Sorting injection molded bbs nondestructively by internal mass distribution

[fortjp500]

I am looking to replicate the work of another who devised the below sorting system which is for a hobby of mine. It sounds like the injection molded bb is being rotated and somehow sorted based on how it dynamically spins or possibly wobbles versus it's original diameter. Does anyone know of a system/product that exists that does this?


"BBs are a major limiting factor in the accuracy of airsoft guns, but how limiting precisely and how much improvement can be made precisely is yet to be properly quantified. To this end we sought to construct a machine which would sort, or segregate into bins, BBs based on their mass distribution consistency and then test its efficacy against a variety of popular brands of airsoft BB. The results were substantial improvements across the board, with universal reductions in average group size and broad improvements in regards to grouping standard deviation, maximum group size, minimum group size. In terms of percentages, improvements averaged approximately 10-25% depending on the specific metric. In summary, injection moulded BBs show marked improvements in regards to precision when binned with our method. This, combined with other research undertaken, have yielded startlingly large performance gains with relative ease, inviting further questions regarding the actual mechanical limits of the 6mm smoothbore-backspin system in regards to ultimate accuracy.

What if I could bin BBs, that is to say sort them by quality, to tighten groups and kick out those pesky fliers? Technically this isn't new, people have been trying to sort BBs for decades now in pursuit of better accuracy. Placing rounds, one by one, on a crack scale and sorting by mass, measuring with calipers for diameter, and so on. Clearly though, for my efforts, such a process would have to be automated as there is no way I'd do the airsoft equivalent of “handloading.” That said it has the possibility to substantially improve accuracy.

There is another problem with this though: it is fairly easy to sort BBs by diameter and weight, in fact a decent factory should have already done this for you with a quality BB. It is possible I could build a machine to do this to even tighter tolerances, but what I wanted to do was test for something much harder: sphericity and density consistency. You see, we all know bubbles are the bane of a BB's existence, and so we regularly break them open in tests to qualitatively assess their consistency. That is only part of the issue though. Plastics don't typically come in sufficient density to make BBs, thus dopants are added to increase the mass. For example when certain ferrous dopants are used, the BBs can become mildly magnetic as some people may have observed. Why do I bring this up? Because the blending of those dopants with the plastic may be imperfect, there may be settling in the injection mould before the plastic hardens, or any of a couple other potential sources of inconsistency, and so the BB may be visibly perfect when split however it may have worse mass distribution than a BB with significant bubbles. How do you test this? How do you even test for voids in the BB without cutting it open? Simple: spin up the BB, nice and fast, and it will be forced to rotate around its center of mass. The worse the BB wobbles, the worse it is. Issues with sphericity can also be readily observed here. So how about a machine that individually spins each BB, measures the difference between geometric center and center of mass, and bins the BBs accordingly?"

 

[SwinnyGG]

Is this just a thought exercise, or is your goal to actually build something?

What you're talking about (measuring very small deviations in radius of gyration to determine the distance between geometric center and center of mass) is technically possible... I think... but are you prepared to spend 6 figures for a machine that will do this? Because that is what it is going to take, on the low end.

I honestly have a hard time believe that someone has already constructed a machine to do this for a reasonable amount of cost; measuring forces this small is very hard to do accurately. Is this beyond theoretical in someone else's hands and you want to duplicate?

 

[MintJulep]

Let them roll down a half cylinder and use machine vision to look for wobble.

 

[fortjp500]

I would like to understand what technologies and input output devices could be used for this. Assuming the bb is already perfectly or near perfectly spherical as plastic ball bearings are, would there actually be wobble that would be perceivable? The machine proposed exists I'm very certain. It sounds like high rpm perhaps is a prerequisite before measurement?

 

[Compositepro]

Just shoot the BB's at a high rate through a barrel and gently capture them with a grid of traps. Keep the ones that consistently hit the same trap, which, presumably, would be the one at the center of the target. That will then sort them by any and all factors that affect accuracy.

 

[fortjp500]

I thought about that too. But distance would probably need to be pretty far. Thanks.

 

[EdStainless]

I have heard of devices that use air to 'float' parts and then ones that spin off center fall into 'reject' bins.
The balls roll down a "Vee" trough with a slot in bottom and a small air jet blows them into the air. With everything aligned so that balanced ones will fall straight in line and 'wobbly' ones off to the sides. It is very delicate to get lined up.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed

 

[fortjp500]

The air float idea seems possible to implement even with 3d printing/ rapid prototyping.

 

[fortjp500]

I am thinking that it would be reasonable to have a V section with no air to get the bbs rolling and then have a section with the bottom gone and air flow upwards to allow them to ride on a cushion of air. The only questions are:

Does width of the air cushion matter?
Does having a V that is "concave up" better in the air cushion portion or does having one that is concave down better?