Inconsistencies In Ball Density

VEX Robotics Competition Discussion VRC > Nothing But Net (15/16)
#1

Our team received our balls this morning and we found an interesting fact about the tolerances in the balls. Some of the balls we received are very squishy while others are almost rock hard.

The weight of these balls also varies by as much as 12 grams. With the most squishy weighing in at 43.1 grams and the heaviest at 55.3.

This density change also effects the our launcher. With a squishy ball (same squishyness as a clean sweep ball) our launcher hits 20 feet but with a very dense ball it struggles to get 5 feet.

We weighed all of the balls we received and they had an average global weight of about 49 grams. These inconsistency’s may were down through out the season but may reappear at worlds because the use of brand new game pieces.

#2

You should post this in the Q&A to see whether or not they meant there to be inconsistencies in the balls.

#3

5 and 20? I find that hard to believe, what kind of launcher are you using? What is the average weight and how far does that ball go?

So far we have been testing with clean sweep balls, and have been getting decent results. I wonder how different the results will be with actual game balls.

#4

If the balls vary this much in weight, it could be a real problem. I already thought that +/- 10% was a large margin for robot designs to deal with and was hoping they’d fall closer to 50 grams, but if the balls vary by this much it can really be a problem for consistency. I wonder how this could be solved if it is a common problem, because producing the balls again would probably be quite a big undertaking. And if that would be done, they would need to have some colour or printing difference so you can tell which ones have the correct weight.

#5

Its a 2 motor flywheel launcher that compresses the balls when it launches for better energy transfer. It works great with squishy balls but the stiffer ones do not compress as much making it harder for the launcher to launch them.

#6

Wow, I really hope this isn’t a problem throughout the year.

#7

Can you try to mount launcher wheels flexibly such that instead of fixed width between axles you compress balls by roughly fixed force (i.e. using rubber bands)?

#8

This is where testing and fine tuning comes into play. There is hopefully a middle ground. If you grab the balls less tightly you should be able to launch that more dense ball further, and the squishy ball less far.

However, if the balls really vary that much…I am really disappointed. We should not have to design around that large of a variance. I understand tolerance in making objects, but ±15% is a bit crazy.

#9

Manufacturing large quantities of product to tight specs could be both tricky and expensive. You either need to have your fabrication equipment built with those tolerances in mind, or your quality control need to throw out large amount of out-of-spec items into the trash can. That is - if you have quality control test equipment that could test to those tolerances large amount of the items. Think about quality of LEGO vs other no-name-brand blocks that hardly stick.

I think ball production was outsourced to a large manufacturer, who never had clients requesting any strict tolerances. They, probably, don’t even have any quality control equipment at all. They have a person picking a ball of two from each batch and measuring it manually. They, likely, do not need anything more sophisticated. Try to imagine DollarTree complaining that $1 play balls don’t meet +/- 10% weight specification.

In addition to that, the mass production facility, likely, has their processes vary more than a smaller operation that produces samplers for the potential clients…

I am a bit disappointed too, but VRC was never meant to be easy. If those tolerances are, really, that bad, or even worse between the different batches, then the game just turned out harder than anyone expected.

What I afraid of, is if a team or club will practice all season long with roughly similar balls from the same batch and then discover, just before the states, that other batches’ balls have significantly different specs - that would be some bitter disappointment.

I think we should sample the balls, post the measurements, and create a size / weight distribution chart. If balls are out of declared specs then, at the very least, game manual should be revised to reflect that.

However, most of VRC participants are future engineers, and will greatly benefit from learning how to figure out clever ways for handling such situations. I am sure there are at least a few good way.

#10

That is something I am going to have them try. We weighed each ball and sorted them by those that were easy to compress versus those that were a bit stiffer. The majority of the balls go through with little effort just as the Clean Sweep balls have, but there are about five that are difficult to squeeze. So a slight alteration to the launcher will be necessary. I have two in front of me now. One is very easy to squeeze and has a mass of 43.1, while the other is extremely dense and has a mass of 55.3. We can get it through the launcher if we give it a little harder push.

#11

Out of curiosity, I did some research on the foams and learned more that I wanted to know for the NbN purposes. Here are some highlights, relevant to foam and engineering:

Engineering your system in a way that it depends on the consistency of the foam formation will lead to nothing but frustration.

Foam formation is subject to so many variables that getting manufacturing process to be consistent is non-trivial. For example, polyurethane foam manufacturing depends on temperature, pressure, exact timing, accurate mixing, and viscosity of the filler material as well as its interaction with the foaming agent. Don’t expect a large generic toy manufecturer, who was sourced for NbN balls, to do strict control for all those variables.

Even if the flight safety depends on the foam quality - it is still too hard to get it right. This lesson was learned in a hard and tragic way. The loss of the Space Shuttle Columbia was primarily due to the the unpredictability of the foam formation while creating the foam insulation on the external fuel tanks. There was a systemic risk in a way entire system was designed - the shuttle was placed below the fuel tank and small pieces of foam insulation kept falling and hitting thermal protection system tiles on virtually every flight. They have tweaked foam application process for years and still couldn’t ensure consistent quality. It would be too expensive to test the foam on the entire fuel tank with x-ray photography, which appear to be the only way to ensure acceptable foam quality. Entire Shuttle program was scrapped in favor of more reliable designs. If foam coated fuel tanks are ever to be used for the space flight again - the crew capsule or cargo will always be placed on top of the tank to avoid any consequences of the inevitable foam loss.

So, as frustrating as it is, we simply have to accept that the balls will be of different size, mass, and density. The best we could do - is to create a good distribution chart and then design the launchers and game strategies accordingly.

#12

Yes, but this is not the dollar tree. This is a robotics competition in which tolerances are very important.

Did you happen to see any skyrise autonomous routines work perfectly on one field and then not on another? Of course you did.

The official NbN field specs say that balls weigh .11 lbs (49.9 grams) ±10%.

+10% = 54.9g
-10% = 44.9g

Tesla529 measured balls at 43.1g (out of spec) and 55.3g (out of spec).

Assuming Tesla’s numbers are correct, the balls are out of spec by VEX’s own numbers. Yes it’s not much, but specs are specs for a reason.

It will be interesting to see how large samples of balls measure up. As technik said, this will be a part of the design challenge (intended or not).

#13

My box has one at 48 and one at 53. Everything else was in between…

#14

One ball test I am interested in seeing: something like a drop test. Drop the ball from, say, 2 meters and see how far up it bounces. Nice application for the iPhone slow motion app. Such a test might give us some idea of the variability in rapid compression/decompression energy losses, not identical to but at least similar to the kind of energy losses that will be seen in a flywheel shooter and much easier for everyone to make comparisons with around the world. Perhaps humidity and temperature and even weeks of use could affect how the balls behave.

There seems to be a lot of science involved in this year’s game.

#15

Those numbers sound much more reasonable.

#16

If balls have started shipping I don’t think anything could be done… Changing one line in the Field Specifications seems to be the simplest way to “fix” the problem.

The reason OP was getting so much variation in flight distance is, probably, due to fact that most of the flywheel’s angular momentum was spent compressing the ball and later turned into heat.

I already started thinking what would be the best way to run both flywheels off the same pair of motors, without fixing distance between the flywheel axles, so we could push on the balls with constant force.

Also, since the balls are of the various mass, but the size is roughly the same, their Ballistic_coefficient will vary and lighter balls will not fly as far, if launched at the same speed, due to the air drag.

So you either have to overpower the launcher to send the balls directly into target or figure out clever way to vary speed, depending on the ball’s weight.

On one hand, if you have flywheels of just the right moment of inertia it will give lighter balls more speed than to heavier balls.

But will it work only at the specific distance to the target, or at some acceptable distance range?

Also, variable density of the ball makes anything dealing with momentum transfer non-linear. So the only way to quantify all those effects is to test, test, and test with multiple launcher designs and large sample of balls.

Edit: You know you don’t have to launch every ball correctly, just make sure that your launcher works well for 95% of them (or any other number that you like).

Important is that that the intake and launcher will not choke on any outlier ball. Max or min size of the ball is really something that might be a problem.

#17

I’m wondering how consistent the size of the balls is? Having the molds at exactly 4" shouldn’t be a problem, but from what I’ve read the PU foam can shrink or expand slightly (forget which way) right after molding. I doubt that there is much variability, but could someone take some exact measurements of the size of the balls?

#18

The sizes of the balls are consistent, it is the the mass of the balls that is the problem. All of the balls were exactly 4 inches give or take .2 inches. If these sizes change I will make sure to let you know. Then again, we have only had our balls for one day, after a while the sizes of the balls may change.

I am wondering whether you got your balls in a Welcome Kit or a Field Element Kit because this may make a difference (It probably won’t but I am curious).

#19

Game element…

#20

Has anyone asked about the ball inconsistencies in the Q&A yet?