Top/bottom roller shooter vs. bottom roller only

We have been heading toward a shooter that has spinning wheels on both the bottom and top side of the ball. By setting the speeds differently, we can get a desired amount of backspin.

But I’m thinking that in addition to backspin influences, this will also influence angle of exit from the apparatus. I know some designs have only a bottom spinning wheel and use a non-moving hood on the top side (except for possibly angle adjustment) I’m thinking this will result in a more reliable, consistent shot, even though it may not go as far, and the backspin RPMs will increase along with the velocity, vs. being able to independently control speed and backspin with wheels top & bottom.


With one driven wheel at 1500 RPM, you get the same exit velocity as with two driven wheels at 750 RPM. So I am leaning towards the two driven wheel solution.

I am thinking that two FP motors driving one wheel is the same energy into the ball as one FP driving each wheel. Yes, you should manage your spin as well.

So if there’s a vote, I pick the 2 roller shooter.

Not necessarily true. In the fly wheel designs very little of the force is in fact transmitted directly to the balls from the motors. Just look at the torque and how fast that ball gets moving in a short period of time…it just wouldn’t make sense. What it really is, is a conservation of momentum. The angular momentum of the wheels is translated to linear and angular momentum of the ball. Therefore, all that matters in terms of how fast the ball goes is the CHANGE in angular momentum of the wheels. A bigger decrease means a faster release. Simply adding another motor to the axle of the wheel at the same reduction will not cause a faster release because the free speed would remain the same. However, what it would do is cut in half the time it takes for the flywheel to get back to full speed because the torque output of the motors doubles. If you want a faster release, you would need to reduce your reduction on the motors, increasing the free speed of the fly wheel and increasing the change in angular momentum of the wheel.

However if you put the motors on separate wheels, then you will get a greater decrease in angular momentum of the wheels because there are now twice as many wheels transferring momentum to the ball. Then you wouldn’t have to adjust the reductions and you would increase the release speed. (However, your reload speed would remain the same…trade offs!)

We started with a 4 wheel pitcher (like the early ones in videos here), it ran on 2 CIMs while we were waiting for the AM FP gear motors. Each axle had 2 narrow wheels.

Based on some further discussions we prototyped a single wheel, 1 motor design, as seen in other videos.

Real world results - we have found simple is better and we are going that route (hint hint). We get a really nice trajectory and consistent backspin. The key seems to be in tuning the angle and length of the box ‘lid’. And consistent ball insertion.

One thing we found in quick testing is that the 4 wheel method was a lot more touchy, if the ball didn’t contact all 4 wheels smoothly and at the same time it created erratic shots.

Either design will work and I am sure we will see plenty of both, but a single bottom wheel looks good to us.

I think the initial velocity of the ball is a function of the rotational velocity at the edge of the wheel. Wouldn’t your 2-wheel example just produce 2X torque (if driven at the same speed) rather than 2X initial velocity of the ball? I must be missing something…

Our students built a 2-wheel prototype (both on the bottom) that is pretty consistent and decided not to look at a 2-wheel (top & bottom) shooter.

John V.,
Our prototyping experience supports your statements regarding shot to shot consistency.

For any particular motor selected to power your wheeled shooter, you can choose a gear ratio to get your desired shot speed. A single wheeled shooter needs to spin ~2X as fast as a double wheeled shooter. A speed difference between your double wheels, either with speed controlled motors on each wheel, or mechanical coupling, gives a selectable amount of backspin. A single wheeled shooter is slightly less “efficient” in that much more energy is expended putting a higher, non-selectable, amount of backspin on the ball.

Energy transfer happens nearly instantaneously, and is governed by the equations derived in a different thread. The most important factors affecting ball speed are rotational speed of the wheels and their rotational moment of inertia. Motor power has little effect on the shot itself. Rather, the more motor power, the faster the initial spin-up of the wheels, and the shorter the time between shots.

The majority of the time, the motor runs at nearly its free speed, inhibited only by the energy losses in the power transmission mechanics. The reduction in speed after a shot takes the motor closer to its max power speed. This helps the motor power the wheels back up to their full free speed. The higher the rotational inertia of the wheels, the less reduction in speed during a shot.

We found that our single wheeled prototype was more consistent, shot-to-shot, than our double wheeled prototype. Both types of wheeled shooters depend on ball friction and compliance for ball speed. The balls vary in both of these aspects right out of the bag. We expect these variations to only get worse over time. So, there is going to be a limit to the ball to ball repeatability of wheeled shooters that is out of our control. We also expect that the accuracy of a wheeled shooter will decrease over time, as the balls degrade.

A well executed catapult should be less subject to ball to ball variation as long as the shape and mass of the ball is consistent. Our catapult prototype keeps being “almost there”. So, we have decided to go with a single wheeled shooter as our “plan of record” and proceed full speed ahead with detailed design and fab. At the same time, we are letting a couple people finish the catapult in case it proves to be a revolutionary breakthrough in terms of repeatability.

I vote for two-*roller *(NOT two-wheel) option as well for the reasons Don highlighted.

I think that rollers are more tolerant to ball variances than wheels are because they’re wider (either in feed, density, compressibility, etc). Our two-roller shooter this year (and in 2006) was VERY consistent, to within less than a ball radius over 20-30ish shots in our short-range prototype being fed by hand (i.e. inconsistent feed) with 6 different balls.

My $0.02

What diameter roller and how much compression? We are looking at a hybrid with 1.75" roller and 1.5" compression. I am worried that we need to have more compression or risk the dreaded orange snow.

Our tests were done with 1.5" rollers running about 1.5" of compression. No-slip carpet material is a good roller wrap.

I’ve seen how challenging it is to make one wheel spin with controlled speed…I have little desire to double our workload.


I will have to agree with Todd on this one. A one axil shooter has been much more consistant than the 2 axis shooters we prototyped. We found that two
8" AM wheels roughly 2-3 inches apart running at 3500 RPM will shoot 1/2 court with a success rate ~40% with a rough prototype and hand feeding. We think we could get that up to 60%. As Todd stated before, the variation of the balls will make consistant long shot extremely difficult.

Interestingly enough, we prototyped both 1 and 2 wheel shooters and got exactly the same error with both (about 0.4" for every foot of distance). However, the two wheeled design shot much farther, as would be expected. The advantage of a one wheel design is that it gives wickid backspin and we found that this gave a higher shot percentage (with the same overall shot error) because balls that hit the backboard tended to shoot down into the hoop.

Our team went with the one wheel design. We agreed, though, that either could be made to work. We also agreed that a catapult would be immune to the ball error, which I discussed our attempts to quantify in another thread (squishimetrics).

The most interesting test was last night when the local police brought a radar gun over to test our shot speed. After adding aluminum foil to reflect the radar, we found that our shots were coming out at about half of the theoretical value. We, therefore, need to really investigate the energy transfer efficiency at different speeds, and its likely not a linear relationship.

That’s the disadvantage to the wicked backspin. The impulse given to the ball is split between angular and linear momentum. The more backspin, the less forward velocity we will see off our shooter.

Exactly… That, and we can hear that there is some degree of slippage on ours. There is also energy loss in compressing the foam, deforming the device, etc…

You need to have some wicked rpm on a single wheel to get good distance with it, but imo having the ball bounce off the backboard, into the basket, is worth it

My team is going with a one wheel shooter because it is more compact for being on a turret

Ok so, I am design our shooting system and have designed a two roller shooter, however I was informed there is an optimum roller diameter based on the rpms it will spin and the ball diameter and I was told there where calculators. I found jvns design calculator but there wasnt anything about shooting. Does anyone know about any calculations and formulas I can use or of a Design calculator for this?

There should be an optimum diameter, but it will also depend heavily on your design, including variables like how many motors (not just RPM, torque will play a role), moment of inertia (if you have a flywheel, or just the inertia of the roller), friction in the bearings, and perhaps most importantly, how much you compress the ball. An analytic solution for the best diameter would be difficult if not impossible because some of those variables are very hard to measure and quantify. So your best choice is probably just to test a bunch of rollers until you get something that works.

A decent estimation is that the ball will get moving at 60-70% of the ideal free speed of your shooter roller’s tangential velocity. Aim for 1-2" of pinch, with some adjustment (we’re using 1.5" of pinch). Wrap your rollers in a grippy, compliant, material.

Hope this helps.

For a 2-wheel (top and bottom) shooter with both wheels the same diameter and powered and spinning at the same speed: If you assume that the ball exit speed is equal to the tangential velocity of the wheels then

wheel rpm = 720V/(pid)

V is ball exit speed in ft/sec, and

d is wheel diameter in inches

For a one-wheel shooter, the required rpm will be approximately twice that.

As for motor selection and gear ratios, here are some thoughts: