shaft and are mounted with two through bolts per wheel. 

The shafts are primarily designed to resist the loads applied during launching. This includes a load from the compression of the ball, "recoil" from the launch, and torque both from the motor spinning up and the wheel decelerating during launch. As with the other subsystems, reducing costs was a must. Therefore, we chose to machine these using stainless steel round stock from the scrap bin. Stainless is not a ideal material for this application, as the shafts are a challenging part to machine, but was our best option. The shafts also had to constrain the bearing block from moving up/down during launch, as the angle of the wheels may result in the ball axially pushing/pulling on the wheels. To accommodate this, we machined a threaded section on the shaft, which mounts a locknut, compressing the bearing block to a wider section of the shaft. 

For mounting the motor to the wheel, we chose to use a flexible shaft collar. The shaft collar has less than ideal characteristics under high torque conditions, but they provide "wiggle room" for any imperfections in our motor mounting which seemed worth it given our analysis predicted that torque on the shaft to remain low due to the wheel's high inertia.

Our bearing blocks are made of aluminum and the bearings have a clearance fit in the bearing blocks. We are using 1" ID ball bearings and the the inner races of the bearings are press fit in the shaft. We chose to press fit the inner diameter of the bearings rather than the outer to minimize internal moments within the bearing, which would decrease load capacity. We had two ball bearings per shaft and spaced them out as far as our stock would allow to resist the moment from the launch of the football. Each bearing could withstand a radial load of approximately 1000 lb, and radial loads of 100 lb, more than enough for our application.

A crucial aspect of these bearing blocks was that the two bearings were concentric. We took careful note of this when machining the bearing blocks, especially because we would have to flip the part during the fabrication process. 

The bearing block mounts allow our wheel and shaft assemblies to sit at an angle relative to the football launcher. This means that the wheels are offset from one another. This angular offset applies torque in the z-direction, which gives the football the spiral that we desire.  We chose to have the mounts angle our bearing blocks at a 22.5 degree angle. 

To ensure rigidity in our system, we also conducted FEA on both the full system and each individual component. We found that all parts were able to withstand the desired loads and all displacements were under 0.0045 mm, which demonstrates high rigidity. This was experimentally confirmed as, even with the excessive vibration caused by out balance wheels, slow motion video reveals that mounting assembly remains ridged.

We are using VEX CIM motors to spin the flywheels of the launcher. Motor choice and calculations can be found in more detail in the Electrical System.

The motors are mounted using a sheet metal part that bolts into the front of the motor and mounts to the side of the bearing block. The motor mount has a slot rather than simple holes, which makes the assembly process much easier and allows compensation for any inaccuracies in the machining of other parts.

When designing the motor mounts, we also conducted FEA with the stall torque of the motors (significantly higher than the torque the mounts will experience) to confirm their strength. While there is risk of stress fractures at stall torque, our analysis revealed that displacement was not excessive, and the mounts were more than strong enough for our application. 

We used 80/20 aluminum with T-slotted profiles for the mounting rail. This allowed us to mount it to the structure, as well as mount our bearing blocks to it via angle brackets. More importantly, it allowed us to adjust the position of our wheels relative to one another. This means we can set the football launcher for different ball sizes and make adjustments as necessary for the launcher to provide an optimal amount of compression on the football.

We chose to use two pieces of 1.5" 80/20 for their strength, and because we had access to both the extrusion and brackets for free. Reserch into the strength of these components revealed that our extrusion and connectors, 1.5" corner gussets,  can withstand 750 lb of force and 500 lb of torque each. Our calculations revealed this was more than enough to secure our assembly while providing modularity. Despite this high strength, we chose to incorporate two connectors to increase rigidity and provide redundancy incase of high vibration.

To meet our design requirments for spin rate and distance traveled, we first created a mathematical model to estimate the ball's distance traveled, and spin rate. This model centered around the equation above, which estimates the spin rate of the ball based upon the force exerted by each wheel in the y direction, radius of the wheel, contact time, and the ball';s polar moment of inertia. We validated this equation with a initial benchtop prototype, and found that actual ball rotation generally exceeds the model by around 30%, this is due to challenges estimating the coefficient of friction, spring constant of the football, wheel velocity, and ball tumbling. To account for this, we factored the increase into our model. 

 To optimize launch angle, we simulated our football's launch using projectile motion (with air resistance) in MATLAB with a consistent launch velocity. Although the optimal angle for distance occurred at around 43 degrees, we decided this mimicked a "long bomb" rather than the average football throw, so we decided on a 28 degree angle which balanced launch distance and trajectory. Our analysis indicates that our final angle would throw the football around 29 meters, which also provided leeway in our distance requirments. Additionally, the lower launch angle helps to keep the launcher's center of gravity well within the base of the mount, increasing stability and preventing tipping over under recoil.

To choose wheel angle, we considered both the launching distance and the spin velocity, as an increased wheel angle results in more force acting in the ball's y-axis, increasing spin, but also decreasing force in the x-axis, lowering overall velocity and travel distance. To improve model speed, the travel distance metric for this study was approximated without air resistance, which significantly increases predicted travel distance. 

We chose a wheel angle of 22.5 degrees as it provided a good balance of both travel distance and spin rate. The calculated spin rate is below the design requirments, however, 400 RPM is more than enough to stabilize the ball, better than most armatures, and a conservative estimate (the contact time is likely greater than we predicted, increasing ball spin). A 22.5 degree wheel angle also meshes well with our overall assembly and provides clearance from the 80/20 with shorter motor block mounts, reducing moment arms from launching.