Poof Ball Dynamics

If anyone is interested, we have done some simple compression testing on the Nerf balls. We just squeezed the balls to a controlled thickness and measured the required force. We compressed with a 1-inch diameter rod, and a flat plate. We found that:

  • There isn’t much difference in loading with a rod or plate (a surprise to me).
  • Old, beat up balls are not very different from new balls (phew!). They are slightly softer.
  • The balls are decidedly viscoelastic, relaxing about 1/3 of the compressive load after 30 seconds of stasis, where they are essentially stable.

This last finding is an important one for us. It doesn’t matter much on the “long” end, (long duration), but when the load is impulsive, viscoelastic materials behave very differently then when the load is static. I have been very puzzled about our shooting tests. Why does the ball perform better when it is highly compressed through the shooter? We are zeroing in on a single wheel pitching machine, by the way. When a ball drops to the floor, only about 10% of its original height is recovered. This implies a very inefficient bounce, and thus, energy invested in compressing the ball should be mostly lost. It is not. The viscoelastic nature of the material probably explains this. Such materials often behave much more elastically when load rates are fast. In other words, the material modulus is not a constant, but a function of time. This makes the ball a much more efficient projectile.

Does anyone have any experience to share, or further thoughts?

Wow, John. Excellent post. For those of us that understand what you said, this helps A LOT!!!

I really have nothing to add at this point, but will once our shooting tests begin. I am quite interested in the results based on spin rate as well.

I am not sure if you tested for accuracy but if you or someone else did, how does the accuracy of a beaten up squished Poof Ball compare to that of a perfectly new one. This is the problem that our team is having trouble testing, and if anyone else has done anything I would be realy greatful.

I’m also very surprised with the result of loading qith a rod or a plate, as I would have thought that ther would have been more force required to compress witha plate (more surface area being compressed).

We have had very good luck with accuracy (repeatability) with both old, beat up balls, and new balls. We shoot at a target and observe the spread. Once we set up, the spread seems well within 30 inches at reasonable range. The largest variable is impeller velocity. They seem pretty consistent. The skin tends to “shred” off with use, but the compression characteristics seem to remain pretty stable.

We’ve found that a highly used ball is significantly softer than a nrand new one, and a new ball is much bouncier than a used one, but it didn’t appear to have much affect on shooting

We have been testing the same mechanism and discovered much the same effect (more highly compressed balls shoot significantly faster). I believe this is due to slippage when the normal force is not high enough. Though it may not seem that the ball should slip, consider the short time that the ball spends being accelerated (our video analysis suggests that the ball is in contact with the wheel and track (8" wheel with 90deg track) for less than .1 seconds).

Our quick tests have also found that the ball requires a fair amount of squeeze in order to get any velocity out of it. I’m talking about over 1" of compression. The first tests we made launched the ball about 8ft, though we calculated a good 18-20ft. Through some extra compression on the ball, and some higher friction material on the wheel, it was getting a lot closer to the 18ft.

As a reference, this was with a single 6" wheel, directly linked to a small CIM. We had about 45-90 degrees of contact with the wheel, though only about 15 degrees of that would result in the highly compressed area. More testing to follow, including a heavier wheel (flywheel) and a dual wheel shooter.


We are testing with a single shaft and two parallel Skyway wheelchair wheels 2.5 inches apart. The wheels squeeze the ball against a plate which can be moved to vary compression. You bring up a great point with friction, normal force, and slip. Our contact point is very short, so the ball has to be accelerated very quickly. We also improved performance by improving frictional characteristics of the wheel, but the performance change did not correspond with the normal force change, or the torque input versus ball range. I bring up the viscoelastic effect because the ball seems to return more of the energy used to compress it than I would have otherwise expected.

Thanks for all the comments!

would the gains that you would recieve by elongating the contact point be worth the effort of setting it up so that you would have a overall greater velocity? (i.e. some sort of pulleys/belt system)

Or is the small contact point more than enough to accelerate the ball to an ideal velocity?

95 measured the dimensions of 48 balls. While I do not have the raw data, there was one interesting note.

On average, the balls are slightly oblong. In one axis, their diameter is about a 1/16th smaller. I believe this corresponds to the injection spot. The other two axis where on average within a 64th or two of each other.

There where a few outliers, perhaps 4% to 6% that had at least one dimension up to an 8th off from the other two. They where just visibly oblong. We noted one ball that was much stiffer and harder to compress then the others. Many balls where very soft, which likely had much to do with the outer skin being torn on arrival. We noted at on a scale of 1 to 5, with 1 being near perfect and 5 being trashed, the 48 balls we have averaged around a 2. Most defects occurred around the balls equator, and consisted of torn skin. A few voids or dent’s were noted.

We had one ball that consistently preformed poorly in our shooting tests. We had already numbered the balls and #32 was widely unpredictable. Of roughly 1 dozen shots fired where over %80 of other balls scored, it failed to score once. It either would fire much faster or much slower then the other balls. All the other balls where quite consistent. I examined #32, but couldn’t come up with any explanation. With out dissecting it, it seemed identical to every other ball in our collection.

I experimented on one ball quickly by stabbing it perhaps 2 dozen times randomly around it’s surface with my knife. I found that it quickly became softer, as the air in the ball escaped much faster. Competition balls will probably follow a similar route as the skin is ripped off.

So, overall the balls are pretty consistent out of the box. They are on average a little oblong and will soften up over time. If our collection is Representative, then it’s possible that a very small percentage (.5%) of balls are demonicly possessed and simply will not go in the goal.

I’ll see about getting the .xls of the measurement data and methods posted. Right now it is stuck on a Windows ME machine that doesn’t recognize USB hard drives.
-Andy A.

Great work Team 95! Thanks for sharing it!

Andy, any possibility that the center of gravity for Ball #32 is significantly offset (moreso than the other balls)? Kind of like one of those “trick” baseballs that you can’t throw straight?

Our team conducted an environmental test on the ball. We measured the diameter of the ball at ambient temperature and then at 40 degrees Celsius, 95% relative humidity. Testing showed that at the high temp/high humidity condition, the weight of the ball increased from 183.5g to 188.8g and the circumference increased from 21 5/8" to 22".

Hi Steve,

I don’t think a long contact point is a good idea, unless you have to move the balls. We have found that the launch seems to be most efficient with a “point” loading of the balls. Again, we are using a single wheel to accelerate the ball, and a plate to “aim” it. I believe that the longer contact point will just waste energy. Our testing seems to corroborate this, as regards the ultimate range of the shooter. The whole compression, viscoelasticity thing kind of came up just for this reason. I might have thought that a long, gradual acceleration with minimal ball compression (because less normal force is required for friction) would be more efficient than a short, highly compressed shooting zone. Not so, surprisingly to me.

could the air trapped in the ball explain the the behavior of the poof-balls under impulse loads? and if so, do you think that the difference between a new and ‘compromised’ (i.e., missing chunks of foam and skin) poof ball might be enough that their behavior under impulse loads might change (i.e., compromised poofs won’t regain as much energy as was invested in the compression by the rollers, and therefore won’t shoot as far or as fast?)?

I can only hazard guesses at all the physics behind the poofs behavior when varying how compressible they are in a roller system and in flight. I can say that the ball I had stabbed and rendered squishy didn’t seem to behave much differently. admittedly, our testing was using our '02 bot (you may remember it, the popcorn popper). With some modification, it was found that the roller system used to suck up soccer balls happened to shoot the poofs pretty well to. The poofs were not being compressed a great deal, but it seems they don’t have to be.

From what I saw, no, squishy poofs fly about as fast as hard poofs. This may be highly dependent on how they are fired- the method used involved a track of rollers about 8 inches long, so the poofs had plenty of time to accelerate. As you said, impulse loads and sudden accelerations that a single ‘tennis’ ball shooter might impart could affect the poofs very differently. I’ve played with a single wheel shooter, but only with a single ball, so I have no good reference for that type.

I guess the only way to know is to put one of your poofs to the knife and see what happens. My gut feeling is that the longer the acceleration is, the more consistent the balls will behave. Short of mocking up all sorts of shooting mechanisms, my gut is about as far as I can go with this. John has done it, and figured out it doesn’t matter. Go figure. Like I said, this is complicated!

I think FIRST may be inspiring the most accurate and largest knowledge base on the proper handling and hurling of squishy Nerf type balls. Heck, we’ve already got kickballs down pat.

-Andy A.

Edit- As to #32’s misbehavior- I hadn’t thought about the CG being a factor, but it makes some sense. I suppose it’s possible that there is a dense spot in the foam. My best guess had been that the skin was somehow not homogenus and perhaps one color was grippier then the other, or perhaps the dense spot near the injection nipple is larger or harder. I really don’t know. I don’t have access to the balls for a while, but another member of 95 may be able to take a closer look.

It’s obvious, a disgruntled worker at the Poof Slinky(r) factory tossed a rat into the plastic molding machine. I bet if you cut Old 32 apart, you would find the modern equivalent of a Dutch sabot in the works.

I am interested if anyone is considering a multi stage acceleration, either by using multiple wheels or some kind of catapult launcher? Also is anyone thinking about adding a spin to the ball by either a rifled barrel or having paired wheels at a slight skew?

[quote=Dr.BotAlso is anyone thinking about adding a spin to the ball by either a rifled barrel or having paired wheels at a slight skew?[/QUOTE]

We are considering putting spin of the ball, but it would not be due to skew, as it can be done in multiple ways. We are thinking that we could run two wheels, one above the other, with the top going slower, providing backspin, or simply having a plate above one wheel, which also provides backspin. Having the wheels aligned on the same axis allows them to be controlled off of the same motor, which saves weight and complexity.[/quote]

I’m not sure if you said this or not, and if you did, I probably missed it somehow, but do the balls stay slightly compressed permanently?